Fermi National Accelerator Laboratory

Measurement of the Isolated Prom t Photon Cross Section inpp Collisions at 1p s=1.8 TeV

The CDF Collaboration

Fermi National Accelerator Laboratory P.O. Box 500, Batavia, Illinois 60510

February 1992

Submitted to Physical Reuiew Letters.

+operatedbyu. nrversiSes Research A-ssxiation Inc. under Cmtmct No. DE-ACM-76CHCNOOO tilh tie United States bpatbnent of Energy

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accumcy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specifti commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein da not necessarily state or reflect those of the United States Government or any agency thereof,

FERMILAB-PUB-92/31-E CLlF/PUB/.JET/CDFR/1516

Measurement of the Isolated Prompt Photon Cross Section

in fip Collisions at fi = 1.8 TeV

F. Abe,(‘) D. Amidei, G. Apollinari, (*‘V M. Atac,(*) P. Au~bincloss,(‘~) A. R. Bade@)

N. Bacchetta$“) M. W. Bailey,(“) A. Bamberger,(*“) P. de Barbam, B. A. B.m,ett,@)

A. Barbam-Galtieri,(‘O) V. E. Barnes,@*) T. Bauma~,(~) F. Bedescbi,(l3) S. Behrends,(*l

S. Belforte,(13) G. Bellettini,(13) J. Bellinger,(zll D. Benjmnin,(ao) J. Bensinger,

A. Beretvas,(‘l J. P. Berge,(*) S. Bertolucci,(6) S. Bha&a,(‘) M. Bb&ley,(*) D. Blsello,(l’l

R. Blair,(‘) C. Blocker,(‘) A. Bodek,(15) V. Bolognesi,(13) A. W. Booth,(*) C. Boswell,(a)

G. Brandenburg,(e) D. Brown,(e) E. Buckley-Gee+‘) H. S. Budd,(lS) G. Busetto,(“)

A. Byon-Wagner,(‘) K. L. Byrun,(211 C. Campagnari,(3) M. Campbell,(3) A. Caner,(*)

R. Carey,(e) W. Carithers,(lo) D. Carlsmith,(21) J. T. Carroll,(*) R. Ca&more,(*@)

A. Castro,(“) F. Cer~elli,(‘~) K. Chadwick,(*) G. ChiarelX,(s) W. Chinowsky,(‘ol

S. Cihangir,(‘) A. G. Clark,(*) M. Cobal,(131 D. C onnor,@) M. Contrerasf’) J. Cooper,(*)

M. Cordelli,@) D. C rane,(*) M. Cwatolc@ C. Day,(*) F. DeJongh,(*) S. Dell’Agnello,(13)

M. Dell’Orsc~,(‘~l L. Demortier,(a) B. Denby, P. F. Denvent,@) T. De&n,(“)

D. DiBitonto,(18) M. Dickson,(15) R. B. Drucker,(‘o) K. Einsweiler,(lo) J. E. El&,(*)

R. Ely,(“) S. Eno,c3) S. Errede,(‘l B. Esposito,(61 A. Etchegoyen,(*+) B. Flaugher,(*)

G. W. Foster,(*) M. l?rankli11,(~1 J. Freeman,(*) H. Frisch,(3) ‘7. Fuess,(*l y. F&$(Q)

A. F. Gafinkel,(‘*) A. Gautbier,c7) S. Geer,(*l D. W. Gerdes,t3) P. Gi-etti,(‘al

N. Giokaris,(‘e) P. Giromini,@‘) L. Gladney,(lz) M. Gold,@) K. Gotiaos,(le)

Submitted to Physical Review Letters, FebTuary 3, 1992.

II. Grassmann, J. Grieco,(13) C. Grosso-Pilcher,@) C. H&x,(10) S. Il. Hahn,c4)

Il. Handler,(21) K. Hara,(“) B. Hamal, R. M. Harris,(*) J. Hauser,c4) C. Hawk{‘71

T. HessinS, R. Hollebeek,(12) L. Holloway,(‘) P. Hu,(‘~) B. Hubbard,(1o)

B. T. H&mn,(‘*) R. Hughes,(12) P. Hurst,@) J. Huth,@) J. Hylen,(4) M. Imgli,(13)

T. Ino,@) H. Iso,@) H. Jensen, C. P. Jessop,@‘) R. P. Johmon,t4) U. Jo&,@)

R. W. Kadel,(“‘) T. K-on,(“) S. Kanda,(“) D. A. Kardelis,(7) I. Karliner,(7)

E. Kearn~,(~) L. Keeble,(‘*) R. Kephart,(‘) P. K es t en,(‘) R. M. Kemp, H. Keutelian,(4)

D. Kim,t4) S. Kim,@“) L. Kirsch,(‘) K. Kondo,@) J. Konigsberg,(~) E. Kovacs,(4)

S. E. Kuhlmann,(l~ E. Ku+‘) A. T. L aasanen,(“) J. I. Lam~ureux,(~*) S. Leom@

J. Lewis,(‘) W. Li,(‘) P. Limon,c4) T. M. Liss,(‘) N. Lockyer,(‘a) M. Loret& E. Low,(l2)

C. B. Luchini,(‘) P. Lukens,@) P. Maas, K. Maeshim=+) M. Mmgmo,(‘3)

J. P. Mariner,(‘) M. Mariotti,(13) R. Markeloff,(zl) L. A. Markosky,@) R. Mattb&,(z)

P. Mclntyre,(“) A. Menzione,@) T. Meyer,(‘s) S. Mikamo,@) M. Miller,@) T. Mima&,(*g)

S. Miscetti,@) M. Mishina,@) S. Miyashita,(lg) Y. Morita,(ls) S. Moulding,(z) J.

Mueller,(“) A. Mukherjee,(4) L. F. Nakae,@) I. N&mo,(le) C. Nelson,

C. Newman-Hol~nes,(~) J. S. T. Ng,@) M. N’ mnomiya,(‘*) L. Nodulman,@) S. Ogawa,(19)

R. Paoletti,(13) V. Papadimitriou,(4) A. Para, E. Pare,(B) S. Park,(*) J. P&r&,(4)

G. PaUletta,(13) L. Pescara,(“) T. J. PhilIips,(B) F. Ptohos,cB) R. Pl&&t,(*l

L. Pondrom,(“) J. Proudfoot, G. Punai,(13) D. Quade,(*) K. Ragan,(*2) G. R&inger,(3)

J. Rhoades,(“) M. Roach,(“) F. R.imondi,(‘,‘) L. Ristori,(‘s) T. Rod&go,(*) T. R&dy,(**)

A. R~odman,(~) W. K. Sakumoto,(16) A. Sansoni,(6) R. D. Sard,(‘) A. Savoy-Navarro,(4)

2

V. Scarpine, P. Schlabach,@) E. E. Schmidt,(*) 0. Sclmeider,(‘“) M. H. Schub,(‘*)

R. Schwitters,(B) A. Scribano,(13) S. Segler,(*) Y. Seiya,@) M. Shapiro,(lO) N. M. Shaw,(l*)

M. She&(‘*) M. Shochet,(3) J. Siegrist, 0’) P. Sinervo,(‘2) J. Skarha,(‘) K. Sliwa,@“)

D. A. Smith,(*3) F. D. Snider,@) L. Song, (la) M. Spahn,(‘“) P. Sphicas,@) R. St. D&s,@)

A. Stefanini,(13) G. Sullivan,@) R. L. Swartz, Jr.,c7) M. Taka~o,(‘~) F. Tartarelli,(*3)

K. Takikawa,(*‘) S. Tarem, D. Theriot, M. Timko,(le) P. Tipton, S. Tkaczyk,(*)

A. Toll&up,(*) J. Ton&on, (**) W. ‘lXschuk,(@) N. Turini,(13) Y. Tsay,@) F. Ukegawa,(*s)

D. Underwood,(‘) S. Vejcik, III,@) R. Vidal,(*) R. G. Wagner,(*) R. L. Wagner,(*)

N. Waim@ J. Walsh,(=) T. Watts,(*‘) R. Webb,@*) C. Wendt,@‘) H. Wenzel,(-)

W. C. Water, III,(“‘) T. Westhusing, (13) S. N. White,@‘) A. B. Wicklund,

H. H. Williams,(*z) B. L. Winer,@) J. Wyss, (**) A. Yagil,c4) K. Yasuoka,(*B) G. P. Yeh,(*)

J. Yoh,(‘) M. Yokoyama,(‘@) J. C. Yun,(*) A. Zanetti,(*3) F. Zetti,(*3) S. Zucchelli (*,a)

The CDF Collaboration

* A7gonn. National L.b.dOry, Argonne, minoi, so,99

’ Bm.ndria Uniwraity, Waltham, Maamrhuactta 02t64

’ Fnnimrrit~ of Chicsge, Chicago, II&wit 60657

4 Fermi Nationd Acrrlrrdor Ldorotory, Batasia, Illin& SOI*

’ Laborotori Nazimali di From& I&MO Nazion& di Firirn Nurlcarr, Fmacoti, It&

’ Xnrt.ord Crnimrrity, Cambridge, Mmrachuarttr 01198

’ Univrrrity of Illin&, Urbono, Illinois 61801

a The John, Hopbin, Unimrrity, Baltimore, Maryland llll8

3

9 N.aNoti0r.d Laboratory fo, High Enqy Phy,irs (KEK). J.apan

lo Lamwwr Bmkrlry LAw-atov, Bwkcloy, Ccrlifornio e4i’ZO

** Uniscrrita di Padora, lnrtituto Nazion& di Firira Nurlrclrr, Scrionr di Podma, 1.56191 Padova, Itory

*’ Unimraity of Pmnrylmnia, Philadelphia, Pmnq&mia 1810~

*’ Iatitxto Nozionolc di Fir& Nwlrorr, Unitwrity nnd Scuola Norm& Supcriorc of Piro, 1.66100 Pira, IMy

** Purdue Uniocrrily, Weal Lafoyrttq Indiana 17907

l5 Uniowrily of Rochmtw, Rochutcr, New York Id617

*’ Rockrfdler Unixraity, New York, New Ymk 100,U

*’ Autprra Vnircrrity, Piaratawoy, New JIV,LY 08854

** Tezar AUM Unimrnity, Colkge Stdion, Team 7i’8/9

*’ Unirerrily of Tsukvbn. Takuba, lbamki 30.6, Japan

” Tufts University, Mcdfmd, Mauachuntta OZ1.66

‘* hiorrritv of Wiacowin, Mod&m, Wiacon&n 5570.9

4

Abstract

We present a measurement of the cross section for production of isolated prompt

photons in jip collisions at fi = 1.8 TeV. The cross section, measured as a function

of transverse momentum (PT), agrees qualitatively with QCD calculations but has a

steeper slope at low PT.

In this letter we present the first measurement of the cross section for production

of prompt photons in proton-antiproton collisions at fi = 1.8 TeV. Prompt photons

are produced in the initial collision, in contrast to photons produced by decays of

hadrons. This measurement can be used to test Quantum Chromodynamics (QCD).

In &CD, at lowest order, prompt photon production is dominated by the Compton

process (gq + rq), which is sensitive to the gluon distribution of the proton. We

measure prompt photons in a previously unexplored range of fractional momentum

(.016 < z < .Oi'O).

A detailed description of the Collider Detector at Fermilab (CDF) may be found

in Ref. [l]; the components relevant for this analysis are described briefly here. We

use a coordinate system with z along the proton beam, azimuthal angle 4, polar

angle 0, and pseudorapidity 7 = -In tan(O/2). Scintillator-based electromagnetic

(EM) and hadronic (HAD) calorimeters in the central region (171 < 1.1) are arranged

in projective towers of size A7 x Ad N .l x .26. The central electromagnetic strip

5

chambers (CES) are multi&e proportional chambers embedded inside the central

EM calorimeter near shower maximum (6 radiation lengths and 184 cm from the

beam). The CES anode wires measure 4 and cathode strips measure 7; both views

have a channel separation of roughly 8 milliradians for measuring the transverse

profile of electromagnetic showers. The central drift tubes (CDT) are three layers of

gas counters just outside the central tracking chamber (CTC). In this analysis, the

CDT measures electron-positron pairs from photons which convert in either the outer

wall of the CTC or the inner two layers of the CDT (18% of a radiation length total).

Photon data were taken with a high threshold trigger and a prescaled low threshold

trigger. An integrated luminosity of 3.3(0.10) pb-’ was acquired with the high(low)

threshold trigger which required 23(10) GeV of EM transverse energy; the triggers

were 90%(98%) efficient for photons with 27(14) GeV of PT. Throughout this arti-

cle PT is the component of the photon momentum transverse to the beam direction,

and transverse energy is defined similarly. To reject jet backgrounds, these triggers

required that at least 89% of the transverse energy of the photon be in the EM

compartment of the calorimeter, and required the photon to be isolated: the extra

transverse energy inside a cone of radius R = dm = 0.7 centered on the

photon was required to be less than 2 GeV. The efficiency of the isolation cut, inde-

pendent of photon PT, was estimated to be 89.5%; this is the measured probability

that an isolation cone in a minimum bias event will contain less than 2 GeV.

6

A prompt photon candidate is an isolated cluster in the central EM calorimeter

with no charged track pointing at the cluster. The only significant background sources

are the decays of the neutral mesons x0 and 7 into photons. We employ two methods

for statistically subtracting the neutral meson background from our photon candi-

dates: the profile method uses the transverse profile of the electromagnetic shower

in the CES and the conversion method counts electron-positron pairs from photon

conversions which produce hits in the CDT.

For both methods, within the tower boundaries of a calorimeter energy cluster[2],

we formed CES strip and wire clusters, each containing 11 channels (total width z 0.1

radians) centered on a seed channel with at least 0.5 GeV. The highest energy strip

cluster and the highest energy wire cluster were chosen for measuring the transverse

profile and position of the photon candidate. Backgrounds from 7 mesons were re-

duced by requiring that any additional CES clusters within the boundaries of the

calorimeter energy cluster had less than 1 GeV each. The efficiency of this cut, rang-

ing from 95% (at 14 GeV) to 81% (at 63 GeV), was determined from measurements of

electrons in a test beam. Fiducial cuts were imposed to avoid uninstrumented regions

at the edges of the CES; the region remaining was 66% of the solid angle for 171 < 0.9.

To maintain the projective nature of the calorimeter tower, the z position of the event

vertex was required to be within 50 cm of the center of the detector; this cut was 88%

efficient. Finally, cosmic ray events which included bremsstmhlung photons had large

7

transverse momentum imbalance which made them easy to identify and remove. The

data sample for the conversion method came from the high threshold trigger, and

the analysis used the same selection cuts with the following additional cuts to reduce

backgrounds: CES average 2’ < 8 (defined below), azimuthal separation between

CDT hits and the CES cluster Ad < 0.07, and z separation between CDT hits and

the CES cluster AZ < 10 cm (at 1.4 meters from the beam). The total acceptance a,

including efficiency, for the profile ( conversion) method was between 50% (44%) and

43% (40%) depending on PT.

The transverse profile of each photon candidate was compared to that measured

for electrons in a test beam. A measure of the goodness of fit (2” [3]) was usually

larger for a neutral meson (poor fit) than for a single photon (good fit) because a

neutral meson usually produced a wider EM shower. The g* variable is arbitrarily

normalized and has a mean value of approximately 2.5 for electron showers. The

average of 2’ in both views ((strip 2’ + wire 2’)/2) was the variable used to separate

single photons from neutral mesons in the profile method.

After all cuts, the final sample was roughly half signal and half background. The

number of photons (N,) in a bin of PT is obtained in each method from the number

of photon candidates (N), the method’s efficiency for photon candidates (E), and the

corresponding efficiency for true photons (c,.) and background (es), using:

N7 =

8

(1)

Equation 1 comes from EN = c7N-, + tbNb with Nb = N-N-,. For the profile method,

the efficiency E is the fraction of events which have ga < 4 out of all events with

g’ < 20. For the conversion method, e is the fraction of photon candidates which

produce a conversion hit in the CDT.

For the profile method we estimated L, and Eb using a detector simulation. The

simulation employed real electron showers, measured in a test beam, and corrected for

differences between photons and electrons. Figure 1 shows that we obtain good agree-

ment between the data and simulation for the strip 2’ distribution of single photons

from q meson decays, and for the average 2’ distribution of single r’s from p meson

decays. There were three sources of systematic uncertainty in the simulation of the

photon efficiency e7: the transverse shape of photon induced electromagnetic show-

ers, the statistical fluctuations in the transverse profile of photons, and the linearity

of gas gain response of the CES to electrons in the electromagnetic shower. Fig. 2a

shows the measured efficiency c for the profile method, as a function of transverse

momentum, along with the simulated efficiencies L, and Q, and their total system-

atic bounds. To obtain Eb we simulated the neutral mesons r”, 11, and Kz with a

relative production ratio of 1:1:0.4 determined as follows. From the data shown in

Fig la, using three channel CES clusters, we measured the ratio of isolated 7 to no

production at 12 GeV: q/r” = 1.02 f .lS(stat) k .23(sys) [3]. The ratio Kz/# N 0.4

for PT > 3 GeV is implied from our previously reported measurement of Kg/r+ and

9

isospin invariance [4]. The systematic uncertainty induced by the uncertainties in

these measurements has been included in the systematic bound on the background

efficiency; however the background systematic uncertainty is dominated by the three

previously mentioned uncertainties in the photon efficiency. Figure 2b shows that

the simulated g2 distribution of photons and background, combined in the relative

proportions predicted by the profile method, is in good agreement with the data. At

high PT the two photons from the decay of a ?y” are so close together that the 2’

efficiency for the background is almost the same as for a single photon; consequently,

the profile method was only used up to PT = 40 GeV/c.

For the conversion method, we estimate the production ratio T/I?’ at low PT from

the azimuthal separation of the CDT hit and the highest energy CES cluster. From

this ratio, the number of prompt photon candidates, and the total number of CDT

hits associated with the prompt photon candidates, we estimated the efficiency of

photon conversion and detection to be 0.095 f 0.014(&t) f O.OlO(sys). The effi-

ciency expected from the amount of material and the selection cuts is about ll%,

consistent with this estimate. The uncertainty in the efficiency, which is independent

of PT, dominated the uncertainty in the normalization of the cross section from the

conversion method. The efficiency of conversion and detection for the background is

estimated from the photon efficiency and a simulation of the background from KO, 7

and Ki mesons. The background efficiency is the same, within errors, as the naive

10

probability of observing at least one photon conversion from a two photon decay:

eg = 29 - 6;. Figure 2a shows the measured conversion efficiency c, for the con-

version method, as a function of transverse momentum, along with the estimated

efficiencies ct and Eb and their total systematic bounds.

From the number of prompt photons (NY) in a bin of transverse momentum (APT)

and a range of pseudorapidity (A7 = 1.8), using the acceptance (cx) and the integrated

luminosity (L), we obtain the isolated prompt photon cross section:

d2u NT -= d&h APT.AT.c~.L: (2)

which is shown in Fig. 3 and tabulated in Table I. Here we present the profile method

in the low PT region (14 < PT < 40 GeV/c) and the conversion method in the high

Pr region (28 < PT < 68 GeV/c). Th t e wo methods agree in the region in which

they overlap. In Fig. 3, the inner error bars are the statistical uncertainty and the

outer error bars are the PT dependent part of the systematic uncertainty combined

in quadrature with the statistical uncertainty. The PT independent component of the

systematic uncertainty is shown as the normalization uncertainty separately for each

of the two methods.

In Fig. 3 our measurements are compared to a next to leading order QCD cal-

culation [S] using a single set of parton distributions [6] at a renormalization scale

p = PT. The QCD prediction changes by less than 30% when the parton distribu-

tions are varied among commonly used sets; it decreases (increases) by 12% when the

11

renormalization scale is doubled (halved). The calculation includes the experimental

isolation cut. Figure 3 shows that the measured cross section agrees qualitatively with

QCD calculations but has a steeper slope at low PT. Data acquired at the CERN pp

collider [7] (J;; = 630 GeV) show similar behavior. One possible cause of the differ-

ence between the data and the QCD calculation is the bremsstmhlung process [S, 91,

prevalent at low PT, in which a final state quark radiates a photon.

We thank the Fermilab staff and the technical staffs of the participating institu-

tions for their vital contributions. This work was supported by the U.S. Department

of Energy and National Science Foundation, the Italian Istituto Nazionale di Fisica

Nuclear=, the Ministry of Science, Culture and Education of Japan, and the Alfred

P. Sloan Foundation. We also wish to thank J. F. Owens for providing his computer

code for the theoretical calculations.

References

[l] F. Abe et al., Nucl. Inst. and Meth. A271(1988)387.

[Z] F. Abe et al., Phys. Reu. D43(1991)2070.

[3] F. Abe et al., “A Prompt Photon Cross Section Measurement in pp Collisions

at fi = 1.8 TeV”, Fermilab-PUB-92/01-E, 1992 (to be published).

[4] M. Schub Ph.D. Thesis, Purdue Univ., 1989 (unpublished).

12

[5] H. Baer, J. Ohnemus and J. F. Owens, Phys. Mt. B234(1990)127.

[6] Set B. in J. Kwiecinski, A. D. Martin, R. G. Roberts and W. J. Stirling, Phys.

Rev. D42(1990)3645.

[7] J. Alitti et al. (UA2 Collaboration) Phys. Lett. B263(1991)544.

[8] E. Pilon, P. Aurenche, M. Fontannaz, and J. Guillet, in Proceedings of the XXVI

Recontre de Moriond, Les Arcs, France, 1991 (to be published).

[9] E. Berger and J. Qiu, Phys. Rev. D44(1991)2002.

13

Table I: The cross section for isolated prompt photons, the statistical uncertainty,

and the PT dependent component of the systematic uncertainty. There is an

additional normalization systematic uncertainty of 27% in common among the

first eleven entries (profile method), and +32%(-46%) for the last four entries

(conversion method).

I PT Bin [GeV/c)

14-15 15-17 17-19 19-22 22-27 27-28 28-29 29-31 31-33 33-35 35-40 28-38 38-48 48-58 58-68

PT

GeV/c) 14.5 15.9 17.9 20.4 24.2 27.5 28.5 30.0 32.0 34.0 37.3 32.2 42.4 52.5 62.6

3.16 x lo3 1.55 x 10s 1.03 x 103 4.36 x 10’ 1.91 x 102 1.30 x 102 1.13 x 101 7.15 x 10’ 6.98 x lo1 3.78 x lo1 2.23 x lo1 6.05 x 10’ 1.19 x 101 6.53 x 10” 2.22 x loo

‘ Stat.

w 11 12 13 18 22 12 12 12 11 20 20 15 37 41 79

jys.

(%) 21 13

6 2

12 23 26 32 40 50 71

7 6 6 8

14

0.75 I 1.25

Figure 1: a) The invariant mass of two photons frommeson decays (points), a two gaussian

plus background fit (solid), the polynomial-like background (dashes), and the expected

background from misidentified single photons (dots). b) The invariant mass of a neutral

cluster and charged track (points) compered to a Breit-Wigner plus background fit (solid)

and a polynomial background (dashes). c) The 2s distribution of photons (triangles) from q

decays in a), and of ?y”s (circles) from p decays in b), are compared to a detector simulation

(curves).

15

W 0 LL

kl

--

‘b --

EY A A

4 -- --

~ot”“““““““““““““1 10 20 30 40 50 60

fT (GeV/c)

Figure 2: a) The efficiency of photon candidates (points), and simulated photons and

background (solid curves) with systematic bounds (dashed curves) for the profile method

(circles) and the conversion method (triangles). The bounds on l , and q, are completely

correlated within each method. b) The i’ distribution of data (points) with PT < 27 GeV/c

compared to our simulation of photons (dots), background (dashes), and their sum (solid).

16

103

- s 102 s \ a a

a+ 7 b N

-0 loo

‘r CDF & = 1.8 TeV CDF X& = 1.8 TeV (NORM. UNCERTAINTY) (NORM. UNCERTAINTY)

n UA2 Js =630 GeV GeV

- QCD x

- QCD

10-l z- NLO, KMRS-B,, p=PT

\ I I I I I I I I I I I I I I III III

20 40 60 80

PT (GeVk)

Figure 3: The isolated prompt photon cross section, from both CDF (circles and trian-

gles) and UA2 [7] (squares), is compared to recent QCD predictions (cuws). The profile

method (circles) and conversion method (triangles) have separate normal.ieation uncertain-

ties shown in the legend. The next to leading order (NLO) predictions [5] use current parton

distributions (KMRS-& [6]) and a standard renormalization scale (p = PT).

17