Transcript International Muon Ionization Cooling Experiment MICE This talk: Measurements & measuring technique Next talk (Rob Edgecock): MICE – the UK Perspective. MICE Vittorio Palladino, RAL, 17 February.

International
Muon Ionization Cooling Experiment
MICE
This talk:
Measurements & measuring technique
Next talk (Rob Edgecock):
MICE – the UK Perspective.
MICE Vittorio Palladino, RAL, 17 February 2003
1
A detector physicist view of MICE
given a
~
10 % cooling engine
build for it
an unequivocal
..... a mysterious black blox ....
~
1% “diagnostics” tool
two complete (almost) identical detectors
(two spectrometers equipped with timing & PID)
one in front of and one in the back
of the box
each providing a complete measurement of all six parameters
of each particle
to measure the beam emittance
i.e. the ~10%
both in & out of the box
emittance reduction
with ~ 1% uncertainty
i.e. with an absolute error equal to 0.1%
MICE Vittorio Palladino, RAL, 17 February 2003
2
Quantities to be measured in a cooling experiment
cooling effect at nominal input
emittance ~10%
equilibrium emittance
curves for 21 MV, 3 full absorbers, particles on crest
  both output emittance & transmission decreases
...better FOM needed
MICE Vittorio Palladino, RAL, 17 February 2003
3
Quantities to be measured in a cooling experiment (ctd)
number of particles inside acceptance of subsequent accelerator
(for nominal NuFact, 15 mm rad, 30%)
Need to count muons coming in within an acceptance box and muons coming out.
must not only measure emittances, but also count particles
It is difficult to count very precisely particles of a given type in a bunch
and to measure emittance very precisely. =>
MICE Vittorio Palladino, RAL, 17 February 2003
single particle experiment
4
Emittance measurement
Each spectrometer measures 6 parameters per particle
x
y
t
x’ = dx/dz = Px/Pz
y’ = dy/dz = Py/Pz t’ = dt/dz =E/Pz
Normalized emittance n
by using (x, y, t, p/mc.dx/dz, p/mc.dy/dz, p/m.dt/dz)
Determines, for an ensemble (sample) of N particles, the moments:
Averages <x> <y> etc…
Second moments: variance(x)
sx2 = < x2 - <x>2 > etc…
covariance(x,y) sxy = < x.y - <x><y> >
Covariance matrix
 s 2x

 ...

...
M =
 ...

 ...
 ...

s xy
s xt
s xx'
s xy '
s 2y
...
...
...
...
...
...
...
s t2
...
...
...
...
s 2x'
...
...
...
...
s 2y '
...
s xt' 

s yt ' 

s tt ' 
s x't ' 

s y 't ' 
s t2' 
6D

 de t(M xytx 'y 't ' )
Evaluate emittance with:
 4 D  de t(M xyx 'y ' )   2
MICE Vittorio Palladino, RAL, 17 February 2003
Getting at e.g. sx’t’
is essentially impossible
with multiparticle bunch
measurements
Compare in with out
5
Single particle experiment
=>
truely thorough analysis
•Correlations between phase space parameters can be
easily measured.
•The detailed understanding of the role of each beam
parameter
energy
transverse momentum
RF phase (=time)
............
can easily be studied by making selection cuts in the
ensemble. A wide range of parameters being sampled,
any desired input beam conditions can be reconstructed
by appropriate slicing and/or reweighting of the
population of particles observed.
MICE Vittorio Palladino, RAL, 17 February 2003
6
requirements on detector system:
1. tag each particle considered: select muon, in & out !
in …… reject e, p, p
=> TOF 2 stations 10 m flight with 70 ps resolution
& up-Cerenkov, if life were tough
out …… reject e => down-Cerenkov + em Calorimeter
2. measure all 6 particle parameters
i.e. x,y,t, px/pz , py/pz , E/pz
3. each with resolution  10% ( 1% in quadrature)
of width around equilibrium emittance …
… ie adequate to preserve unsmeared estimates
of variances & covariances .. emittance !!!
4. stand possibly severe noise level from RF cavities
MICE Vittorio Palladino, RAL, 17 February 2003
7
10% cooling of 200 MeV muons requires ~ 20 MV of RF
single particle measurements =>
measurement precision can be as good as D (  out/ in ) = 10-3
never done before either….
Coupling Coils 1&2
Spectrometer
solenoid 1
Matching
coils 1&2
Focus coils 1
Focus coils 2
Focus coils 3
Matching
coils 1&2
Spectrometer
solenoid 2
m
Beam PID
TOF 0
Cherenkov
TOF 1
Diffusers 1&2
RF cavities 1
RF cavities 2
Liquid Hydrogen absorbers 1,2,3
Incoming muon beam
Downstream
particle ID:
TOF 2
Cherenkov
Calorimeter
Trackers 1 & 2
measurement of emittance in and out
MICE Vittorio Palladino, RAL, 17 February 2003
8
need of simply matching with the cooling section
and of keeping a large-emittance beam in a small physical volume
=>
solenoidal spectrometer coaxial to cooling channel
Blondel, Janot 01
two “timing helicometers”
Each measures, very precisely, & time stamps
the helicoidal trajectory of each individual muon
Each a precise
tracker device
inside a 4 T solenoidal magnetic field
matching the transverse diameter of the muon beam being cooled (30 cm)
capable also of precise (70 ns)
Difference is in the
PID systems
timing
emphasis on p/m separation in front of the front detector
fighting beam contamination
on m/e separation in the back of the back detector
fighting muon decay
MICE Vittorio Palladino, RAL, 17 February 2003
9
Tracking & timing the helices
need of simply matching with the cooling section
keeping a large-emittance beam in a small physical volume
=> chose
solenoid magnets coaxial to the cooling channel
Blondel, Janot 01
B Field
(x,y) measurements at 3 different z values in principle sufficient
length such that average m makes about (2/3 of) a turn
ie about 1 m for 200 MeV/c in a 4 T solenoid
70 ps TOF measures phase of m to 5° wrt the 201 MHz RF
MICE Vittorio Palladino, RAL, 17 February 2003
10
TRANSVERSE MOMENTUM RESOLUTION spt
Performance
= 110 keV
Pt/Pz
E/Pz

Pz resolution degrades at low pt :
0.06 % resolution
resolution in E/Pz is much better behaved

0.06 % resolution
measurement rms is 4% of beam rms
11
MICE Vittorio Palladino, RAL, 17 February 2003 … also ≈100 mm spatial resolution real small
MICE simulations & general sofware effort
A fast simulation (DWARF4), including dE/dx & MS,
was used for the basic design …..
emittance generation
tracking
particle identification
................
since the time of the LOI
A more complete GEANT4 simulation (G4MICE)



long term foundation of the MICE sofwtare, including everything
true end-to-end simulation of BOTH cooling AND detectors
complete description of the detectors, down to digitization
first repository of reconstruction algorithms
now already providing first results
MICE Vittorio Palladino, RAL, 17 February 2003
12
background
from the nearby high-gradient
solenoidal geometry...
one major
RF cavities
drawback for trackers
exposed almost directly to a large dark current
and x-ray
however
background
i) background moderate at RF gradient of 8.3 MV/m
baseline due to limited availability of RF power
ii) LiH absorbers do absorb dark current e- completely,
only x-rays only through
iii) the detectors are built of low-Z material
x-rays Compton hits are less frequent
& distinguishable from m triplets
it appears that
the performance of the detectors will not be affected
studies of this background continue .... with high priority !
did we ever have the RF power ...
MICE Vittorio Palladino, RAL, 17 February 2003
13
Estimating the background
measure 4 x 104 Hz/cm2 e rate at 8 MV/m in lab G
805 MHz pillbox cavity
scaling to 201 Mhz by the volume ratio of cavities
(emitter area proportional to surface, energy to length)
gives total e rate of 30 MHz
for detector radius 15 cm (an area of 700 cm2)
converted to photons by absorber (efficiency = Re/X0 ~ 0.07)
Re e range, X0 rad length
0.26% Compton scatters in the detector per plane
6 fibre ribbons 0.035 cm thick X0 45 cm
rate about
0.5 MHz/plane
(admittedly large uncertainty)
about 2-3 orders of magnitude below levels
found tolerable by G4MICE studies
... while materials & surface treatments also being investigated
MICE Vittorio Palladino, RAL, 17 February 2003
14
Backgrounds
measured dark currents
real background
reduced by factor
L/X0(H2) . L/X0(det)
0.07
0.0026
Extrapolation to MICE (201 MHz):
scale rates as (area.energy) X 100
and apply above reduction factor 2 10-4
Dark current backgrounds measured
on a 805 MHz cavity in magnetic field
with a 1mm scintillating fiber at d=O(1m)
MICE Vittorio Palladino, RAL, 17 February 2003
4 104 Hz/cm2 @ 8 MV/m @805 MHz
0.8 kHz/cm2 per sci-fi
 500 kHz/plane
! within  one order of magnitude !
15
=> 6.70 MV/m
=> 4.65 MV/m
.43 X 4 cells = 1.7 m  11.5 MV for 1X 4 = 6.70 MV/m
16 MV for 2X4 = 4.65 MV/m
16 MV for 1X4 = 9.3 MV/m
MICE Vittorio Palladino, RAL, 17 February 2003
16
Tracker
Baseline Option:
Sci Fi tracker – 5 stations with 3 crossing double
planes of 350 mm fibres
readout by VLPC (High Q.E. and high gain) as in D0.
Pro: we are essentially sure that this will perform well enough for MICE.
Con: 43000 channels make it quite expensive (4.1 M€)
2 cost saving alternatives, under energic investigation
:
Option 1 Sci-Fi
-- Reduce channel count by multiplexing channels (1/7)
Pro: cheaper by factor 4.
Con: less sure to work in presence of Bkg and a few dead channels.
Option 2
Use a Helium filled TPC with GEM readout (‘TPG’)
Pro: very low material budget (full TPG = 2 10-4 X0) and lots of points/track (100)
cheaper (<0.5 M€ of new money)
Con: long integration time (50 ms vs 10 ns) => several muons at a time + integrated noise
effect of x-rays on GEMs themselves unknown.
MICE Vittorio Palladino, RAL, 17 February 2003
17
Tracker Baseline Option:
Sci Fi tracker – 5 stations with 3 crossing double
planes of 350 mm fibres
readout by VLPC (High Q.E. and high gain) as in D0
… see MUSCAT also
Pro: we are essentially sure that this will perform well enough for MICE.
Con: 43000 channels make it quite expensive, if read out individually (4.1 M€)
MICE Vittorio Palladino, RAL, 17 February 2003
18
Sci-Fi studies
with
G4MICE
concept
background simulated 1000 times larger
that extrapolated from measurements
does NOT degrade resolution seriously
(No multiplexing here!)
in
further improvements from the RF cavity
will be seeked (TiN coating) etc…
out
MICE Vittorio Palladino, RAL, 17 February 2003
looks good.
if this is confirmed, one could envisage
running with higher gradients.
This is possible if
-- 8 MW power to one 4-cavity unit
-- LN2 operation
19
Cost saving alternatives:
Option 1
Multiplex Sci-Fi -- Reduce channel count (1/7)
Seven 350 mm fibres multiplexed by a 7:1 multiplexing wave guide into one
VLPC pixel.
only be about a 10% loss of light in the outer ring of fibres due to a
slight mismatch.
Pro: cheaper by factor 4 …. save 3 M !!!!!
Con: less sure to work in presence of bkgd and of a few dead channels.
OK if backgrounds from the RF cavities stays at a low level
MICE Vittorio Palladino, RAL, 17 February 2003
20
Cost saving alternatives:
TPG
………… a 90% Helium filled
Novel:
GEM amplification
TPC
Option 2
1 m long and 30 cm Ø
and
hexaboard R/O
Pro: very low material budget (full TPG = 2 10-4 X0)
lots of points/track (>100) oustanding pattern recognition
cheap (<0.5 M€ of new money)
Con: long integration time (50 ms vs 10 ns)
=> several muons at a time + integrated noise
effect of x-rays on GEMs themselves …. still to be ested
MICE Vittorio Palladino, RAL, 17 February 2003
21
at noise rate similar to that
simulated for fibers,
no difficulty finding tracks
and measuring them.
resolution somewhat better
than sci-fi (which is good
enough)
difficulty: nobody knows the effect of RF photons on the GEM themselves
tests in 2003, decision October 2003
MICE Vittorio Palladino, RAL, 17 February 2003
22
MICE Vittorio Palladino, RAL, 17 February 2003
23
basic TOF design :
upstream t0
tin
tout
1212
4040
cm2
trigger ... simple doubles & triples ....
70 ps rms ..... 1% p/m ID ( 1400 ps DTOF ..)
+ 5° timing wrt the RF phase
TOF0
TOF1&2
Bicron BC-420 plastic scintillator Hamamatsu R4998 PMTs
BC-404
+ dedicated laser calib system [HARP]
MICE Vittorio Palladino, RAL, 17 February 2003
fine mesh Hamamatsu R5505
OK in 1 Tesla fringe fields
thou reduced gain BESS]
24
MICE Vittorio Palladino, RAL, 17 February 2003
25
Particle Identification
Both residual beam pions
and muon decay electrons
fake different kinematics
spoil and bias the emittance measurements
must have < 0.1% of either in the sample
PID rejection must be strong & redundant
both up and down stream
MICE Vittorio Palladino, RAL, 17 February 2003
26
Upstream Particle Identification
Even with a solenoid decay channel, the beam is not pure.
Entering electrons and protons have different cooling properties and must be
rejected (easy)
Pions are more of a problem since they can decay in flight with a daugther of
different momentum ………………. big bias in emittance!
need less than 1 ‰ pion contamination in final sample.
TOF in the beam line with 10m flight path and 70 ps resolution provides
p/m separation better than 1% @ 300 MeV/c (Dt = 1400 ps ! )
 it is sufficient that the beam has fewer than 10% pions
A small beam Cherenkov is foresen to complete the redundancy in this system
(overlaps …. or higher p/m …)
a 5 ” Ø vessel of liquid C6F14 (n=1.25), 30 cm long in total
MICE Vittorio Palladino, RAL, 17 February 2003
27
Downstream PID
0.5% of m decay in flight ………… large bias if a forward-decay e is kept as m
two systems are foreseen to get to eliminate electrons below 10-3 :
Electromagnetic Calorimeter (mip = em shower 27 MeV, use also z profile)
Aerogel Cherenkov ( n=1.02, m blind , threshold well above beam momentum )
Positive Identification of a
particle in the calorimeter
consistent with a muon
AND no electron signal in
the Cherenkov
MICE Vittorio Palladino, RAL, 17 February 2003
28
PRECISION
1. statistical errors
--emittance is measured to 10-3 with ~106 muons.
-- ratio of emittances to same precision requires much fewer (105)
a nice unexpected surprise
(re)measuring again the same m , after little material
… in & out strongly correlated
statistical fluctuations largely cancel in the ratio
-- Due to RF power limitations we can run about 10-3 duty factor 1ms/s
-- To avoid muon pile up we want to run at ~1 muon per ISIS bunch (1/330ns)
3000 muons /s
-- The emittance generation must cover cooling acceptance uniformly
-- 25% of incoming muons within acceptance
-- about 1/6 mimic a bunch on crest => MICE cools
100 muons/s .
A 10-3 measurement of emittance ratio will take 103 s … 1/3 hour !!
N.B. this assumes a beam line with a solenoid to be obtained from PSI.
A quadrupole channel has more background and less rate, and would lead to a
time longer by a factor ~10.
MICE Vittorio Palladino, RAL, 17 February 2003
29
1. statistical errors
--emittance is measured to 10-3 with ~106 muons.
--ratio of emittances to same precision requires much fewer (105)
a nice surprise, unexpected
measuring again the same m, after little material
… in & out strongly correlated
statistical fluctuations largely cancel in the ratio
-- To avoid muon pile up we want to run at ~1 muon per ISIS bunch (1/330ns)
3000/s
-- The emittance generation must cover keeps
25% of incoming muons within acceptance
about 1/6 are on crest … a bunch … =>
100 good muons per second.
-- Due to RF power limitations we can run about 10-3 duty factor 1ms/s
A 10-3 measurement of emittance ratio will take 106 s < one hour
N.B. this assumes a beam line with a solenoid to be obtained from PSI.
A quadrupole channel has more background and less rate, and would lead to
a time longer by a factor ~10.
MICE Vittorio Palladino, RAL, 17 February 2003
30
ACCURACY
systematic errors
MICE-FICTION:
. MICE measures e.g.
compares with
(out / in)exp
= 0.894 ± 0.001 (statistical)
(out / in)theory. = 0.885
and tries to understand such a ~ 0.01 difference
in emittance reduction
SIMULATION
A. theory systematics:
modeling of cooling cell
is not as reality
REALITY
B. experimental systematics:
modeling of spectrometers
is not as reality
MEASUREMENT
MICE Vittorio Palladino, RAL, 17 February 2003
31
No systematic must affect the expected 10% cooling effect
by more than 10–3 absolute, i.e., 1% of its value.
The errors of class A (modeling the cooling cell )
Errors in this category include:

Uncertainties in the thickness or density of the liquid-hydrogen absorbers
and other material in the beam

Uncertainties in the value and phase of the RF fields

Uncertainties in the value of the beta function at the location of the absorbers
Misalignment of the magnetic elements
Uncertainty in the beam energy scale
Uncertainties in the theory (M.S. and dE/dx and correlation thereof)
All errors of type A become more important near the equilibrium emittance.
The errors of class B (modeling the pair of detectors ):
systematic differences between incoming and outgoing measurement devices
different efficiency
different misalignments
possible differences in the magnetic field of the two spectrometers
will constrain them heavily
.... got a full arsenal of ancillary measurements
Step III or
run MICE empty with no RF or
analyse cooling vs muon phase (free) or .....
MICE Vittorio Palladino, RAL, 17 February 2003
32
Most critical is the control of the magnetic fields.
For this reason MICE will be equipped with a set of magnetic measurement devices
that will measure the magnetic field with a precision much better than 10–3.
Magne t ic
senso rs
SC Coils
3 hall probes
Magnetic measurements
design of system and
procedures by Saclay
(Rey, Chevallier)
Posit ioning holes
MICE Vittorio Palladino, RAL, 17 February 2003
NIKHEF will provide the probes
(Linde)
33
m
-
STEP I:
2004
STEP II: summer 2005
STEP III:
winter 2006
STEP IV: spring 2006
STEP V: fall 2006
STEP VI:
2007 full power
MICE Vittorio Palladino, RAL, 17 February 2003
34
The statistical precision will be very good and there will be many handles against
systematics.
We believe that the systematic errors on the measurement of the ratio of
emittances can be kept below 10-3
This will require careful integration of the acquisition of data from the
spectrometers and from the cooling cell.
A lot remains to be done in this area, admittedly, to make sure
that MICE has foreseen the necessary diagnostics by the time it turns on.
MICE Vittorio Palladino, RAL, 17 February 2003
35
MICE Vittorio Palladino, RAL, 17 February 2003
36
MICE Vittorio Palladino, RAL, 17 February 2003
37
Next Future
1. Prototype critical detectors (trackers) in 2003
Test beds … Fermi … KEK …. CERNPS …
bkgrounds
…… Sci Fi
…..TPG
2. Choose
helicometer option Oct 31 (CollMeet)
3. Start build late this year … more standard items
………trigger&TOF, PID subdetectors
4. Be there first ….
…. well understood by the time
the cooling engine arrives
MICE Vittorio Palladino, RAL, 17 February 2003
38
Conclusions
The MICE detector system
an integrated tool
…
a pair of challenging particle detectors…
capable to measure reduction of emittance
at 1% level
appears feasible
first to be studied
a full cell of the US NuFact design studyII
including a number of beam & optics conditions beyond the baseline
The collaboration has organised responsibilities, determined the basic costs and time
line of the construction and commissioning of the two detectors .
MICE Vittorio Palladino, RAL, 17 February 2003
39
requirements on detector system:
1. tag each particle considered: select muon, in & out !
in …… reject e, p, p
=> TOF 2 stations 10 m flight with 70 ps resolution
& up-Cerenkov 1/2” Ø, 30 cm liquid C6F14 (n=1.25)
out …… reject e => down-Cerenkov + em Calorimeter
cell
2. measure all its six parameters
i.e. x,y,t, px/pz , py/pz , E/pz
3. each with resolution better than 10% of
width at equilibrium emittance (correction less than 1%) …
… ie adequate not to smear (spoil sensitivity to)
variances & covariances
…...
emittance !!!
4. Be robust against noise from RF cavities
MICE Vittorio Palladino, RAL, 17 February 2003
40
Further Explorations
We have defined a baseline MICE, which will measure the basic cooling
properties of the StudyII cooling channel with high precision, for a moderate
gradient of ~8 MV/m, with Liquid Hydrogen absorbers.
Many variants of the experiment can be tested.
1. other absorbers: Various fillings and thicknesses of LH2 can be envisaged
The bolted windows design allows different absorbers to be mounted.
2. other optics and momentum: nominal is 200 MeV/c and b  42 cm.
Exploration of low b (down to a few cm at 140 MeV/c)
Exploration of momentum up to 240 MeV/c
will be possible by varying the currents.
3. the focus pairs provide a field reversal in the baseline configuration, but
they have been designed to operate also in no-flip mode which could have larger
acceptance both transversally and in momentum (Fanchetti et al)
(We are not sure this can be done because of stray fields…)
4. Higher gradients can be achieved on the cavities, either by running them
at liquid nitrogen temperature (the vessel is adequate for this) (gain 1.5-1.7)
or by connecting to the 8 MW RF only one of the two 4-cavity units (gain 1.4)
MICE Vittorio Palladino, RAL, 17 February 2003
41
Time Lines
… to be reformulated
Time lines for the various items of MICE have been explored
– procurement delays and installation –
the critical items are solenoids and RF cavities.
If funding is adequate, the following sequence of events can be envisaged * ->
consistent with the logistics of the beam line upgrade at RAL and of the various
shut downs.
Muon Ionization cooling will have been demonstrated ands measured precisely by
2007
At that time
MINOS and CNGS will have started and mesured Dm132 more precisely
J-Parc-SK will be about to start up (Q13)
LHC will be about to start as well
It will be timely (…and not too soon!) to have by then a full design for a
neutrino factory, with one of the main unknowns (practical feasibility of
ionization cooling) removed.
MICE Vittorio Palladino, RAL, 17 February 2003
42
m
-
STEP I:
2004
STEP II: summer 2005
STEP III:
winter 2006
STEP IV: spring 2006
STEP V: fall 2006
STEP VI:
2007
MICE Vittorio Palladino, RAL, 17 February 2003
43