A High Statistics Measurement of the Lambdac+ Lifetime

5 Pages·0.14 MB·English

by  J. M. Link

FERMILAB-Pub-02/015-E
A High Statistics Measurement of the Λ+ Lifetime
c
J. M. Link,1 M. Reyes,1 P. M. Yager,1 J. C. Anjos,2 I. Bediaga,2 C. G¨obel,2 J. Magnin,2 A. Massafferi,2
J. M. de Miranda,2 I. M. Pepe,2 A. C. dos Reis,2 S. Carrillo,3 E. Casimiro,3 E. Cuautle,3 A. Sa´nchez-Herna´ndez,3
C. Uribe,3 F. Vazquez,3 L. Agostino,4 L. Cinquini,4 J. P. Cumalat,4 B. O’Reilly,4 J. E. Ramirez,4 I. Segoni,4
J. N. Butler,5 H. W. K. Cheung,5 I. Gaines,5 P. H. Garbincius,5 L. A. Garren,5 E. Gottschalk,5 P. H. Kasper,5
A. E. Kreymer,5 R. Kutschke,5 S. Bianco,6 F. L. Fabbri,6 A. Zallo,6 C. Cawlfield,7 D. Y. Kim,7 A. Rahimi,7
J. Wiss,7 R. Gardner,8 A. Kryemadhi,8 Y. S. Chung,9 J. S. Kang,9 B. R. Ko,9 J. W. Kwak,9 K. B. Lee,9 H. Park,9
G.Alimonti,10 M.Boschini,10 P.D’Angelo,10 M.DiCorato,10 P.Dini,10 M.Giammarchi,10 P.Inzani,10 F.Leveraro,10
S. Malvezzi,10 D. Menasce,10 M. Mezzadri,10 L. Milazzo,10 L. Moroni,10 D. Pedrini,10 C. Pontoglio,10 F. Prelz,10
M. Rovere,10 S. Sala,10 T.F. Davenport III,11 V. Arena,12 G. Boca,12 G. Bonomi,12 G. Gianini,12 G. Liguori,12
M. M. Merlo,12 D. Pantea,12 S. P. Ratti,12 C. Riccardi,12 P. Vitulo,12 H. Hernandez,13 A. M. Lopez,13 E. Luiggi,13
H. Mendez,13 L. Mendez,13 A. Mirles,13 E. Montiel,13 D. Olaya,13 A. Paris,13 J. Quinones,13 C. Rivera,13
2 W. Xiong,13 Y. Zhang,13 J. R. Wilson,14 K. Cho,15 T. Handler,15 R. Mitchell,15 D. Engh,16 M. Hosack,16
0
0 W. E. Johns,16 M. Nehring,16 P. D. Sheldon,16 K. Stenson,16 E. W. Vaandering,16 M. Webster,16 and M. Sheaff17
2 (FOCUS Collaboration)

n 1University of California, Davis, CA 95616
a 2Centro Brasileiro de Pesquisas F´isicas, Rio de Janeiro, RJ, Brazil
J 3CINVESTAV, 07000 M´exico City, DF, Mexico

1 4University of Colorado, Boulder, CO 80309
3 5Fermi National Accelerator Laboratory, Batavia, IL 60510
6Laboratori Nazionali di Frascati dell’INFN, Frascati, Italy I-00044

1 7University of Illinois, Urbana-Champaign, IL 61801
v 8Indiana University, Bloomington, IN 47405
1 9Korea University, Seoul, Korea 136-701
0 10INFN and University of Milano, Milano, Italy
0 11University of North Carolina, Asheville, NC 28804
2 12Dipartimento di Fisica Nucleare e Teorica and INFN, Pavia, Italy
0
13University of Puerto Rico, Mayaguez, PR 00681
2
14University of South Carolina, Columbia, SC 29208
0
15University of Tennessee, Knoxville, TN 37996
/
x 16Vanderbilt University, Nashville, TN 37235
e 17University of Wisconsin, Madison, WI 53706
-
(Dated: February 3, 2008)
p
e AhighstatisticsmeasurementoftheΛ+c lifetimefromtheFermilabfixed-targetFOCUSphotopro-
h
ductionexperimentispresented. Wedescribetheanalysistechniquewithparticularattentiontothe
:
v determinationofthesystematicuncertainty. Themeasuredvalueof204.6±3.4(stat.)±2.5(syst.)fs
i from 8034±122 Λc →pKπ decaysrepresentsa significant improvement overthepresent world av-
X
erage.
r
a
PACSnumbers: 13.30.Eg,14.20.Lq,14.65.Dw
Experimentalmeasurementsofcharmparticlelifetimes applicability.
havebeenusedinthestudyofstronginteractionphysics. Precise charm lifetime measurements are now begin-
The measurements provide some guidance for theoret- ning to emerge from e+e− collider experiments [3, 4].
ical calculations of non-perturbative strong interaction The effects of lifetime and vertex resolution are also im-
processes. The steady improvement in the precision of portantin mixing andCP violationmeasurements[5,6].
the measurements has not only helped to improve our Itiscrucialtohaveaccuratelifetime measurementsfrom
theoreticalunderstandingofstronginteractions,butalso fixed-target experiments to act as a standard to evalu-
to help stimulate the development of better theoretical ate any relative systematic differences. The Λ+ lifetime
c
tools. These have progressed from the spectator model presentedinthispaperrepresentsthemostaccuratemea-
to various quarks models and currently to Heavy Quark surementof this quantity to date andis a significantim-
Expansionmethods[1]. Thesecalculationaltoolsarethe provement over the present world average.
same orsimilar to those usedinother areas,for example The data used were collected by the FOCUS collab-
to determine the size of the V CKM element through oration in the 1997 fixed-target run at Fermi National
ub
inclusive semileptonic B decays [2]. More precise mea- AcceleratorLaboratory. The FOCUS spectrometer is an
surements of all of the charm particle lifetimes will help upgrade of the spectrometer used in the E687 photopro-
continue this process of improvement and extension of duction experiment [7]. The vertex region consists of
2
sures that only a small acceptance correctionto the life-
2 2500 timedistributionisneeded. Theaveragepropertimeres-
c
V/ 2000 Yield(L +)=8034±122 olutionforthisdecaysample(42fs)issmallenoughcom-
Me c pared to the lifetime to use a binned likelihood method
5 1500
s/ [9].
nt 1000 The t′ distributions for the decays in the signal and
e
Ev 500 sidebandregionsarebinnedintotwoseparatehistograms
0 from 0–1 ps in 20 fs bins. The observed number of de-
2.1 2.15 2.2 M(2p.2K5p ) (G2.3eV/c22).35 2.4 2.45 2.5 cays in the ith t′ bin is si for the signal region and bi for
′
the sideband region. The t distribution of the sideband
region is used as a measure of the lifetime distribution
FIG. 1: pKπ invariant mass plot for data (points) fitted
of background events in the signal region. Thus the ex-
withaGaussiansignalandquadraticbackground(solidline).
Theshadedareaindicatesthefittedlevelofbackground. The pected number of decays in the ith t′ bin of the signal
vertical dotted lines indicate the signal and sideband regions region is given by:
(see text) used in thelifetime analysis.
Expected =n =S f(t′i)e−t′i/τ +B bi . (1)
four BeO targets and 16 planes of silicon strip detec- Events i Pif(t′i)e−t′i/τ Pibi
tors (SSD). Two of the SSD planes were placed imme-
The likelihood that is maximized in the fit is given by
diately downstream of the second target, and two im-
mediately downstream of the fourth (most downstream) nsie−ni (αB)Nbe−αB
target. Momentumanalysiswasmadepossiblebytheuse Likelihood=Y i × (2)
s ! N !
i b
of 5 multiwire proportional chambers and two magnets i
with opposite polarities. Hadronic particle identification
whereS isthe totalnumberofsignaleventsandB isthe
was achieved using three multicell threshold Cˇerenkov
total number of background events in the signal region
counters [8]. The data for this measurement were taken
and S +B = Σs . The total number of events in the
using a photon beam with averageenergy of ∼180 GeV i
sideband region is N = Σ b and α is the ratio of the
b i i
for triggered events.
number of events in the sideband region to the number
The Λ+c → pK−π+[13] candidates are reconstructed ofbackgroundeventsinthesignalregion. Thevalueofα
using a candidate driven algorithm which is highly effi- isobtainedfromthefittotheinvariantmassdistribution
cientforalldecaysincludingshortlivedones. AllpK−π+ and is very close to 2. B and τ are the fit parameters.
candidates are tested to see if they form a vertex with a Theeffectsofgeometricalacceptance,detectorandre-
confidencelevelgreaterthan1%. Thecandidate Λ+c mo- constructionefficiencies,andabsorptionaregivenbythe
mentum vector is then projected to search for a produc- f(t′) correction function. The f(t′) is determined us-
tion vertex with one or more tracks. As many tracks as ing a detailed Monte Carlo (MC) simulation of the ex-
possible are included in the production vertex so long as periment where the production (using Pythia [10]) was
the vertex confidence level is larger than 1%. The pro- tuned so that the production distributions for data and
duction vertex is required to be within one of the four MCmatched. Notethatonlytheshapeofthef(t′)func-
targets. The separation L between the production and tion is important and it is obtained by dividing the ob-
decayverticesisrequiredtobelargerthan6σL whereσL servedMCt′ distributionbyapureexponentialwiththe
is the erroron L calculated on a candidate-by-candidate MC generated lifetime. The f(t′) distribution is shown
basis. In addition, each track in the pK−π+ candidate in Fig. 2(a).
combination must also satisfy the appropriate Cˇerenkov Using the likelihood function givenabove we obtained
particle identification criteria. afittedlifetimeof204.6±3.4fs. Thelifetimedistribution
The pKπ invariant mass plot for data is shown in ofalldecaysinthe signalregionisshowninFig.2(b)to-
Fig. 1. The fit shown uses a Gaussian signal and a getherwiththe fitandthelevelofbackgroundcontained
quadratic background function which yields 8034±122 in the signal region.
reconstructedΛc decays. The lifetime analysisuses pKπ Detailed studies were performed to determine the sys-
candidates within the signaland symmetric sidebandre- tematic uncertainty in this measurement.
gions as shown in the figure. All three regions are 4σm The uncertainty in the absolute time scale was inves-
wide and the centers of the sideband regions are lo- tigated by studying the absolute length and momentum
cated±6σm fromthemeanofthefittedGaussian,where scales in the experiment. For the length scale, compar-
σm =8.2 MeV/c2 is the width of the fitted Gaussian. isons were made between measurements of the distances
For the lifetime analysis we use the reduced proper between silicon planes in the target region. The values
time, t′ =(L−6σL)/βγc [14], where βγ =pΛc/mΛc and obtainedusingvertexpositionsinthedatawiththestan-
require it to be less than 1 ps to reduce long-lived back- dard vertexing code agree well with those obtained us-
grounds. Thisrequirementwasalreadymadeforthedata ing precision instruments. The absolute momentum and
shown in Fig. 1. The use of the reduced proper time en- massscaleswerecheckedbycomparingthereconstructed
3
s 1.3 enedtheCˇerenkovrequirementsonthedataandusedthe
20 f 1.2 (a) MC efficiencies to extrapolate to tighter particle identi-
n/ fication criteria. From this we found the above three
o 1.1
cti decays respectively contribute 0.5%, 1.3% and 2.7% of
e 1
rr the total background in the signal region. The small
o 0.9
C contributionofthese reflectionbackgroundsandthe fact
f(t) 0.8 that they are distributed fairly uniformly across the sig-
0.7 nal and sideband mass regions mean they give rise to
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Reduced Proper Time (ps) insignificant uncertainties. This was verified in a test by
explicitly eliminating them by cutting out the appropri-
(b) ate mass regions. Using variations in particle identifica-
103
tion and vertexing selection to significantly change the
s
0 f signal/background ratio also showed no significant un-
2
nts/ 102 certainties.
e Thebackgroundlifetimeuncertaintywasfurtherinves-
v
E tigatedbyusing symmetric sidebands ofdifferentwidths
(4–16σ ), and located at different separations from the
10 m
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 signalregion(±4to±16σ ). Theeffectofusingonlythe
m
Reduced Proper Time (ps)
loworonlythehighmasssidebandwasalsostudied. The
s effectofhavingthefitparameterBtrulyfreebyeliminat-
20 f (c) ing the background term in the likelihood (second term
e)/ 104
al in Eq. (2)) was studied and found to be inconsequential.
c
ary s 103 NinottheethbaatcktghreourensdulttesrmofitnhethpeKliπkemlihaossodfi.t are only used
bit
ar 102 Finally, anindependent analysiswhich didnot relyon
nts ( knowledge of the background lifetime distribution was
ve 10 performed. In this analysis the data were split into
E 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Reduced Proper Time (ps) twenty 50 fs wide reduced proper time bins from 0–1 ps.
The number of Λ+ → pK−π+ decays in each bin was
c
FIG. 2: (a) The f(t′) correction function. Deviation from determinedinamassfitandtheyieldsfittedtoanexpo-
′
a flat line indicates the correction from a pure exponential; nential decay distribution modified by a f(t) correction
′
(b) The lifetime distribution for all decays in the data sig- function. This f(t) function was obtained separately
nal region (points), and the fit (histogram). The shaded dis- for this analysis from the MC, doing the same split into
tribution shows the lifetime distribution of the background twenty time bins and fitting the mass distributions for
component in the signal region; (c) The lifetime distribution each MC bin. This f(t′) correction function agrees well
forΛc decays(points),i.e. thesidebandsubtractedandf(t′) with that obtained in the standard analysis method.
corrected yield. The line is a pure exponential with the fit-
Fromthese studies we assigna backgroundsystematic
ted lifetime and the shaded region gives the background. An
uncertainty of ±0.77%.
arbitary yield scale is used because of the particular normal-
ization of f(t′). Uncertaintiesinthef(t′)correctionincludeuncertain-
ties from the geometrical acceptance, the detector and
reconstructionefficiencies,theproductionmodel,theab-
masses of charm and strange mesons and hyperons with sorption cross-sections,and the decay dynamics.
established values. Our studies showed no evidence of With our chosenselectioncriteria,the f(t′) correction
any scale offset, but due to the limited statistical pre- reducesthefittedlifetimeby1.19%. Anumberofstudies
cision of these comparisons we assign an uncertainty of were performed to study the uncertainty in this correc-
±0.11% to the absolute time scale. tion. Since the correction function is obtained from MC
The backgrounds are composed of a non-charm and simulations,carewastakentoensurethatthissimulation
a charm component; these two background components correctly reproduces a very large number of data distri-
are approximately equal in our sample and fairly evenly butions. In particular the MC reproduces the data Λ+
c
distributed across the signal and sideband mass regions. longitudinalandtransversemomenta,the multiplicity of
The level and lifetime distribution of the background in the production vertex, and the decay length and proper
the signal mass region is assumed to be well represented timeresolutions. Asensitivecheckoftheacceptanceand
by symmetric mass sidebands close to the signal region. efficiency part ofthe MC correctionwasdone using high
Theuncertaintiesthatarisebecauseoftheseassumptions statistics K0 → π+π− decays. Short-lived K0 decays
S S
were determined by a large number of studies. were reconstructed using the same analysis methods in
The contamination from D+ → K−π+π+, D+ → the same decay region as the Λ+ decays. Since the K0
c S
K−K+π+ and D+ →K−K+π+ decays misidentified as lifetime is well knownwe candetermine the f(t′) correc-
s
pK−π+ decaysweredeterminedin oursample. We loos- tion in data and compare it to that obtained in our MC
4
simulation. The agreement is excellent but was limited We have measured the Λ+ lifetime to be 204.6 ±
c
bybothdataandMCstatisticstoasensitivityof±2%of 3.4 (stat.)±2.5 (syst.) fs using 8034±122 Λ → pKπ
c
the correction. Using this as the level of the uncertainty decays from the Fermilab FOCUS photoproduction ex-
′
inthef(t)correction,wecanassignasystematicuncer- periment. This measurementrepresentsa significantim-
tainty due to this correction of ±0.83%. Possible time provement in accuracy and special care was taken to in-
dependent systematic effects werelookedfor by splitting vestigate and properly quantify possible systematic un-
the data into different time periods and comparing the certainties. Table II compares our measurement with
fitted lifetimes. We also compared the separate fitted previousrecentpublishedresults. Thedifferencebetween
lifetimes for decaysoriginatingfromeachofthe fourtar- this measurementand the measurement from the CLEO
gets. No systematic uncertainties were found in these e+e− experimentmaypointtotheemergenceofpossible
two comparisons. relative systematic effects [12]. Any such systematic dif-
Our limited knowledge of the production and decay
of the Λ+ could contribute to a systematic uncertainty. TABLE I:Contributions to the systematic uncertainty.
c
This was studied using different MC simulations where Contribution Systematic (%)
the production parameters and the resonance substruc- Time scale ±0.11
ture of the decay were varied over reasonable ranges. Backgrounds ±0.77
Productionsystematicswerealsostudiedbysplittingthe Acceptance ±0.83
dataintodifferentbinsoflongitudinalandtransverseΛ+ Production ±0.38
c Resolutions ±0.12
momenta,primaryvertexmultiplicity,andbycomparing
Absorption ±0.23
the fitted lifetimes for particles and anti-particles. We
assign a systematic uncertainty of ±0.38% due to our Total ±1.23
limited knowledge of Λ+ production and decay.
c
In order to use the reduced proper time we must be
able to correctly model our proper time resolution. This TABLE II: Comparison of recent Λ+c lifetime measurements.
was verifiedby comparing the distributions for data and
MC and by studying splits of the data sample that can Experiment Type τ(Λ+c) fs
E687 [9] FT 215±16±8
besensitivetoresolutioneffects. Thedataweresplitinto
SELEX [11] FT 198.1±7.0±5.6
bins of proper time resolution and reconstructed invari- CLEO II.5 [4] e+e− 179.6±6.9±4.4
ant mass. Variations of the proper time bin width from FOCUS (this result) FT 204.6±3.4±2.5
10 to 100 fs were also studied as was changing the fitted
range from 0–0.6ps to 0–1.4 ps, and from 0–1 ps to 0.2–
1 ps. We assign a systematic uncertainty of ±0.12% to
the lifetime due to resolution uncertainties. ference would be important to resolve given the number
The systematic uncertainty due to absorption of the of recent and future mixing and CP-violation measure-
Λ+ and daughter particles was studied by varying the ments that rely on accurate knowledge of lifetime distri-
c
charm interaction cross-sectionby 100% and the daugh- butions.
ter particle interactioncross-sectionsby 50% in the MC. We wish to acknowledge the assistance of the staffs
It was also studied by comparing the lifetimes of decays of Fermi National Accelerator Laboratory, the INFN of
occuringinsideandoutsideofthetarget,andbycompar- Italy, and the physics departments of the collaborating
ingthelifetimesfordecayswheretheΛ+wasproducedin institutions. This researchwas supported in part by the
c
the upstream half of each target with those produced in U.S.NationalScienceFoundation,theU.S.Department
the downstream half of the same target. We determined of Energy,the Italian Istituto Nazionale di Fisica Nucle-
a systematic uncertainty of ±0.23% due to absorption. are and Ministero dell’Universit`a e della Ricerca Scien-
Contributions to the systematic uncertainty are sum- tifica e Tecnologica, the Brazilian Conselho Nacional de
marized in Table I. Taking contributions to be uncorre- Desenvolvimento Cient´ıfico e Tecnol´ogico, CONACyT-
latedweobtainatotalsystematicuncertaintyof±1.23% M´exico, the Korean Ministry of Education, and the Ko-
or ±2.5 fs. rean Science and Engineering Foundation.
[1] G. Bellini, I. I. Y. Bigi, and P. J. Dornan, Phys. Rept. Sect. A 320, 519 (1992).
289, 1 (1997). [8] J. M. Link et al., FERMILAB-Pub-01/243-E, hep-
[2] M. B. Voloshin, Phys. Lett. B515, 74 (2001). ex/0108011, to be published in Nucl. Instrum. Methods
[3] G. Bonvicini et al., Phys.Rev.Lett. 82, 4586 (1999). Phys. Res., Sect. A.
[4] A.H.Mahmoodetal.,Phys.Rev.Lett.86,2232(2001). [9] P. L. Frabetti et al., Phys.Rev.Lett. 70, 1755 (1993).
[5] R.Godang et al., Phys. Rev.Lett. 84, 5038 (2000). [10] T. Sj¨ostrand et al., Comput. Phys. Commun. 135, 238
[6] K.Abeet al., hep-ex/0111026. (2001).
[7] P.L.Frabettietal., Nucl.Instrum.MethodsPhys.Res., [11] A.Kushnirenkoetal., Phys.Rev.Lett. 86,5243 (2001).
5
[12] H. W. K. Cheung, FERMILAB-Conf-01/351-E, hep- minimum allowed time for each decay candidate. In the
ex/0111050. absence of other corrections, it does not matter where
[13] The charge conjugate mode is implicitly implied unless along the decay exponential one starts the clock, hence
otherwise stated. thereduced proper timealso follows a pureexponential.
[14] The reduced proper time has the clock started at the