RESULTS FROM THE PLUTO EXPERIMENT

ON E+E- REACTIONS AT HIGH ENERGIES

Ch. Berger

1. Physikalisches Institut der RWTH

Sornmerfeldstr.

5100 Aachen, Germany

Summary

Results on e+e- interactions at CM ener­gies of the e+e- system> 10 GeV are dis­cussed. A QED test down to very small distan­ces (I" -= 3·1 0-16 cm ) wa s performed. The tota Icross section for e+8- annihilation into ha­drons was measured. The event topologies werestudied in detail. From these data we can ex­clude the contribution of a standard topquark. This result is supported by our mea­surement of inclusive muon production. Evi­dence for hard gluon bremsstrahlung in e+e­annihilation into hadrons is presented. Theevidence comes from jet broadening and theproduction of events with 3 well separatedha dro n i c jets •The so called two photon reactions e+e-~e+e-+X

where X is a hadronic state or a lepton pai~

were measured for the first time at high in­variant masses WX' The results for lepton pairproduction agree very well with QED predic­tions The dependence of the hadro nic era ss sec­tion·on 02 (the mass of one of the virtualphotons) and WXis discussed.

Introduction

In this talk I will report on data takenwith the detector PLUTO at the storage ringPETRA (Hamburg, Germany) since November 1978.The CM energy of the e+e--system ranges from13 GeV up to 31.6 GeV 14,15,16.

PLUTO at PETRA is run by a collaboration ofphysicists from Germany, Norway and the US.The present list of authors is given below.

PLUTO CollaborationChi Berger, H. Genzel, R. Grigull, W. Lackas,and F. RaupachI. Physikalisches Institut del" RWTH Aachen1 ,GermanyA. Klovning, E. Lillest6l, E. Lillethun andJ.A. SkardUniversity of Bergen 2 , NorwayH. Ackermann, G. Alexander3 , F. Barreiro,J. Burger, L. Criegee, H.C. Dehne~ R. Deve­nish 10 , A. Eskreys~, G. FIOgge, G. Franke,W. Gabriel, Chi Gerke, G. Knies, E. Lehmann,H.D. Mertiens, K.H. Pape, H.D. Reich,8. Stella 5, T.N. Ranga Swamy6, U. Timm,w. Wagner, P. Waloschek, G.G. Winter andW. ZimmermannDeutsches Elektronen-Synchrotron DESY, Ham­burg, Germany

O. Ach~erberg, v. B~obe113, ~. Boesten, 12H. Kapltza, B. KOPPltz, W. Luhrsen, R.MaschuwR. van Staa and H. SpitzerII. Institut fur Experimentalphysik derUniversitat Hamburg 1 , Germany

e.Y. Chang, R.G. Glasser, R.G. Kellogg,K.H. Lau~ B. Sechi-Zorn, A. Skuja, G. Welchand G.T. Zorn?University of Maryland B, Collegepark, USA

A. Backer, S. Brandt, K. Derikum, A. Diek­mann, C. Grupen, H.J. Meyer, B. Neumann,M. Rost and G. ZechGesamthochschule Siegen 1 , Germany

T. Azemoon 9, H.J. Daum, H. Meyer, M. Rossler,D. Schmidt and K. Wacker 11Gesamthochschule Wupperta1 1 , Germany

I will talk on the following topics

I Description of the detectorII QED resultsIII Search for a new flavour thresholdIV Investigation of gluon bremsstrahlungV First results on two photon interactions

I Description of the detector

The main components of PLUTO are(A) a central detector with 13 cylindricalproportional chambers operating in a magneticfield of 1.65 T. The momentum resolution forcharged tracks is op/p = 3% p (p in GeV)for p > 3 GeV.

(B) barrel and end cap shower counters withproportional tubes for position measurementof the showers. The energy resolution forelectrons and photons with energy E > 1 GeVis 0 IE - 35%/IE (E in GeV) in the barrel and- 19~/iE in the endcaps. The geometricalacceptance of (A) and (B) is 87% and 94% of4 11' sterad.(el a muon identifier with a 1 m iron absor­ber for hadrons. The tracks are sampled attwo depths within the absorber by a set ofproportional and drift chambe~s.

(D) Forward spectrometers on each side'of thedetector for luminosity measurements and forselection of reactions coming from two pho­tons interactions. Because these spectrometersare relatively new and some understanding oftheir operation is essential for the resultsof chapter VI I will describe them in moredetail.

The layout of the PLUTO detector is shown infig. 1.

do [ 2]S dcose nb GeV

• ECM=27.4 GeV• ECM = 13 GeV

104

II QED results

A high energy e+e- storage ring is anideal tool for testing the validity of QED inpurely electromagnetic processes at large mo­mentum transfers.At PETRA we have analyzed until now Bhabhascattering (e+e-+e+e-) at CM energies of 13,17 and 27.4 GeV.In the first step of the analysis Bhabha even~

were defined mainly by requiring two collinearand coplanar (with the beam axis) showers inthe barrel shower counter and the endcap. Werequired at least 1/3 of the beam energy de­posited in each shower, the acollinearity andacoplanarity angle being less than 200 • Forthe final angular cuts we used the track in­formation. In addition the track informationserved for the discrimination of very narrowhadronic jets from Bhabha scattering. Becausethese jets normally have a high charged multi­plicity/events with more than 4 tracks origi­nating from a common vertex were rejected. Wecompare our results to radiativelycorrectedQED predictions. The radiative correction~7

were computed according to Berends et al. 'including hadronic vacuum polarisation andheavy lepton contributions.I want to point out, that due to the strongmagnetic field and good momentum resolutionPLUTO can distinguish between forward andbackward Bhabha scattering.For a quantitativecomparision with QED we used the data betweencose = -.75 and cose = +.75 (e is the polarangle of the tracks, measured with respectto the beam axis).

PLUTO

o Earlier results

• Present results

- - - - - - -udscbt.OCO- ---

- SAT

tr!

o _ ~TRADORIS,

front vlews-

forward detector

Fig.1 Layout cf the PLUTO detector

(only one half shown)

_-£~~=~C=~;;;~~~~~iinteractionside view POInt

PLUTO

Each arm of the forward spectrometersconsists of a 'large angle tagger' (LATl anda 'small angle tagger' (SATl. The LAT coversthe polar angle region between 70 and 260 mradThe energy of electrons and photons is deter­mined with a lead scintillator shower counterof 14.5 radiation length thickness. The posi­tion of charged particles is determined bysix planes of proportional tube chamberswith a wire spacing of 1 cm. The SAT coversthe angular region between 23 and 70 mrad.Energy information of electrons and photonsis obtained from a lead glass shower countermatrix. It consists of 96 blocks (each witha front area of 6.6 x 6.6 cm2 ), in a concen­tric arrangement around the beam pipe. Thethickness of this counter is 12.5 radiationlength. Tracking of charged particles isachieved by a set of four planar proportionalwire chambers (wire distance 0.3 cm). In atest beam the energy resolution of the LATwas measured to be 11%I/E (rms) and of theSAT 8.5 liE (rms), E in GeV. These valueshave been reproduced by analyzing small angleBhabha scattering.

e"',; ,.c "-

,I b"-~

.~

b~ 2

10 20 Ecm IGeVI 30

Fig.3 The ratio ahadla~~ versus ECM ' Model

predictions are included.

~ ~ ~ ~ 0 2 A .6 B

cosSFig.2 Angular distribution for Bhabha

scattering at two energies

-20-

The energy scaled angular distributions forECM = 13 and 27.4 GeV are shown in fig.2. Theagreement with QED is excellent.For establishing quantitative limits on thevalidity of QED we used a form factor ansatzfor the timelike (FT) and spacelike (FS) ampl~

tude contributing to Bhabha scattering:

FT(s) = 1 ::.z:!, 1± s (for A2T»s)

1 ; s A +2AT2 T-

--~- ~1±2

- .9.--1 + A 2

s.±.

(Herein s=ECM2 =4 Ebe;m and q2= -s.sin2 e/2)

The approximation on the right hand sidegives a model independent parametrizationof a possible QED breakdown.Note# that other groups use different para­metrizations 18.

Our results are summarized in table 1 forAS = AT' We have also included the DORIS dataat 9.4 GeV on Bhabha scattering and one+e-~ ~+~-. A combined fit yields A+ = 71 GeVand A- = 67 GeV# thus provin~ QED valid downto distances of about 3.10- 1 cm.

III. Search for new flavour thresholds

In this chapter and in chapter IV I willconcentrate on our data on e+e- annihilationinto hadrons. The trigger conditions used areessentially the same as described in ref. 14 •To select hadronic e+e- annihilation eventsfrom our raw data we applied the followingcutsa) number of charged tracks in the central

detector> 2. We requ ired a distance <' 20 mmfrom the beam axis and <50 mm from the in­teraction point, when measured along thebeam (z - axis).

b) difference in azimuthal angle (~) for twoprongs ~~ <150 0

c) total observed energy (charged + neutrals)>0.5 ECM' In the analysis of our 13 and 17GeV data we actually used Eneutra1>O.3 ECM'

The observed events have a non negl~gib1e con­tribution from beam gas reactions (e1ectro­production). two photon annihilation intohadrons# and T+T- pair production.a) The small «5%) background from beam gas

interactions was estimated by measuring thedistribution of reconstructed event verti­ces along the z-a~is.

b) Hadron production from the so called twophoton reactions i.e. e+e-+e+e- +hadronsis discussed in some detail in chapter V.We used the results obtained there for,estimating the background to the annihi­lation channel.

c) The relative contribution from T+T- pairproduction based On the prong number andneutral energy distribution was also esti­mated. It turned out to be rather small

( < 1%) • I n tab I e 2 the number 0 fob s erv edhadronic events (corrected for beam gas back-ground) is given for various CM energiestogether with the expected background from twophoton reactions and T+T- pair production.

In column 2 of the table the correspondingintegrated luminosities are given. The lumi­nosity values have been determined fromBhabha events in the central detector. Theyagree with the results from the forward spec­trometers within 7%.

The acceptance E of the detector (column 4of table 1) has been calculated from a MonteCarlo simulation. I will discuss this pro­gram below. Knowing the corrections frominitial state radiation) the total hadroniccross section a had can now be calculatedwithout difficulty. As usual we compute thedimensionsless quantity R = ahad/alllJ. (column 7of table 1). The result is plotted in fig.3versus ECM' We have also included PLUTO dataat lower energies.

As indicated in the figure the R values above22 GeV agree very well with the expectedvalue of 3.9 obtained from udscb quarks andQCD corrections 19. On the other hand theyare clearly below the expectation includinga charge 2/3 top quark. I think that a topmass of about 10 GeV (e = 2/31) can be ex­cluded by the R measurement alone«R> = 3.88 ± 0.22).The topological details of hadronic eventsprovide an independent evidence for thepresence of any new flavour t~reshold. Thisis especially important for top masses around15 GeV, where the total cross section datasuffer from the limited statistics.We have analyzed the "jetiness" of oureve nt sin te rm s 0 f the two rna s t pop uI a r va ria b­les sphericity

S 3/2 min ZPf,i/Z pt and thrustCo l

T max! /PIIJ LIf.Llp dWe used both charged and neutral particlesfor the determination of Sand T. The resultfor the observed S versus ECM is shown infig. 4. The data agree well with the expec­tation from a Monte Carlo simulation includingudscb quarks (and gluon bremsstrahlung).The horizontal bars in the figure indicatethe Monte Carlo values including a 'standardtop quark' (t{ threshold 2 GeV below ECM).These expectations are definitely in con­flict with the data. (~5 std.dev.) Notehowever,that a charge of 1/3 would reduce thedifference in the expected S values by afactor 4! One arrives at the same conclusionby studying the observed thrust distribution(fig. 5). As usual we plot 1-T versus ECM'The figure also includes the Monte Carloexpectations with and without the top quark.

At this stage ,some explanations about theMonte Carlo program are in order. The programgenerates events which are subsequently pass~

through a complete detector simulation pro­gram and the same event recognition and ana­lysis chain as used for the data. Initial

-21-

.9 1.

dNECM =30 GeV dT

60

.6 .1

• PLUTO- Monte CarlQ--- "top"

.1

PLUTOdN

<S) observed • ECM=216 GeV dT

- Monte Carlocharged + neutral -----·top·· 60

.4 PLUTO•-Monte

Carlo

+40

.331 expectedwith top

"top" 16 obNrwd 20

.2

.7 .8 .9 1.0

TDiffer~ntial thrust distributionsat various e ergies. Solid linecorresponds. 0 the st'anrlard model.Dashed line ncludes top quarks.

30

ECM[GeV]10 20

Observed 1-T versus ECM •Model predictiora with andwithout top quark included.

.1

Fig.5

0 10 20 30

ECM[GeV]40

Fig.4 Observed sphericity v~rsus E~M. Modelpredictiorewith and wlthout op quarkincluded.

20

<1-T> observed --"'",,'charged + neutral .","",""

".4 " .,,'• PLUTO

/'"

-Monte .6 .7 .8 .9 1.0

Carlo dN

• PLUTO ECM=31.6 GeV dT

.3 - Monte Carlo 30--- "top"

+.2 "top" 20

-22-

1fABLg 1: Limits on the QED cutoff parameters (95% confidence level)

Data used E Model at 95% c.l. (GeV) xL/ IDF

ems(GeV)

9.4 A > 43 A > 40 19.4/23+ -ee only 13 A = AT A > 29 A > 17 10.4/14

s + -17 A > 55 A > 45 8.55/13+ -27.4 A > 40 A > 60 4.26/8+ -

loll! only 9.4 AT+> 16 AT-> 31 8.66/11

> 71 67 74/64j

ee and all As = AT A A >+ -abovel!l! energiss

combi.ned j

TABLE 2: Relative hadronic cross section R = (ahad/a )at specified e+e- c.m. energies. ~~Background subtraction, correction for initialstate radiation and T subtraction are incorporated.The data at 13 and 17 GeV have been analyzed usingslightly different methods 14 and are therefore notincluded in this table.

ECM 1= JL dt (b) Nhad expected bac kground radiative .R Ca )£ correct: io ns

(GeV) (nb- 1 ) observed N(2y) N' (T+ - ) ~T

22.0 47 ± 5 29 0.86 0.8 0.7 0.1 3.41 ± 0.73

27.6 408 ± 15 169 0.84 7.0 3.9 0.1 3.64 ± 0.31

30.0 561 ± 19 227 0.81 9.4 4.2 O. ~I 4.38 ± 0.37

31.6 219 ± 13 66 0.80 3.7 1 .5 o• ~I 3.59 ± 0.52

E2[(Nha~ - N2y )(a ) R cm - N' _}

87X. T+T ~ ) I

N + -£ detection efficiencYJ N' + - --!.........!-

T T £T

(b) Luminosity values as determined from the central shower counters. The errorsare statistical. There is an additional systematic error of 5%.

TABLE 3: Comparison of inclusive muon signal with expectations fromu,d,s,c.b and u.d.s.c.b.t quarks

ECM events with computed corrected expecteda muon back- signal udscb udscbt

(GeV) Cp>2 GeV) ground

27.6 7 2.9±O.8 4.1±2.7 4.7 14.630.0 8 4.1±1.2 3.9±3.1 5.1 15.931 .6 2 1 •7±O. 5 O.3±1.5 1 .7 5.3

total 17 8.7±2.6 8.3±4.9 11 • 5 35.8

-23-

state radiative effects are also included inthe program. The qq generating pro~ram isbased on the Field-Feynman model 2 • In itsstandard form the program contains u,d,s,cand b quarks. The top quark can be added also.For uds quarks we use t~e fragmentation func~

tion fen) = 1- A + 3 An with A = 0.77J forcb (t) quarks f (n) = constantJ n = 1 - z andZ = PH/Pq.where PH and Pq are the momentum ofthe primary meson and the quark respectivel~(ref. 20,21). The transverse momentum distrl-bution for the quarks is assumed to be exp(_q2 /2o q2 ) with Og = 247.5 MeV. Some timeswe changed 0q to higher values. This will beindicated in the text. The decay of knownprimary mesons with only udsc quarks aretaken from the particle data tables 22. Thedecays of the bottom and top mesons are fromref. 23,24.

The program 1s also able to simulate gluonbremsstrahlung 25 (q~g program). FollowingHoyer et al. 26 gluons are radiated only byu,d,s,c quarks, gluon jets are treated as acombination of quark-antiquark pairs, theQCD cross section is taken for qqg but acut-off on thrust is used. (This prescrip­tion yields a gluon jet in addition to theq and q jets in 15% (25%) of the generatedevents at 17 GeV (30 GeV) eM energy).

The need to include gluon bremsstrahlung fora quantitative description of our data willonly become clear (hopefully!) at the endof chapter IV. Nevertheless I use the qqgprogram already now, because leaving outthe gluon effects changes the conclusionsabout top quark contributions in no way, butonly describes the sphericity and thrustdistributions less accurately.

Returning now to the question of a new fla­vour threshold we plot in fig. 6 a,b,c thedifferential thrust distributions forECM = 27.6, 30 and 31.6 GeV. The data arein very good agreement with our standardmodel (qqg, udscb quarks). In contrast,having passed the threshold for open tt pro­duction,one would expect much more events atlow thrust (say T<0.8) and much less athigh thrust. ·For example at E = 27.6 wewould expect 37 events but observe only16 ± 4 events.

As a final check we study the inclusive muonsignal in the" hadron sample. Qualitativelyone would expect an increase in the numberof muons per event from the cascade decayof the top quark (fig.?). In order to re­duce the background from punch throughhadrons we demand a minimum momentum of2 GeV for the associated muons. The numberof events with an associated muon and theexpectations without and with a top quarkare given in table 3. These expectations arecalculated under the assumption of a cascadedecay for the heavy quarks (t ~ b ~ c ~ d)with a 10% branching ratio into muons ateach step. The data in the ~nergy range27.6 to 31.6 GeV agree well with the ex­pectations for udscb quarks and are signifi­cantly (- 5.6 std. dev .) below the valueswhen top quark states are included.

-24-

There are other interesting things, whichone can study in our hadron sample, butwhich are not so closely related to thetop search. As an example the averaged char­ged multiplicity n h (corrected for accep­tance losses) is ptotted versus ECM in fig.B.At high energies the increase is steeperthan expected from a simple extrapolation(27)of the low energy data «n> = 2 + 0.7 Ins )It seems to follow the trend of the pp data.

In conclusion of this chapter, the value ofR, the distribution in thrust and spheri­city, and the data on inclusive muons ex­clude open te production due to a charge 2/3top quark~with 'standard' decay and fragmen­tation at eM energies below 32 GeV. Thestatistical limitations of the present datado not allow any conclusions regarding acharge 1/3 quark heavier than the b-quark.

IV. Investigation of gluon

bremsstrahlung

Hadron production in e+e-collision via jetsis now widely believed to be 'objectiveevidence for quarks'. Let us recall thebasic features of a jet: The average trans­verse momentum of the particles in a jet(with respect to the jet axis) is limited«PoL> ~300 MeV), whereas the longitudinalmomentum <Pu> scales with the energy ECM.In fig. 9a the mean observed P.L and Pl\for the charged particles is plotted versusECM. The axis is taken from the thrust cal­culus including information from the neutralpart io les. (The same co nvent io.n ho ld s truefor fig. 10 and fig. 12). Although thedifference in the energy behaviour of <PL>and <p II > iss t r i king, 0 ne 0 bs erv esal soa slight increase of <P.L > with ECM ' Inorder to investigate this effect ln moredetail we study a highe2 moment of the p~

distribution, namely<p~ >, which is moresensitive to high p~ effects, but stillrather insensitive to experimental errors.

Jet broadening is predicted in any fieldtheory and thus is a necessary but notsufficient condition, for the validit~ ofOeD, which predicts an increase of <pt>like

<p1.2> - a ('s)·sS

due to gluon bremsstrahlung (e+e-~q~g).The PLUTO data ~re shown in fig.9b. Theincrease in <p~ > is nicely followed bythe Monte Carlo prediction including gluonbremsstrahlung (solid line), whereas theMonte Carlo without gluon bremsstrahlungpredicts a much weaker increase of theobserved <p~2> with the center of massenergy ECM. The model predictions are alsoincluded in fig.9a and obviously <Pu > and_<PL> are not very discriminative between qqand qqg. The coupling constant as (5)(as/w = 6.3% at ECM - 30 GeV 19) is rathersmall and thus th~ process e+e-~qqgg isstrongly supressed compared to e+e-~qqg.

One expects therefore oniy one of the jetstDbe broadened due'to gluon bremsstrahlung.

15

PLUTOt b c 5

• ADONE~

)( MARK 1 ? N~/.s:. 10 o PLUTOW W W u

• DASP /C ? ?/"./ppv

/'¢ /' .----

,/ ---5

¢ ".---?-'. _;-"",.. <n> =2+0.7 Ins

.~)(')()('K,..~~~)(~_ ......

q q' q q' q q'~ V ~ V I.l V 0 I I I I I I

1 2 3 5 10 15 20 30 40 SOe V e V e V

{S' [GeV]

Fig.? Cascade decay of the top quark

1.S -.

IGeV/c) ~<PII> PLUTO

, <Pi><p> ----qq

-qqg1.0

0.5

Fig.B Average charged multiplicity versusthe center of mass energy~ Data fromother experiments at low energiesare also included.

IGeV/cl 2 I (GeWe1 2

06 ____ qq Slim Jet Fat Jet 06

O.t -qqg

~O~

<Pt>

~<pl>

02 . -- 0.2

0 I(a)

0~O 20 20 40

rGeV) -E ern Eern -IGeVl

Fig.10 Average observed<p~> of charged par­ticles in the slim and fat jet asfunction of CM energy.

0.1

• • • __e_~-'-=----=----- - -- ---

0.5

1.01.0

o.s

----qq 127-32 GeV-qqg

1.5 -.-._'-qq (aq=OJ5GeVJ'cJ

<PL2>

(GeVIc) 2Slim Jet , Fat Jet f _] fGeVlC) 2

113-17~V ..-t-

0.6 i 0.&<p 2>

,/ <p2>1. 0.4

~/ 0.4 1-

v0.2 v 0.2

(b)

o --1-__----'I-- (c_)-010 0 S 00 05 1.0

-xp Xp-

(b)

30IGeV)20

10

0.2

(a)o...............-.- ---L- ----L_-.-J

Eern 30 (GeV)(GeV/c)2

O.~

< P12;>

OJ

Fig.9a,b <Po >, <p~> and<pI> versus ECM in­cluding model predictions with(qqg) and without (qq) gluon brems­strahlung

2Fig.12 Seagull plots - <p~> of charged par-ticles as function of xp for slimand fat jets at low and high energies.The solid and dashed 1in~s are qqgand qq predictions, respectively.

-25-

An even more sensitive measure is obta~ned

by investigating the dependence of <PL >on the scaled hadron momentum xp = Ph/Pbeam.A strong dependence of <p~2> on xpis againpredicted from hard gluon bremmstrahlung.This follows intuitively from fig. 11.

The result is shown in fig. 10. The slim jetpart is equally well described by the MonteCarlo with an without gluons. ~n the otherhand the strong increase of<p~ > between thelowest and highest energy for the fat jetcannot at all be explained by the Monte Carlosimulation without gluon radiation.

12

..

116

3

PLUTO e+ e-~q q 9

J..5.

Fig. 13a

9

-1IIPol (hlII

_~.__ .__ .__ .__ Ijet axis

Experimentally one can search for this'one sided jet broadening! by sorting eventby event the jet with the higher p~ into the'fat jet' class and the jet with the lower p~

into the 'slim jet' class. Although this pro~

cedure introduces a natural bias, the ener-gy de pen den ce 0 f <PJ. 2> i n bo t h cIa sse s s ha uI dbe very different if compared to qq and qqgmodEl predictions.

hadron (fromgluonfragmenta ­tion)

5

.2".~U·

Fig.11

Because of limited statistics we combinedfor this research the low energy (13 and 17GeV) and the high energy (27.6, 30, 31.6 GeV)runs. No gluon effects are visible in the socalled 'seagull plot' fig. 12 for the fatand slim jet at low energies. At high ener­gies the steep increase of <p~2> for the fatjet can easily be explained by the qqg model,but not by qq. We investigated the influenceof the transverse quark momentum on our re­sults,Trying to fit the inclusive PL distri­bution with the qq model one needs 00=350 MeVat ECM = 30 GeV. It is obvious from ~he

dotted line in fig.12, that even the qq modelwith the artificially increased 0q cannot fitour data.

Hard gluon bremsstrahlung should finally leadto triple jets. A nice example of a planart rip 1 e qfi~.. jet(ECl'r 30 G8 V) iss hown i n fig.13a which reminds one to the very similar

Fig. 13b

-26-

situation in radiative Bhabha scattering(eey),~shown in fig,13bJ

For a quantitative study of the event shape weuse the triplicity method 28. The final statehadrons with the momenta Pl ...PNCfig.14aJ aregrouped into 3 non empty c asses C1,C2 ,C3 withthe total momenta ~ (Cl)=Ipi 1=1,2,3,if: C1Triplicity T3 is then defined by

NT3=C11 Ip.l) max IP<C

1)1+IPCC

2)1+IPCC

3)1 (1)

i=1 1 C1,C2

,C3with the bounds T3=1 for a perfect 3 jetevent and~T3=.65 for a spherical event. Thoseclasses Cl of particles yielding the maximumT3 are identified with the hadrons origina­ting from quark and gluon fragmentation. Thusthe ~et momenta are the ~ (Ci). We rename themP1' P2' P3 with P1>P2>P:3 ('fastest jet',

second fastest jetJ,etc.J The convention weuse for the angles between the jets is indi­cated in fig. 14b.

8

0° 180 0 00

Dalitz plot for 81 .82 .93

00 < 81 < 120 0

90 0 < 82 < 1800

1200 < 8 3 < 180 0

Fig,15 Prinziple of Oalitz plot analysis forjet axis arientation

...P,

(a) ct (b)~

63

C;\ ...P2

61

Fig.16 Event density in angular Dalitz diagran

Fig.14 a,b

Triple jet events are characterized by lowthrust and high triplicity. Selecting allevents with T3>O.9 and T<.8 leaves us with48 events (for the high energy data), whichhas to be compared to predicted 43 eventswith gluon bremsstrahlung and 11 eventswithout.Perturbative QCD makes quantitative predic­tions on the cross section

aa + - -de de (8 8 + qqg)

1 2We investigated the jet axis orientation ofour high energy data (27.6~ECM~31.6 GeV)in an angular Oalitz plot. The principle isexplained in fig.15. Due to the orderingprocedure in fastest, second fastest andthird fastest jet only the triangle ABC isfilled with entries. Triple jet events arecentered near C, whereas along the line AScollinear events are concentrated. A cut in83 separates triple jets from qq jets. Wehave studied the number of entries for twodifferent e3 (fig. 16) cuts and compared itto model expectations with and without gluonradiation. The results are shown in tabla 4.Again the ~greement between experiment andthe prediction based on gluon radiation is

1200

1000

800 , . , . ..' .:

...

-27-

striking, although for the rather weak cute 3 <t620 the q~ model with oq=350 MeV liesc~ose to the data.

-1

i!

TAIL

(e) 13-17GeV

TAil

O~~--..I_-05 10 15

(d) 27 -32 GeV

10~

O' 6 0.8

b)

27 - 32 6eV

-qqg---qq

n04

0.2 0.1 0.2 0.3<p2 > GeV/c)2

out

a)

13 -17 GeV

0.2

,-\ 0 2 _, , --qqg, \ ----qqI ,

\

10

20

20

10

C1J.0

E ]0~

::z::

Fig. 18 shows the distributions of <pL >and <P:iQ> for the two energy regionsO~~d for~ompar1s10n the predictions of qqg and qq. It1S clear, that at high energies the <p2. >distribution develops a tail. This tail 1ncorresponds to planar events which are pre­dict:d by qqg but cannot be accounted forby qq.

Fig~18a,b,c~d Distributions of <p2 > and<p2 in > for the lower and hi~~~r ener­gy regions. Solid and dashed linesare q~g a~d q~ predictions respective­ly

\,,,\

'f-'I

'I"'-,,

'I'I

III

eventsexpecteda =350MeV

eventsexpecteda =245MeV

".',, ,

", ~, I, I

, I

'I _--­-------'I

I/1,I

II

II

to

o

1.0o

-1.0

-to

IGeVlc)

Table 4:eventsobserved

qq qqg qq

e3<1500 52 19 51 31

03<1620 120 74 130 101

Triple jets define a plane. Therefore2

we ex­pect the momentum out of t2 8 plane <p t> tobehave differently from <P~i >, the outransverse momentum in the p~ane with respectto a given axis e.g. the axis of the fastestjet. Fig. 17 shows a typical 3 jet event, de­monstrating that Pout is rather small, butP~,in is large. CIt is by the way not the same

1.0 0-1.0

-to 0 4.0Momentum Component (6eV/c)

Fig.17 Momentum.vectors of an event CEcm =31.6GeV) with high triplicity and low thrust pro­jected onto the triplicity plane Ctop left),onto a perpendicular plane normal to thefastest jet (top right) and onto a plane con­taining the direction of the fastest jet(bottom). Solid and dotted lines correspondto charged and neutral particles, respective­ly. The directions of the jet axis are indi­cated as fat bars near the margins of thefigures.e~ent ~s in ~ig. 131) For a quantlt~tive stu­dy.of planarity we used the meth~d alreadyemployed in our analysis of the T decay 29.We construct the conv~ntional sphericitytensor30 N

rae (pf 6ae_p~ p~) (2)i=l

where the p. are the momentum vectors of all(charged ana neutral) hadrons and a,a are thecoordinate indices. We now order the eigen­values Ak of Taa so that A1> A2>A 3 and callthe corresponding eigenvectors ~1' n2' n3'The sphericity axis is n3. If the events aredisklike the normal to the disk plane is n1.The vector n2 lies in the disk plane and isnormal to the sphericity axis.Wen 0 w2form the a vera g es <p2' >= <(p · ;1 ) 2>and <PJ.. > = <cp·rr2 )2> overo~!l charged par­tic les 18f a n event as a· measure of t he momentumout of the plane and in the plane in a direc­tion perpendicular to the sphericity axis.

-28-

V First results an two photon

reactions

It was suggested some years ago3~ thatin high energy e+e- reactions hadron ~ro­

duction via the so called two photon mecha­nism (fig. 19) becomes more and more impor­tant compared to the usual 1 photon mechanisnThe importance of experiments covering two­photon processes lies in the fact that onecan hope to extract from the measured crosssection

the genuine two-photon cross section foreither real or virtual photons. Depending ondifferent kinematical conditions one can ex­plore the hadron like or the point like be­haviour of photon-photon scattering in thesame reaction35 • The specific signature ofreaction (3) as compared to electron-p~'.. i­tron annihilation into hadrons is the occur­ence of two leptons in the final state. whichare peaked at high energies and very smallforward angles36 • In order to select the yyreactions, the PLUTO detector at PETRA hasbeen equipped with two forward spectrome­ters for identifying ('tagging') the out­going electrons and positrons (see chapterfor a description).

two photon reactions. This behaviour isnaively expected from the bremsstrahlungs~

spectrum of the interacting photons. The'2y interpretation' becomes evident, if onelooks at the energy spectrum of all eventswith a 'tag' (Eta >3 GeV) in one of the for­ward spectrometer~. This distribution is gi­ven by the shaded area in fig. 20.

In order to test our quantitative understan­ding of two photon initiated events i~ thePLUTO detector, we have studied the reactione+e-+e+e- 2 prongs. The obvious 2 prong can­didates are e+e-,p+p-,n+n-.Because the n+n- contributions is expectedto be small compared to the lepton channel,one can calculate the 2 prong cr~ss sec­tion via high order (amplitude e!) QED.Fig.22 shows the distribution of the 2 pronginvariant mass. The thin curve is the QEDprediction using a program by Vermaseren 37.The agreement is very good. I think the num­ber of events in the plot and the invariantmasses reached (W2 = 16 GeV2!) already indi­cate 'a big step forward' in t~e field oftwo photon physics. The analysis of the ~dronic reactions was restricted to eventswith a tag in the SAT only, in order to keepthe 02 of the virtual photon small.

Fig.19The vertex distribution of the tagged events(fig. 21) shows a very clear peak around theinteraction point, thus excluding the possi­bility, that the tagged events are comingfrom electroproduction, with the electronscattered into the forward spectrometers.Theshaded area in fig. 21 contains our candi­dates for hadronic events. These are definedby requiring three or more tracks in the de­tector or two tracks and at least oneshower not associated with the tracks. Thelow multiplicity hadronic events have beenscanned by hand, to make sure that they arenot contaminated by OED background.

(3 )+ - +-e e + e e + hadrons

We have also studied the distributions of<p2out > and <p2o n> computed with respect tothe normal on t~e triplicity plane and withrespect to a unit vector in the triplicityplane perpendicular to the fastest jet axis,respectively. The results are very similar tothose shown in fig.18. However, all distribu­tions are somewhat broader, since diagonali­zation of (2) minimizes p2L with respect tort'3 whereas (1) maximizes P" with respect tothe 3-jet axis.

In the h~gh en~rgy data sa~p~e we observe 68events wlth <P~in> >.5 GeV ln good agree­ment with the qqg prediction of 56. In con- _trast the qq mod el (23 events) and even the qqmodel with oQ=3S0 MeV (37 eve~ts)ocan~ot ex­plain the tall of the p~,in dlstrlbutlon.

To summarize this chapter, I emphasize thatthe evidence for gluons which has been accu­mu la ted duri ng the .past 2 years 31,32,33,espec ia lly by the 'work of the PLUTO group onthe T resonance 29, gets very strong supportfrom the p~esent experiment. On the otherhand we still do not have 'objectiveevidence for gluons'. To arrive at thisgoal one has (just giving one example)to be able to tell, which of the 3 jetsin fig.13 a iR the gluon. In the case ofOED (fig.13bJ it is immediately obvious,which of the 3 particles is the photon.With more data to come, we will hopefullysolve this problem.

In fig.20 the distribution (beam gas-back­ground subtracted!) of the total visibl~

energy is shown for our data at 13.8 GeVbeam energy. The cut used for separating~'1y events' is also indicated. The steepincrease towards small energies of the lowenergy part of the distribution stronglysupport the idea, that these events come from

For the question of extracting a hadroniccross section from the measured data I pointout, that only the photon radiated from theuntagged electron is close to the mass shell.The tagged electron,. however, (e' >20 mrad)radiates ph~tons which h~ve 02»m2 «02>up to ,5 GeV l. The proper descrip~ion of thisexperimental situation is electron scatter-

-29 -

Events

100

4

W2prang[Gev)

PLUTO

6 7 GeV5

3

<0 2) =O.~ GeV 2

<Eb)-15 GeV

Ebeam=13.8 GeV

3 4Wvis

e+ e--e+e- e+ e­e+e- 11+11-

+++-----~-_ _l

2

PLUTO

+----

t------~., ++}-......, =::::r:=-.-__1~__,

--_ T

PLUTO<Q2> =0.1 GeV2

Distribution of the invariant mass of2 prong events in the central detectorwith a tag in the forward spectrometers.The data are given by the fat histogra~

the QED expectations by the thin curve.

o

10

250

500

750

20

30 Events

500

Total hadronic cross section for two 2photon initiated events at <Q2>~.1GeV(Ebeam = 6.5 and 8.5 GeV). Modelexpectations included.

at +e:aL [nb)1000~--'::"":'_---- -..,..,

[rb

1000

1500

Fig.22

Fig.23

r~RMultiprong

DAII

ECM =27.6 GeVEbeam=13.8 GeV

VERTEX DISTRIBUTIONSINGLE TAG

150

Fig.2D Distribution of the measured energyper event (charged + neutral) inunits of ECM= 2 Ebeam

50

100

150

Fig.21 Distribution of event vertices forevents with a tag in the forwardspectrometers

Fig.25 Total hadronic cross section for twophoton initiated events at <02>~.4GeV2(13.7<Ebeam<15.6 GeV)

50 o 50

Z[mm]

-30-

2 3 4 5 6 7 8

Wvis [GeV)

ing off a free photon target. The crosssection for ey scattering can be written verysimilar to inelastic electron-nucleon scatter­ing. Because we want to interprete our datain terms of photon-photon cross sections ra­ther than in tems of structure functions weadopt, what is known as 'Hands formula" inelectroproduction38 •

q

dodO' dE'

ey

r t iE aflux factor for the virtual photons,£ the polarization parameter and at and 01are respectively the tota.! crossA ..~ections forhadron production via virtual transverseand longitudinal photons off a free photontarget. The differential cross section fore+e-~ e+e- + hadrons is then given by

do ee~ee hadrons =r t (Ot +£0 1 ) N(Ey ) dEydO' dE' (4)

where NCEy ) dEy is the number of photons perelectron radiated from the untagged lepton.Assuming 01 to be zero (which is very likelyfor small q2) (4) reduces to the one termformula discussed in ref.39 but with a diffe­rent flux factor. The validity of formula(4) has be~B discussed by ~arimalo, K~ssler

and Parisi .The cross section versus W. at<Q2>=.1GeV2(Eb=6.5,e.5GeV) is shown in ¥i~.23.Wvis is the invariant mass of the hadronicsystem as determined in the central detector,assuming pion masses for all charged partic­les. The range of W that contributes (FWHM)is indicated by the horizontal bars. Besidesthe statistical error, which is given in thefigure, we estimate an overall systematicerror of ±25% mainly coming from the un­certainty in the acceptance calculation.The solid line is the expectation forOt(q2,W2) assuming a pure Regge asymptoticbehaviour for yy scattering extrapolated tolow energies via duality and factorization 41

2

(.24nb + .27nb/W) ( ~ 2) (5)l+Q 1m

p

having included a p form factor ansatz forthe virtual photon. In the highest W binsthe data lie close to the model, takinginto account the rather large statisticaland systematic errors. This a posterorijustyfies our 2y background calculationdiscussed in chapter III#because the Monte­Carlo simulation used for the calculationof the 2y contribution is based on thisRegge model. In the lower bins there is anexcess in the measured cross section. It isunlikely that all this excess is due tolongitudinal contributions. In a recentpaper42 Greco and Shrivastava have arguedthat both for real and virtual photons onehas to include contributions from thepoint-like coupling of real photons toquar~ (quark-loop diagrams, fig.24). Follo­wing this suggestion and assuming an effec­tive quark mass of 100 MeV we calculate acontribution which is given by the dottedline in fig.25. Obviously this improvesthe agreement with the data, but does notac co unt fo r t re 0 bs erved cro s ssect ion qua nt i ­tatively.

-31-

q

Fig.24

The hadronic cross section measured at beamenergies from 13.8 to 15.6 GeV (fig.26) lieconsistently below the data of fig.25. Thereason is very simple. At high beam energiesthe average Q2 of the electrons scatteredin to the SAT is about .4 GeV2 and thus thecross section is reduced due to form factoreffects. The difference in the measured crosssections is (within the errors given) welldescribed by the simple p pole ansatz of eq.( 5) •To summarize this chapter, we have measuredfor the first time two photon initiatedhadron production at center of mass energies>1 GeV. At the highest eM energies the re­sulting cross section is close to the ex­paction from Regge like exchange processes,leaving room for pointlike contributions atlow energies. We have also extended themeasurement of two photon QED reactions torather high invariant masses of the leptonpairs. Two photon physics will clearly bea very exciting field of research at e+e­machines.

I want to thank the organizers of thesymposium for the very nice atmosphere at theconference. I have also to thank many of myPLUTO colleagues for very helpful discussionson their work.

Conclusions

Because I have given a summary at theend of each chapter I only emphasize the mostimportant results again.

QED is valid down to very small distan­c es (~3. 10 - 1 6cm )

It is very unlikely that the continuumfor tt production is below 30 GeV center ofmass energy.

There is evidence for gluon bremsstrah­lung. Jet broadening and triple jet produc­tion rate agree with QeD predictions.

The cross section for hadron productionvia 2y interactions agrees with Regge assymp­totic behaviour at high CM energies. At lowenergies there is room for pointlike contri­butions.

1 •2.

3.4.

5.

6.

7.

8.

9.10.11 •

12.13.14.

15.

16.

17 •

18 •

19.

20.

21 .

22.

References and footnotes

Supported by the BMFT, GermanyPartially supported by the NorwegianResearch Council for Science andHumanitiesOn leave from University Tel Aviv,IsraelOn leave from Institute of Nuclear PhysicsKrakow, PolandOn leave from University of Rome, ItalYJpartially supported by INFNOn leave from the Tata Institute, Bombay,IndiaUniversity of Maryland General ResearchBoard Grantee for 1978Partially supported by Department ofEn§rgy, USANow at University College, London,EnglandNow at University of Oxford, EnglandNow at Harvard University, Cambridge,Mass., USANow at Technische Universitat, KarlsruheNow at CERN, Geneva, SwitzerlandPLUTO Kollaboration, Ch.Berger et.al.,PL 81B (1979) 410PLUro-Kollaboration, Ch. Berger et.al.,DESY 79/56PLUTO Kollaboration, Ch.Berger et al.,DESY 79/57F.A. Berends et ale NP B63 (1973) 381

NP 868 (1974) 541PL 63B (1976) 432

L.H. O'Neill et.al.,PRL 34 (1975) 233Mark~ Kollaboration, o. Barber et al.,PRL 42 (1979) 1110

2 as(s)We used R= = 3 L Q

1" (1 + ---7T-- );expec

a s (s)=127TC33-2Nf )lnCs/A 2 ) -1;A=O.5GeV

R. Field and R.P. FeynmanNP, 8136 (1978) 1J. EIIIS, M.K. Caillard, O.V. Nanopoulos,S. Rudaz, NP 8131 (1977) 285A. Ali, J.G. Korner, G. Kramer and J.Will­rodt, OESY 79/16Particle properties bookletJ Particledata group April 1978

-32-

23.

24.

25.

26.

27.28.29.

30.

31 .32.

33.

34.

35.36.

37.

38.

39.

40.

41 •

42.

M.K. Gaillard, 8.W. Lee and J.L. RosnerRev. Mod. Phys. 47 (1975) 277C. Quigg and J.L:-RosnerPRO 17 (1978) 239A. Ali, Z.Physik.C. Particles and Fields1 (1979) 25 and private communicationA. Ali, J.G. Karner, G. Kramer and J.Willrodt, DESY 78/67 and Z.Physik C1(1979) 203J. Ellis, M.K. Gaillard and G. RossNP B111 (1 976) 253P. Hoyer, P. Osland, H.G. Sander, T.F.Walsh and P.M. Zerwas, DESY 79/21,to be publishedG. Wolf, DESY 79/71S.8randt and H.D. Dahmen Z.Phys. C1 (1979)61PLUTO Collaboration, Ch. Berger et al.,PL 82B (197~) 449J.D:-Bjorken and S.J. Brodsky, PRD1(1970) 1416 --C. Bromberg et al., PRL 43 (1979) 565J.G.H. de Croot et al., PL 828 (1979)292 and 456, Z. Physik C1 (1979) 143P.C. Bosetti et al., NPB149 (1979) 13S .. I. Ei"delman et al., P LB"2B .(197.9) 278

N. Arteago Romero, A. Jaccarini andP. Kessler, C.R. Acad.Sci. ~ (1969)153, and C.R. Acad. Sci. ~ (1969) 1129,V.E. Balakin, V.M. Budnev and I.F. Ginz­burg, Zh. Eksp. Thear. Fiz. Pis'ma Red11 (1970) 559 (JETP Lett. 11, 338);s:J. Brodsky, T. Kinoshita-and H. Tera­zawa, PRL 25 (1970) 972.E.g. S.J. Brodsky, SLAC PUB 2240 (1979)If not defined in the text, the symbolsused are explained in fig.1We are gratefu 1 to J. A.11,- Va.rmaseren forhelping us with this programL.N. Hand, PR ~ (1963) 1834,

S.J. 8rod~ky, T. Kinoshita and H. Tera­zawa, PRD4 (1971) 1532C. Carimalo, P. Kessler and J. Parisi,College de France preprint LPC 79-17E.g. S.J. Brodsky, .j. Physique, C-2,Suppl.3 (1974) 69 .M. Greco, Y.Shrivastava, N.C. 43A (1978)88

Q and A session

Speaker: John Yah-ColumbiaQ. I have a question about one of your earlier

plots on Evis divided by 15. It seems to m~that if I remember properly}there is a largenumber of events with Evis significantlyless than Is. Could you show that plotagain? Just how do you account for thedifference between your Evis and Is?

A. This plot? (Fig.20)Q. That's right. It seems that it peaks around

0.8. Is that because you"re missing a lotof events?

A. No that is due to particle 10ss8s in anevent. It is a resolution effect.

Q. Oh, its resolution, I see.Speaker: V. LOth-SLACQ. Is the beam gas background subtracted?A. Yes, in fig.20 this background is subtrac-

ted.Speaker: J. Rosner-MinnesotaQ. What is the relation between Wvis and W

for the two photon events?A. It is typically about 1-2 GeV below N,

but it is hard to ~isentangle it complete~

ly right now, We ran a Monte Carlo

and found that it is typically 1-2 GeV be­low. (Note added: T-hE~ ranqe of tf contr; but; noto Wv;s is now included in fig. 23 and 25) v

Speaker: Arie Sadek-RochesterQ. Can you comment on the mean multiplicity as

a function of energy and how you expect itto affect the search for new specific fi­nal states?

A. We did not look for specific final statesat the high energy, but I showed you thegraph of the multiplicity. The multiplici­ty is growing rapidly, so to say. If youwant, if you really wa~t to extrapolatefrom the low energy regime, it seems torise faster than Ins. The charged multi­plicity is now around 11 in our data atthe highest possible energy.

Speaker: G. Belletini-FrascatiQ. Did you look for speclflc flnal states

in the 2y events?A. We did a study, we looked in two prongs

for example for resonances. For three andfour prongs we do not find any mass peak.For two prongs we have some results.

-33-