Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
C. C
iobanu
Page 1W
IN’03
Single Top Quark Physics at the Tevatron and
Beyond
Catalin
Cio
ban
u, U
niversity o
f Illino
is, CD
F
Fo
r the L
HC
part, valu
able h
elp received
from
:
Serg
ey Sab
losp
itsky, IH
EP
Pro
tvino
, CM
S
Du
gan
O’N
eil, Sim
on
Fraser U
niv., A
TL
AS
WIN
’03, Octo
ber 7, 2003
C. C
iobanu
Page 2W
IN’03
Single Top Productionl
Why look for single top:
‰C
ross section ~ | V
tb | 2: the only w
ay todirectly m
easure CK
M m
atrix el. Vtb
‰S
ingle top is background to othersignals: e.g. H
iggs searches.‰
Test non-S
M phenom
ena:ÿ
Heavy W
’ bosonÿ
Anom
alous Wtb
couplingsÿ
FC
NC
couplings tug or tcg
lA
t the Tevatron relevant channels are:
‰t-channel W
-gluon fusion:ÿ
1.98 pb at ÷s=
1.96 Te
Vÿ
Hard b-jet, W
decay products, soft b(usually lost), light quark jet
‰s-channel W
*:ÿ
0.88 pb at ÷s=1.96
TeV
ÿT
wo hard b-jets, W
decay products
‰B
.W. H
arris et al.: Phys. R
ev. D 66, 054024
(theoretical calculations)
C. C
iobanu
Page 3W
IN’03
From Run I to Run II
lR
esults at ÷s = 1.8 T
eV from
CD
F and D
∅ :
‰S
ingle top has not been observed; 95% C
L limits w
ere set:‰
t-channel: D∅
limit: 22
pb, C
DF
limit: 13
pb (theoretical x-section: 1.40 pb
)
‰s-channel: D
∅ lim
it: 17 pb
, CD
F lim
it: 18 pb (theoretical x-section: 0.76
pb)
‰com
bined s- and t- channels search: CD
F lim
its: 14 pb (H
T , Mlnb )
‰m
ost recent Run 1 C
DF
study finds a 2.2s excess and lim
it of 24 pb (7-inp N
N)
‰D
∅:P
hys. Lett. B 517, 282 (2001); C
DF
: PR
D65, 091102 (2002), C
. Ciobanu
’s Ph.D
. thesis
lF
or Run II:
‰H
igher rate: 32% increase in the com
bined s and t channel x-section:ÿ
still small: 2.9 pb versus 6.7 pb for top pair production
‰M
ore luminosity – right now
CD
F has 180 pb
-1 – that is 70% m
ore data!‰
Better detector acceptance
‰P
reliminary C
DF
analysis done for 107 pb
-1 of data:ÿ
Search for s- and t-channels com
bined productionÿ
Seaparate
search for t-channel
C. C
iobanu
Page 4W
IN’03
Combined Search
lLook in the W
+2 tight jets channel:
‰“tight” jet: E
T >
15 GeV
, |h| <
2.8‰
at least one SV
X B
-tag‰
exactly one lepton with E
T >20 G
eV, |h
|<2.0
‰m
issing energy: ET >
20 GeV
‰veto Z
’s, dilepton events‰
apply a top mass w
indow cut: 140 <
Mlnb <
210 GeV
/c2
lM
onte Carlo sam
ples:‰
single top signal:ÿ
Pythia
(1M events)
ÿM
adevent+P
ythia (~
200k)
‰H
ER
WIG
tt background‰
ALP
GE
N+
HE
RW
IG W
bb used to model the shapes of the
non-top backgrounds: Wbb, W
cc, Wc, m
istags, non-W events
C. C
iobanu
Page 5W
IN’03
Expected Yieldl
For L=
107.1 pb-1:
‰expect 2.4 signal events
‰expect 16.1 backgrd events
‰observe 19 (expect 18.5)
lT
he variable with the m
ostdiscrim
ination power is H
T ,the total transverse energy inthe event:‰
Will fit the H
T shape in data asthe sum
of weighted M
Cshapes.
‰U
se maxim
um likelihood
method w
ith Gaussian
background constraints
ProcessN
in 107.1 pb-1
W-gluon
1.61 ± 0.51
W*
0.80 ± 0.18
tt2.33 ± 0.70
Wbb
3.5 ± 1.5
Wcc
1.4 ± 0.8
Wc
3.1 ± 1.0
Mistags
2.9 ± 0.9
Non-W
2.2 ± 0.7
Diboson
0.7 ± 0.2
Z+bb
0.07 ± 0.03
Total non-top
13.8± 2.8
C. C
iobanu
Page 6W
IN’03
Fitting HT distributions
lH
T distributions for signal and backgrounds:‰
signal shapes similar for s- and t- channels
‰fit yields 2.9
±4.5 signal events (exp. 2.4 events).
C. C
iobanu
Page 7W
IN’03
95% C.L. lim
itl
Limit at 95%
C.L. is 17.5 pb
lT
his accounts for the systematics:
‰Jet E
T scale: 20.9%‰
ISR
/FS
R: 4.0%
‰T
op Mass: 6.2%
‰P
DF
, signal generator, backgroundm
odel 1-2%
lC
ompare to R
un I:‰
Sam
e luminosity 107 vs 106 (p
b-1)
‰14 pb lim
it for 2.16 pb x-section‰
Now
: 17.5 pb limit for 2.86 pb x-s.
marginally better lim
it!
‰S
everal ways to im
prove this will be
discussed later.
C. C
iobanu
Page 8W
IN’03
Separate Searchl
Looking for the two signal channels individually is desired:
‰D
ifferent sensitivities to new physics:
ÿs-channel: heavy charged vector bosons W
’, CP
-violation effects within
MS
SM
, Kaluza
-Klein excited W
-boson within M
SS
M
ÿt-channel: F
CN
C couplings, anom
alous V+
A contributions to the W
-t-bvertex, etc.
ÿT
o go back to extracting Vtb w
e will have to know
the individual rates
lW
e use the same selection as in the com
bined search.‰
How
ever, using HT is no longer appropriate
‰A
variable with good W
* - W-gluon separation potential is Q
x h, w
hereQ
is the lepton charge and h the pseudorapidity of the non-b jet.
‰N
ow treat W
* as a background, therefore perform a 4-com
ponent fit
C. C
iobanu
Page 9W
IN’03
Fitting Q¥h
lt-channel exhibits an asym
metry tow
ard Qxh
positive values (twice as
many events than in the negative Q
xh range)
l fit to data yields: 1.9± 3.7 t-channel signal events (exp. 1.6 events).
C. C
iobanu
Page 10W
IN’03
95% C.L. lim
itl
Limit at 95%
C.L. is 15.4 pb
lT
his accounts for the systematics:
‰Jet E
T scale: 21.5%
‰S
ignal generator: 21.1%
‰B
ackground model: 43.0%
‰IS
R/F
SR
: 6.9%
‰T
op Mass: 6.7%
‰P
DF
4.3%
lC
ompare to R
un I:‰
13 pb limit for 1.40 pb x-section
‰N
ow: 15.4 pb lim
it for 1.98 pb x-s.
again a better limit!
C. C
iobanu
Page 11W
IN’03
Future Prospectsl
Neural N
etwork w
ith 5 input variables: HT , E
T (j1), ET (j2), E
T , PT (j1-j2),
Looking in the W+
2j channel, no M
lnb cutA
ssuming the background uncertainties w
ill scale as L-1/2
lR
ight now: 3 sigm
a with 2.5 fb
-1
lIm
provements to com
e. We try:
‰S
oft lepton tagger, Jet Prob. tagger
‰U
se of forward electrons
‰B
etter discriminant variables:
ÿM
atrix elements (or ratios of these)
‰U
se the top mass hypothesis to
constrain the eventÿ
Will lead to better M
lnb resolution
‰B
etter understanding of thesystem
atic uncertainties is crucial.
C. C
iobanu
Page 12W
IN’03
Single Top at the LHC
lLH
C w
ill be a top quark factory:‰
8 million top pairs per experim
ent per year (10 fb-1 / year)
‰S
ome cross section values for ÷s =
14 T
eV:
ÿtop pair production: ~
800 pb (mostly via gluon-gluon fusion)
ÿt-channel single top: 153 (top) and 90 (antitop
) = 243 pb
ÿe.g
. per d
ay ~6000 even
ts, at 1033cm
-2s-1
ÿs-channel single top: 6.6 (top) and 4.8 (antitop
) = 11
pb
ÿassociated W
t production: 50-60 pb
ÿN
eglig
ible at th
e Tevatro
n
ÿW
ill discuss each of the three modes in w
hat follows
C. C
iobanu
Page 13W
IN’03
t-channel single top (CMS)
lC
MS
study (hep-ph/0003033):
‰F
ull calorimeter sim
ulation + b-tag efficiency param
etrization
‰P
ythia 5.72 signal, tt, W
Z; V
ecbos+H
erwig W
+jets. S
election:ÿ
One isolated lepton w
ith PT >
20 GeV
/c, |h|<
2.5
ÿM
issing ET >
20 GeV
, 50 < M
ln <100 G
eV/c
2
ÿT
wo jets w
ith ET
> 20
GeV
, |h| <
4.0
ÿO
ne jet: E
T >
20 G
eV, |h
| < 2.5, th
e oth
er ET
> 50 G
eV, 2.5<
|h| <
4.0
ÿL
eadin
g jet E
T <100 G
eV (to
redu
ce tt)
ÿE
xactly on
e b-tag
ged
jet – the cen
tral on
e
ÿR
eject WZ
candidates with: 80 <
Mjj <
100 GeV
/c2
‰S
ignal peak visible in reconstructed Mtop distribution
‰W
ith 10 fb-1, ÷(S
+B
)/S =
1.4%
lT
he quest continues – newer det. sim
ulation‰
use fancier generators: TopR
ex, Single T
op‰
Work-in-progress, unofficial results ds
/ s ~
10%
V.Abram
ov, CMS Week 9
/17/03
C. C
iobanu
Page 14W
IN’03
lA
TLA
S study of W
-gluon single top (Dugan O
’Neil’s P
h.D. thesis):
‰A
TLF
AS
T param
etrized detector simulation
‰O
netop+
Pythia
5.72 signal, tt; M.E
.+H
ER
WIG
Wbb
, Wjj
‰S
elections are rather similar to C
MS
study, plus:ÿ
150<M
top <200
GeV
/c2
ÿR
esult: ÷(S+
B)/S
= 0.9%
with 10 fb
-1.
lW
hat about Vtb ? E
xtracting |Vtb | 2 from
x-section picks up uncertainties:‰
x-section uncertainty:ÿ
Statistical: 5%
is a perhaps reasonable/optimistic guess for L=
10 fb-1
ÿS
ystematic <
20%. H
ere things can easily be off by a lot, especially in the 1st year…
‰Lum
inosity uncertainty: traditionally 4-5%
‰T
heoretical uncertainty:ÿ
From
factorization and renormalization scale dependence (3%
from N
LO calculation)
t-channel single top (ATLAS)
C. C
iobanu
Page 15W
IN’03
lC
MS
study (S. S
ablospitsky, CM
S w
eek 9/17/03)‰
disclaimer: w
ork-in-progress, not yet ‘official’ results
‰use T
opRex to generate single top, tt, W
bb. Pythia W
+jets. U
se CM
SJE
T 4.801
‰selection:ÿ
one lepton with E
T >10 G
eV, |h
|<2.5
ÿ2 B
-tags with E
T >20 G
eV, no other jets w
ith ET >
20GeV
ÿV
ector sum all final state P
T : |SP
T |<15 G
eV
ÿR
econst. top mass: 150 G
eV<
Mtop <
200 GeV
/c2
ÿR
esults: ÷(S+
B)/S
= 12%
, S/B
= 8%
for L = 30 fb
-1
ÿw
hich implies d|V
tb |=8.3%
s-channel single top (CMS)
C. C
iobanu
Page 16W
IN’03
s-channel single top (ATLAS)
lA
TLA
S study (hep
-ph/0003033)‰
Onetop
+P
ythia 5.72 signal, tt; M
.E.+
HE
RW
IG W
bb, W
jj
‰S
election:ÿ
One isolated lepton w
ith PT >
20 GeV
/c, |h|<
2.5
ÿ2 b-jets w
ith ET >
75 GeV
ÿS
calar sum all final state P
T : |SP
T |<175 G
eV
ÿR
econst. top mass: 150
GeV
<M
top <200 G
eV/c
2
‰R
esults: ÷(S+
B)/S
= 5.5%
for L = 30
fb-1, d|V
tb |~1%
lW
t single top production at AT
LAS
:‰
65< M
jj < 95 G
eV/c
2
‰E
vent invariant mass
< 300 G
eV/c
2
‰÷(S
+B
)/S=
4.4% for L =
30 fb-1
C. C
iobanu
Page 17W
IN’03
Top Polarization in Single TopEvents
lE
W-produced top quarks are highly polarized:
‰In the top rest fram
e, its spin points along the direction of the down (d) quark.
‰R
estrict to t-channel single top:80% (69%
) of the top (antitop) events have the
‘d’ quark in the final state. The angular distrib
. of the charged lepton:
‰used signal sam
ples with P
=+
1 and P=
-1. Consider only W
+jet background.
‰C
hisquare fit, letting the two P
contributions float.
‰T
he error on polarization measurem
ent is 1.6% for 10 fb
-1 (Dugan
’sPh
.D.thesis)
lW
boson helicity measurem
ent:‰
Again restrict to t-channel signal.
‰M
easure cos(yl ) distribution. y
l between the direction of the lepton in W
restfram
e and the direction of the W in the top rest fram
e.
‰T
hree component fit: fL , fR , fLong , w
ith fL +fF
R +fLong =
1
‰U
ncertainties of the order 2-3% for 30 fb
-1 of data.
()l
lq
qcos
12 1
)(cos
Pf
+=
C. C
iobanu
Page 18W
IN’03
Conclusionsl
CD
F and D
ø accum
ulated more events than in entire R
un I (factor of 2).l
Prelim
inary CD
F study from
107 pb-1 sets prom
ising single top limits
lH
owever, m
any improvem
ents needed if we are to observe single top in
the next 2-3 years.l
Precise m
easurements of |V
tb |, as well as top polarization are not w
ithin(im
mediate) sight
lLH
C w
ill see very exciting results in the first (few) years of running:
‰10 –30
fb-1 needed for observation in m
ost searches.‰
Higher collider energy =
considerable higher cross sections
lA
TLA
S and C
MS
single top studies are very mature
‰T
he Monte C
arlo efforts are particularly strong: Top R
ex, SingleT
op.
lC
DF
/Dø
can help with insight into system
atics. Search strategies are
already similar betw
een Tevatron and C
ER
N, but big im
provements are
just an idea away!