- Email: [email protected]
Pion pair production in photon-photon collisions at DCI
Volume 194, number 4
PHYSICS LETTERS B
P I O N PAIR PRODUCTION I N P H O T O N - P H O T O N
20 August 1987
C O L L I S I O N S AT DCI
Z. A J A L T O U N I , A. F A L V A R D , J. JOUSSET, B. M I C H E L J.C. M O N T R E T , Ph. M U T T E R , D. P A L L I N Laboratoire de Physique Corpusculaire, Universit~ de Clermont II, BP 45, F-63170 Aubibre, France
A. C O U R A U , J.E. A U G U S T I N , A. C O R D I E R , G. COSME, F. C O U C H O T , B. D E L C O U R T , B. D U D E L Z A K , F. F U L D A , B. G R E L A U D , G. G R O S D I D I E R , J. H A I S S I N S K I , B. J E A N - M A R I E , S. J U L L I A N , D. L A L A N N E , V. L E P E L T I E R , F. M A N E , C. PAULOT, R. R I S K A L L A , Ph. ROY, F. R U M P F , G. S Z K L A R Z Laboratoire de l'Acc~l~rateur Lin~aire, Universitd de Paris-Sud, F-91405 Orsay, France
R. B A L D I N I , S. C A L C A T E R R A , G. C A P O N Laboratori Nazionali di Frascati delI'INFN, CP 13, 1-00044 FrascatL Italy
D. BISELLO, G. BUSETTO, L. P E S C A R A , P. S A R T O R I a n d L. S T A N C O Dipartimento di Fisica deH'Universit~ di Padova, e INFN, Sezione di Padova, 1-35131 Padua, Italy
Received 20 March 1987
This paper presents experimental results on n ÷n - production near threshold from the collision of two quasi-real photons. The data, obtained at the e+e collider DCI, are a combination of the results from the DMI and DM2 experiments. Using the e+e and n ÷n - production for normalization and cross-checks, we observe a pion pair yield at low invariant mass ( W< 500 MeV/c 2) which is approximately twice the one expected from Born terms.
U p to now, only the DM1 [1] a n d P L U T O [2] Collaborations have r e p o r t e d results on the n ÷ ~ p r o d u c t i o n near threshold in 77 collisions. Both experiments show a high yield for low i n v a r i a n t masses: In the case o f the DM1 experiment, the measured cross section for a mass W ~ < 500 MeV/c 2 is about twice the one expected from the Born terms alone, while P L U T O also shows a similar effect for W ~ < 460 MeV/c 2. This p a p e r reports on a statistical i m p r o v e m e n t o f the D M 1 result by adding the very similar one [ 3 ] also o b t a i n e d at D C I with the D M 2 detector [4]. Both experiments were looking at the e÷e - - - , e ÷ e - X ÷ X - reactions involving collisions o f two quasi-real photons. In both cases, a coincidence is required between a p a i r o f charged particles prod u c e d in the central detector ( D M 1 or D M 2 ) a n d at
0370-2693/87/$ 03.50 © Elsevier Science Publishers B.V. ( N o r t h - H o l l a n d Physics Publishing D i v i s i o n )
least one o f the two scattered electrons observed in a zero-degree tagging device [ 5 ]. In all cases the Q2 o f both p h o t o n s remains strictly limited, either by the tagging [ Q Z < 2 0 0 (MeV/c2) 2] or, for the notagged ones, by a kinematical cut [ 0 2 < 1 0 4 (MeV/c2)2]. The similarity o f the experimental arrangements and analyses allows a s u m m a t i o n o f the D M 1 a n d D M 2 samples in o r d e r to o b t a i n a single result with a higher statistical significance. Table 1 c o m p a r e s the acceptances a n d resolutions o f the central detectors a n d o f the tagging devices used in the two experiments. The m a i n difference between t h e m comes from the s t a n d a r d triggering c o n d i t i o n s involved in the central detector: (i) In the D M I case, the trigger only requires one charged track hitting three cylindrical multiwire prop o r t i o n a l chambers. It leads to a transverse m o m e n t u m cut o f 75 MeV/c. 573
Volume 194, number 4
PHYSICS LETTERS B
20 August 1987
Table 1 Characteristics of the DM I and DM2 apparatus.
Central detector polar acceptance azimuthal acceptance angular resolution (mrd) transverse m o m e n t u m acceptance (MeV/c) m o m e n t u m resolution tT(p)/p Tagging system Vertical acceptance (mrd) horizontal acceptance (mrd) X = E~,/Ebeam angular resolutions (mrd) energy resolution a(y)y= 1-x
DM 1
DM2
Icos0 I ~<0.64 0~<0~<2n
0~<~<2n
a(O) = 18
a(0)=5
a(O)=18
tT(O) = 2 see text
p~/> 75
0.05
~f(0.025p~) 2 +0.0152 (Pt in GeV/c)
10vl ~< 12.5
10v1~<12.5
10,1~<10.0
IOH[ <~10.0
0.20~
0.20~
(ii) In the DM2 case, two triggers were used: the single-tag trigger and the double-tag one. Both require that a charged particle track hits at least eight drift chambers and one time-of-flight scintillator located inside the magnetic coil of the central detector. When only one scattered electron is detected in the tagging device, the single-tag trigger also imposes a hit in a scintillation counter of the photon detector located behind the magnetic coil of the DM2. This induces a Pt cut which depends on the nature of the charged particle and its energy loss in the material located before the photon detector (respectively 75, 170 and 190 MeV/c for electrons, muons and pions). If there is a coincidence between the two tagging hodoscopes, no condition is imposed on the photon detector output. This trigger favours the acquisition of double-tagged events at low invariant masses, the cut on Pt of any charged particles being then reduced to 75 MeV/c. The event analysis requires two (and only two) tracks of opposite charge reconstructed in the central detector with an acolinearity angle larger than 4 deg. and at least one scattered electron reconstructed on the tagging device. When only one scattered electron is observed, the kinematical parameters of the other 574
Icos01 ~<0.71
one are determined from momentum conservation with some uncertainties due to resolution and radiative effects. When both electrons are measured, radiative effects are corrected for by assuming that the missing momentum is carried away by one radiated photon. In both cases, the momenta of the four prongs being defined, we can determine the invariant mass of the pair (Wvv) and the squared mass of its components (M2). There are two main sources of background which contaminate the single-tagged events. One arises from random coincidences between an annihilation event within the central detector and one electron which has undergone some Bremsstrahlung and which is detected in the tagging device. The other one is due to beam-gas electroproduction in which the scattered electron is detected by the tagging device. Both sources of background can be practically eliminated [ 1 ] by a kinematical cut on the momenta P~ and P2 of the charged particles observed in the central detector: IP, I + I P z I + IP, +P2 [ ~<0.77x/s, and another cut on the squared mass Q2 of the reconstructed untagged photon, assuming the event
Volume 194, number 4
PHYSICS LETTERS B
to be due to a 77 collision. The small remaining contamination ( < 4 % ) has been estimated from the events leading to a M 2 reconstructed mass outside the expected range [ M 2 < - 1 0 0 0 0 (MeV/c2) 2, M2> 30000 (MeV/c2) 2] and from Monte Carlo simulations. These few background events have been subtracted in all the following kinematical spectra (respectively 20 in DM1 and 16 in DM2). Figs. 1a and 1b show the experimental M 2 spectra thus obtained in the D M I and DM2 experiments respectively.
100
DM1
(~)
75
was fitted to histograms la and lb respectively.
f~(M2),f,(M 2) and f~ ( M 2) are the normalized M 2
25
~ ;'5 ~ 50
~ 2s ®
© 200
0 -10
The Q2 value of both photons being strictly limited (Q2 << w2), the Monte Carlo simulations were performed using the complete differential cross sections obtained by factorizing two quasi-real transverse photon spectra with the relevant real photon-photon cross section (DEPA) [ 6 ]. The latter was derived from QED for lepton pair production and from Born terms for pion pair production. These simulations include all experimental effects and uncertainties [ 1 ] (acceptances, resolutions, radiative corrections [ 7 ], etc.). Let us notice that they differ essentially by the photon-photon cross sections, so that a normalization by the lepton pair production should suppress systematics coming from acceptances and effective luminosity measurements. Then the following expression
Norc(M ~) + Nj~( M 2) + N~f~ ( M 2)
5C
z
20 August 1987
~
0
10 20 M~ [,103(MeV/c2)2]
30
Fig. 1. Square mass distributions for the two opposite charged particles. Full line in (a) and (b): best fit to the data where the numbers o f e + e -, ~t+~t- and n + n - pairs are free parameters. Dashed line in (a) and (b): lepton tail deduced from the best fit up to M 2 = 16000 (MeV/c2) 2 with lepton alone. Full line in (c): global fit to the lepton kinematic spectra combined with the distribution of n +Tz- from both experiments. Dashed line in (c): lepton tail decuded from the global fit to the lepton kinematic spectra.
distributions expected from the simulation of the e+e-,~t+lx-,n +n - channels, and the free parameters Ne,N~, and N~ are the corresponding numbers of pairs to be fitted. The results of these adjustments as well as the integrated luminosity deduced from the number of electron pairs are given in table 2, for both the DM 1 and DM2 samples. By selecting the events in the relevant M 2 ranges [ - 10000 < M e < 6000 (MeV/c 2) 2, 6000 < M 2 < 16000 (MeV/c 2) 2, 16000 < M 2 < 30000 (MeV/c 2) 2, respectively for electrons, muons and pions ], one gets the experimental invariant mass distributions of the various pairs Wee,W~, and W ~ . These distributions are compared to those expected from the given luminosities in figs. 2a, 2b, 3a, 3b and 4a, 4b where the electron tails are included in the ~t~t spectra but the lepton ones are subtraced in the rcn spectra. Let us emphasize that the invariant mass distributions of both leptonic pairs are in good agreement with the expected ones though the normalization has been derived from the number of electron pairs only. In contrast, one observes an excess of pions with respect to the expectation derived from the Born terms alone. derived from the Born terms alone. To draw such a conclusion one has to ensure that radiative and resolution effects are well taken into account in the simulations. In particular, the M 2 leptonic tail should be well determined in the pion mass 575
20 August 1987
PHYSICS LETTERS B
Volume 194, n u m b e r 4 Table 2 Summary of the results for the various channels. N¢~
N~,.
N,,,,
L~ ( nb - ' )
Separated fits a DMI DM2
363+20 343+ 19
197+ 14 135+ 12
29+8 24+7
452+25 347+ 19
DM1 + D M 2
706+28
332+ 18
53+ ll
Global fit b) DMI +DM2
704+27
340+ 13
51+ 11
359+ 19 344+20
177+9 163+9
DMI DM2
448 + 23 349+20
a~ The number of events is derived from the fit to the M 2 spectrum and the luminosity from the number of electron pairs. b~ The number of leptons and the luminosity are derived from a m a x i m u m likelihood fit to the W and M 2 spectrum of the combined DM 1 and DM2 e+e - and ~t+la- pairs.
50
®
DM1 t~l
40
®
:30
20
40
%
2O
10
% o60 u~
~2G ca
~4o "6
g g~ "5
20
i
L
I~11÷DIVI2
©
®
~12C
90 30
60
30
100
200
300
400
500
600
700
800
Wyy(MeWc') Fig. 2. Electron pair invariant mass distributions. Data: invariant mass distribution of events with - 1 0 0 0 0 < M e < 6 0 0 0 (MeV/c2) 2. Full line in (a) and (b): expected distribution normalized to the n u m b e r of events. Full line in (c): global fit to the lepton kinematic spectra. 576
0 100
200
300
400
500
600
Wyv(MeV/c~)
700
800
900
Fig. 3. Muon pair invariant mass distributions. Data: invariant mass distribution of events with 6000 < M 2 < 16000(MeV/c2) 2. Full line in (a) and (b): expected distribution using the luminosity deduced from the number of electron pairs. Full line in (c): global fit to the lepton kinematical spectra. The contamination due to the M 2 tail of electrons is included in these spectra.
Volume 194, number 4
PHYSICS LETTERS B
20 August 1987
®
16 14
DM1 I0C
12 10 8
>
6
0
4
g 75
2
DM2
~
8
o
4
~
@ ~0 Z
---:
f
i~
--7 ',
--i
2~
2 i 'r
Z DM1 + DM2
30 25 20
0
15 10
10
20
Mx= (MeV/c=)2 x 103
;i 300
400
500
600
700
800
900
Fig. 5. M 2 distributions o f e + e - identified pairs (full line) and of all the pairs (dashed line) observed in the DM2 experiment. The curve is the M 2 distribution expected from the e÷e simulation.
W.rrn ( M e V / c 2)
Fig. 4. Pion pair invariant mass distributions. Data: invariant mass distribution of events with 16000 < M 2 < 30000 (MeV/c2) 2 after lepton substraction. Full line in (a) and (b): expectation from the Born terms using the luminosity deduced from the number of electron pairs. Full line in (c): expectation from the Born terms using the luminosity deduced from the global fit to the lepton kinematic spectra.
range. These points are checked by the good agreement of the fitted M 2 distributions. They can be further checked by studying the sample of electron pairs well identified in the DM2 by the showers produced in the photon detector. Fig. 5 shows the M 2 distributions for all pairs and for the sample of identified electron pairs. It also compares the latter spectrum to the one obtained from the Monte Carlo simulation. Good agreement is observed up to large values of M 2 with only a small contamination due to a few pion pairs which have undergone nuclear interaction or charge exchange and simulate electromagnetic showers.
Figs. 1c, 2c, 3c and 4c show the M 2 and invariant mass distributions obtained by just adding the corresponding DM1 and DM2 spectra. Table 2 shows in its upper part the total numbers ofe+e-,~t+kt- and n + n - events also obtained by just adding the corresponding values which came out of the independent DM1 and DM2 analyses. Let us recall that in these separate analyses the three sample sizes were deduced from fits to the M 2 distributions (and only to these distributions), the normalizations were derived from the number of electron pairs only. The yields and the kinematic distributions of both leptonic channels were well reproduced by the Monte Carlo simulation. Then, a global analysis of the c o m b i n e d data has been performed. As far as the lepton pairs are concerned, the only free parameters are the integrated luminosities of the DM1 and DM2 experiments. A maximum likelihood fit to the Wee, W,, and M 2 distributions for M 2 < 16000 (MeV/c 2) z leads to a new determination of these integrated luminosities and 577
Volume 194, number 4
PHYSICS LETTERS B
of the corresponding yields and kinematical distributions for each kind of lepton pairs. The total number of pions is deduced from the excess of observed events in the range 16000 < M 2 < 30000 (MeV/c 2) 2 over the leptonic tail in the same range. The results of this global analysis are given in the lower part of table 2. One can see that they are very close to those previously obtained by mere addition of the two samples and lead to a good agreement of the predicted M 2 and Ww distributions with those observed for the lepton pairs (see figs. lc, 2c and 3c). Again one observes a significant excess of pions with respect to what is predicted by the Monte Carlo simulation using for each experiment the fitted value of the integrated luminosity and the ~7--'n + ~ - cross section derived from the Born terms alone (fig. 4c). From the different analysis, the total number of pions obtained, Not, . . . . .
d(~+~
-
20 August 1987
proposed to explain such a result including, in particular, the possible production of scalar isoscalar mesons with an important gluon component [8-10]. Recently, a model based on an attractive potential between pions has also been proposed by Barnes et al. [11] in order to reproduce this threshold enhancement. We are very indebted to the DCI staff for the good operation of the machine during the course of DM 1 and DM2 experiments. We are very grateful to all people who have been involved in building and operating the detectors, and especially Dr. M. Brossard for his technical assistance in running the zero degree tagging system. We would like to thank Dr. G. Mennessier and Professor T. Barnes for their great interest in our result and their illuminating discussions.
)=52+_ 11 ,
References is nearly twice the one expected, NHorn(~ + ~ - ) = 2 9 + _ 2 . In summary, we have studied the production of charged particle pairs of low invariant masses in quasi-real photon collisions. Lepton and pion pairs have been kinematically identified. We observe a good agreement for the e+e - and n + n - yields and mass distributions with respect to the QED expectations. These events being used for normalization, we find that the Born terms are not sufficient to explain the somewhat large production of pion pairs at low invariant masses. Several phenomenological models based on nn phaseshift analyses have been
578
[ 1 ] A. Courau et al., Phys. Lett. B 96 (1980) 402; Nucl. Phys. B 271 (1986) 1. [2] Ch. Berger et al., Z. Phys. C 26 (1984) 199. [3] M. Brossard et al., Nucl. Instrum. Methods 174 (1980) 157. [4] J.E. Augustin et al., Phys. Scr. 23 (1981) 623. [ 5 ] Z. Ajaltouni, Thrse de Doctorat d'Etat-Universit6 de Clermont II (1986). [6] P. Kessler, Acta Phys. Austr. 41 (1975) 141; A. Courau, SLAC-PUB 3363 (1984). [7] Etim et al,, Nuovo Cimento B 51 (1967) 276. [8] G. Mennessier, Z. Phys. C 16 (1983) 241. [9] A.Z. Kaloshin and V.V. Serebryakov, Z. Phys. C 32 (1986) 279. [ 10] K.L. Au et al., Phys. Rev. D 35 (1987) 1633. [ 11 ] T. Barnes et al., Phys. Lett. B 183 (1987) 210.
PHYSICS LETTERS B
P I O N PAIR PRODUCTION I N P H O T O N - P H O T O N
20 August 1987
C O L L I S I O N S AT DCI
Z. A J A L T O U N I , A. F A L V A R D , J. JOUSSET, B. M I C H E L J.C. M O N T R E T , Ph. M U T T E R , D. P A L L I N Laboratoire de Physique Corpusculaire, Universit~ de Clermont II, BP 45, F-63170 Aubibre, France
A. C O U R A U , J.E. A U G U S T I N , A. C O R D I E R , G. COSME, F. C O U C H O T , B. D E L C O U R T , B. D U D E L Z A K , F. F U L D A , B. G R E L A U D , G. G R O S D I D I E R , J. H A I S S I N S K I , B. J E A N - M A R I E , S. J U L L I A N , D. L A L A N N E , V. L E P E L T I E R , F. M A N E , C. PAULOT, R. R I S K A L L A , Ph. ROY, F. R U M P F , G. S Z K L A R Z Laboratoire de l'Acc~l~rateur Lin~aire, Universitd de Paris-Sud, F-91405 Orsay, France
R. B A L D I N I , S. C A L C A T E R R A , G. C A P O N Laboratori Nazionali di Frascati delI'INFN, CP 13, 1-00044 FrascatL Italy
D. BISELLO, G. BUSETTO, L. P E S C A R A , P. S A R T O R I a n d L. S T A N C O Dipartimento di Fisica deH'Universit~ di Padova, e INFN, Sezione di Padova, 1-35131 Padua, Italy
Received 20 March 1987
This paper presents experimental results on n ÷n - production near threshold from the collision of two quasi-real photons. The data, obtained at the e+e collider DCI, are a combination of the results from the DMI and DM2 experiments. Using the e+e and n ÷n - production for normalization and cross-checks, we observe a pion pair yield at low invariant mass ( W< 500 MeV/c 2) which is approximately twice the one expected from Born terms.
U p to now, only the DM1 [1] a n d P L U T O [2] Collaborations have r e p o r t e d results on the n ÷ ~ p r o d u c t i o n near threshold in 77 collisions. Both experiments show a high yield for low i n v a r i a n t masses: In the case o f the DM1 experiment, the measured cross section for a mass W ~ < 500 MeV/c 2 is about twice the one expected from the Born terms alone, while P L U T O also shows a similar effect for W ~ < 460 MeV/c 2. This p a p e r reports on a statistical i m p r o v e m e n t o f the D M 1 result by adding the very similar one [ 3 ] also o b t a i n e d at D C I with the D M 2 detector [4]. Both experiments were looking at the e÷e - - - , e ÷ e - X ÷ X - reactions involving collisions o f two quasi-real photons. In both cases, a coincidence is required between a p a i r o f charged particles prod u c e d in the central detector ( D M 1 or D M 2 ) a n d at
0370-2693/87/$ 03.50 © Elsevier Science Publishers B.V. ( N o r t h - H o l l a n d Physics Publishing D i v i s i o n )
least one o f the two scattered electrons observed in a zero-degree tagging device [ 5 ]. In all cases the Q2 o f both p h o t o n s remains strictly limited, either by the tagging [ Q Z < 2 0 0 (MeV/c2) 2] or, for the notagged ones, by a kinematical cut [ 0 2 < 1 0 4 (MeV/c2)2]. The similarity o f the experimental arrangements and analyses allows a s u m m a t i o n o f the D M 1 a n d D M 2 samples in o r d e r to o b t a i n a single result with a higher statistical significance. Table 1 c o m p a r e s the acceptances a n d resolutions o f the central detectors a n d o f the tagging devices used in the two experiments. The m a i n difference between t h e m comes from the s t a n d a r d triggering c o n d i t i o n s involved in the central detector: (i) In the D M I case, the trigger only requires one charged track hitting three cylindrical multiwire prop o r t i o n a l chambers. It leads to a transverse m o m e n t u m cut o f 75 MeV/c. 573
Volume 194, number 4
PHYSICS LETTERS B
20 August 1987
Table 1 Characteristics of the DM I and DM2 apparatus.
Central detector polar acceptance azimuthal acceptance angular resolution (mrd) transverse m o m e n t u m acceptance (MeV/c) m o m e n t u m resolution tT(p)/p Tagging system Vertical acceptance (mrd) horizontal acceptance (mrd) X = E~,/Ebeam angular resolutions (mrd) energy resolution a(y)y= 1-x
DM 1
DM2
Icos0 I ~<0.64 0~<0~<2n
0~<~<2n
a(O) = 18
a(0)=5
a(O)=18
tT(O) = 2 see text
p~/> 75
0.05
~f(0.025p~) 2 +0.0152 (Pt in GeV/c)
10vl ~< 12.5
10v1~<12.5
10,1~<10.0
IOH[ <~10.0
0.20~
0.20~
(ii) In the DM2 case, two triggers were used: the single-tag trigger and the double-tag one. Both require that a charged particle track hits at least eight drift chambers and one time-of-flight scintillator located inside the magnetic coil of the central detector. When only one scattered electron is detected in the tagging device, the single-tag trigger also imposes a hit in a scintillation counter of the photon detector located behind the magnetic coil of the DM2. This induces a Pt cut which depends on the nature of the charged particle and its energy loss in the material located before the photon detector (respectively 75, 170 and 190 MeV/c for electrons, muons and pions). If there is a coincidence between the two tagging hodoscopes, no condition is imposed on the photon detector output. This trigger favours the acquisition of double-tagged events at low invariant masses, the cut on Pt of any charged particles being then reduced to 75 MeV/c. The event analysis requires two (and only two) tracks of opposite charge reconstructed in the central detector with an acolinearity angle larger than 4 deg. and at least one scattered electron reconstructed on the tagging device. When only one scattered electron is observed, the kinematical parameters of the other 574
Icos01 ~<0.71
one are determined from momentum conservation with some uncertainties due to resolution and radiative effects. When both electrons are measured, radiative effects are corrected for by assuming that the missing momentum is carried away by one radiated photon. In both cases, the momenta of the four prongs being defined, we can determine the invariant mass of the pair (Wvv) and the squared mass of its components (M2). There are two main sources of background which contaminate the single-tagged events. One arises from random coincidences between an annihilation event within the central detector and one electron which has undergone some Bremsstrahlung and which is detected in the tagging device. The other one is due to beam-gas electroproduction in which the scattered electron is detected by the tagging device. Both sources of background can be practically eliminated [ 1 ] by a kinematical cut on the momenta P~ and P2 of the charged particles observed in the central detector: IP, I + I P z I + IP, +P2 [ ~<0.77x/s, and another cut on the squared mass Q2 of the reconstructed untagged photon, assuming the event
Volume 194, number 4
PHYSICS LETTERS B
to be due to a 77 collision. The small remaining contamination ( < 4 % ) has been estimated from the events leading to a M 2 reconstructed mass outside the expected range [ M 2 < - 1 0 0 0 0 (MeV/c2) 2, M2> 30000 (MeV/c2) 2] and from Monte Carlo simulations. These few background events have been subtracted in all the following kinematical spectra (respectively 20 in DM1 and 16 in DM2). Figs. 1a and 1b show the experimental M 2 spectra thus obtained in the D M I and DM2 experiments respectively.
100
DM1
(~)
75
was fitted to histograms la and lb respectively.
f~(M2),f,(M 2) and f~ ( M 2) are the normalized M 2
25
~ ;'5 ~ 50
~ 2s ®
© 200
0 -10
The Q2 value of both photons being strictly limited (Q2 << w2), the Monte Carlo simulations were performed using the complete differential cross sections obtained by factorizing two quasi-real transverse photon spectra with the relevant real photon-photon cross section (DEPA) [ 6 ]. The latter was derived from QED for lepton pair production and from Born terms for pion pair production. These simulations include all experimental effects and uncertainties [ 1 ] (acceptances, resolutions, radiative corrections [ 7 ], etc.). Let us notice that they differ essentially by the photon-photon cross sections, so that a normalization by the lepton pair production should suppress systematics coming from acceptances and effective luminosity measurements. Then the following expression
Norc(M ~) + Nj~( M 2) + N~f~ ( M 2)
5C
z
20 August 1987
~
0
10 20 M~ [,103(MeV/c2)2]
30
Fig. 1. Square mass distributions for the two opposite charged particles. Full line in (a) and (b): best fit to the data where the numbers o f e + e -, ~t+~t- and n + n - pairs are free parameters. Dashed line in (a) and (b): lepton tail deduced from the best fit up to M 2 = 16000 (MeV/c2) 2 with lepton alone. Full line in (c): global fit to the lepton kinematic spectra combined with the distribution of n +Tz- from both experiments. Dashed line in (c): lepton tail decuded from the global fit to the lepton kinematic spectra.
distributions expected from the simulation of the e+e-,~t+lx-,n +n - channels, and the free parameters Ne,N~, and N~ are the corresponding numbers of pairs to be fitted. The results of these adjustments as well as the integrated luminosity deduced from the number of electron pairs are given in table 2, for both the DM 1 and DM2 samples. By selecting the events in the relevant M 2 ranges [ - 10000 < M e < 6000 (MeV/c 2) 2, 6000 < M 2 < 16000 (MeV/c 2) 2, 16000 < M 2 < 30000 (MeV/c 2) 2, respectively for electrons, muons and pions ], one gets the experimental invariant mass distributions of the various pairs Wee,W~, and W ~ . These distributions are compared to those expected from the given luminosities in figs. 2a, 2b, 3a, 3b and 4a, 4b where the electron tails are included in the ~t~t spectra but the lepton ones are subtraced in the rcn spectra. Let us emphasize that the invariant mass distributions of both leptonic pairs are in good agreement with the expected ones though the normalization has been derived from the number of electron pairs only. In contrast, one observes an excess of pions with respect to the expectation derived from the Born terms alone. derived from the Born terms alone. To draw such a conclusion one has to ensure that radiative and resolution effects are well taken into account in the simulations. In particular, the M 2 leptonic tail should be well determined in the pion mass 575
20 August 1987
PHYSICS LETTERS B
Volume 194, n u m b e r 4 Table 2 Summary of the results for the various channels. N¢~
N~,.
N,,,,
L~ ( nb - ' )
Separated fits a DMI DM2
363+20 343+ 19
197+ 14 135+ 12
29+8 24+7
452+25 347+ 19
DM1 + D M 2
706+28
332+ 18
53+ ll
Global fit b) DMI +DM2
704+27
340+ 13
51+ 11
359+ 19 344+20
177+9 163+9
DMI DM2
448 + 23 349+20
a~ The number of events is derived from the fit to the M 2 spectrum and the luminosity from the number of electron pairs. b~ The number of leptons and the luminosity are derived from a m a x i m u m likelihood fit to the W and M 2 spectrum of the combined DM 1 and DM2 e+e - and ~t+la- pairs.
50
®
DM1 t~l
40
®
:30
20
40
%
2O
10
% o60 u~
~2G ca
~4o "6
g g~ "5
20
i
L
I~11÷DIVI2
©
®
~12C
90 30
60
30
100
200
300
400
500
600
700
800
Wyy(MeWc') Fig. 2. Electron pair invariant mass distributions. Data: invariant mass distribution of events with - 1 0 0 0 0 < M e < 6 0 0 0 (MeV/c2) 2. Full line in (a) and (b): expected distribution normalized to the n u m b e r of events. Full line in (c): global fit to the lepton kinematic spectra. 576
0 100
200
300
400
500
600
Wyv(MeV/c~)
700
800
900
Fig. 3. Muon pair invariant mass distributions. Data: invariant mass distribution of events with 6000 < M 2 < 16000(MeV/c2) 2. Full line in (a) and (b): expected distribution using the luminosity deduced from the number of electron pairs. Full line in (c): global fit to the lepton kinematical spectra. The contamination due to the M 2 tail of electrons is included in these spectra.
Volume 194, number 4
PHYSICS LETTERS B
20 August 1987
®
16 14
DM1 I0C
12 10 8
>
6
0
4
g 75
2
DM2
~
8
o
4
~
@ ~0 Z
---:
f
i~
--7 ',
--i
2~
2 i 'r
Z DM1 + DM2
30 25 20
0
15 10
10
20
Mx= (MeV/c=)2 x 103
;i 300
400
500
600
700
800
900
Fig. 5. M 2 distributions o f e + e - identified pairs (full line) and of all the pairs (dashed line) observed in the DM2 experiment. The curve is the M 2 distribution expected from the e÷e simulation.
W.rrn ( M e V / c 2)
Fig. 4. Pion pair invariant mass distributions. Data: invariant mass distribution of events with 16000 < M 2 < 30000 (MeV/c2) 2 after lepton substraction. Full line in (a) and (b): expectation from the Born terms using the luminosity deduced from the number of electron pairs. Full line in (c): expectation from the Born terms using the luminosity deduced from the global fit to the lepton kinematic spectra.
range. These points are checked by the good agreement of the fitted M 2 distributions. They can be further checked by studying the sample of electron pairs well identified in the DM2 by the showers produced in the photon detector. Fig. 5 shows the M 2 distributions for all pairs and for the sample of identified electron pairs. It also compares the latter spectrum to the one obtained from the Monte Carlo simulation. Good agreement is observed up to large values of M 2 with only a small contamination due to a few pion pairs which have undergone nuclear interaction or charge exchange and simulate electromagnetic showers.
Figs. 1c, 2c, 3c and 4c show the M 2 and invariant mass distributions obtained by just adding the corresponding DM1 and DM2 spectra. Table 2 shows in its upper part the total numbers ofe+e-,~t+kt- and n + n - events also obtained by just adding the corresponding values which came out of the independent DM1 and DM2 analyses. Let us recall that in these separate analyses the three sample sizes were deduced from fits to the M 2 distributions (and only to these distributions), the normalizations were derived from the number of electron pairs only. The yields and the kinematic distributions of both leptonic channels were well reproduced by the Monte Carlo simulation. Then, a global analysis of the c o m b i n e d data has been performed. As far as the lepton pairs are concerned, the only free parameters are the integrated luminosities of the DM1 and DM2 experiments. A maximum likelihood fit to the Wee, W,, and M 2 distributions for M 2 < 16000 (MeV/c 2) z leads to a new determination of these integrated luminosities and 577
Volume 194, number 4
PHYSICS LETTERS B
of the corresponding yields and kinematical distributions for each kind of lepton pairs. The total number of pions is deduced from the excess of observed events in the range 16000 < M 2 < 30000 (MeV/c 2) 2 over the leptonic tail in the same range. The results of this global analysis are given in the lower part of table 2. One can see that they are very close to those previously obtained by mere addition of the two samples and lead to a good agreement of the predicted M 2 and Ww distributions with those observed for the lepton pairs (see figs. lc, 2c and 3c). Again one observes a significant excess of pions with respect to what is predicted by the Monte Carlo simulation using for each experiment the fitted value of the integrated luminosity and the ~7--'n + ~ - cross section derived from the Born terms alone (fig. 4c). From the different analysis, the total number of pions obtained, Not, . . . . .
d(~+~
-
20 August 1987
proposed to explain such a result including, in particular, the possible production of scalar isoscalar mesons with an important gluon component [8-10]. Recently, a model based on an attractive potential between pions has also been proposed by Barnes et al. [11] in order to reproduce this threshold enhancement. We are very indebted to the DCI staff for the good operation of the machine during the course of DM 1 and DM2 experiments. We are very grateful to all people who have been involved in building and operating the detectors, and especially Dr. M. Brossard for his technical assistance in running the zero degree tagging system. We would like to thank Dr. G. Mennessier and Professor T. Barnes for their great interest in our result and their illuminating discussions.
)=52+_ 11 ,
References is nearly twice the one expected, NHorn(~ + ~ - ) = 2 9 + _ 2 . In summary, we have studied the production of charged particle pairs of low invariant masses in quasi-real photon collisions. Lepton and pion pairs have been kinematically identified. We observe a good agreement for the e+e - and n + n - yields and mass distributions with respect to the QED expectations. These events being used for normalization, we find that the Born terms are not sufficient to explain the somewhat large production of pion pairs at low invariant masses. Several phenomenological models based on nn phaseshift analyses have been
578
[ 1 ] A. Courau et al., Phys. Lett. B 96 (1980) 402; Nucl. Phys. B 271 (1986) 1. [2] Ch. Berger et al., Z. Phys. C 26 (1984) 199. [3] M. Brossard et al., Nucl. Instrum. Methods 174 (1980) 157. [4] J.E. Augustin et al., Phys. Scr. 23 (1981) 623. [ 5 ] Z. Ajaltouni, Thrse de Doctorat d'Etat-Universit6 de Clermont II (1986). [6] P. Kessler, Acta Phys. Austr. 41 (1975) 141; A. Courau, SLAC-PUB 3363 (1984). [7] Etim et al,, Nuovo Cimento B 51 (1967) 276. [8] G. Mennessier, Z. Phys. C 16 (1983) 241. [9] A.Z. Kaloshin and V.V. Serebryakov, Z. Phys. C 32 (1986) 279. [ 10] K.L. Au et al., Phys. Rev. D 35 (1987) 1633. [ 11 ] T. Barnes et al., Phys. Lett. B 183 (1987) 210.