physics results with the argo-ybj experiment

Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 51 (2013) 268–279

Physics results with the ARGO-YBJ experiment

Benedetto D’Ettorre Piazzoli 1

Dipartimento di Fisica, Universita “Federico II”, Napoli and Sezione INFN, Italy

Available online 17 October 2011

Abstract

ARGO-YBJ is a multipurpose experiment consisting in a dense sampling air shower array with 93% sensitive area located at very highaltitude. The apparatus is in stable data taking since November 2007 at the YangBaJing Cosmic Ray Laboratory (Tibet, PR China,4300 m a.s.l., 606 g/cm2). In this paper we report the main results in Gamma-Ray Astronomy and Cosmic Ray Physics after about3 years of operation.� 2011 Published by Elsevier Ltd. on behalf of COSPAR.

Keywords: Extensive air showers; Cosmic rays; ARGO-YBJ experiment

1. Introduction

The ARGO-YBJ detector is in stable data taking sinceNovember 2007 with a duty cycle >85% and with excellentperformance. The large field of view (�2 sr) and the highduty cycle allow a continuous monitoring of the sky inthe declination band from �10� to 70�.

The experiment, promoted and funded by the ItalianInstitute for Nuclear Physics (INFN) and by the ChineseAcademy of Sciences (CAS) in the framework of theItaly–China scientific cooperation, is operating at theYangBaJing Cosmic Ray Laboratory (Tibet, PR China,4300 m a.s.l., 606 g/cm2).

Location and detector features make ARGO-YBJcapable of investigating a wide range of fundamental issuesin Cosmic Ray and Astroparticle Physics at a relatively lowenergy threshold:

� high energy c-ray astronomy, at an energy threshold of afew hundred GeV;� search for emission of gamma ray bursts in the full

GeV–TeV energy range;

0273-1177/$36.00 � 2011 Published by Elsevier Ltd. on behalf of COSPAR.

doi:10.1016/j.asr.2011.09.037

1 On behalf of the ARGO-YBJ Collaboration.E-mail address: [email protected]

� cosmic rays physics (energy spectrum, chemical compo-sition, �p=p ratio measurement, shower space-time struc-ture, multicore events, p-air and pp cross sectionmeasurement) starting from TeV energies;� Sun and heliosphere physics above �1 GeV.

In the following sections, after a description of thedetector and its performance, the main results achievedafter about 3 years of stable data taking are reviewed.

2. The ARGO-YBJ experiment

The ARGO-YBJ experiment is currently the only airshower array with a dense sampling active area operatedat high altitude, with the aim of studying the cosmic radia-tion at an energy threshold of a few hundred GeV. The largefield of view (�2 sr) and the high duty cycle (P85%) allow acontinuous monitoring of the sky in the declination bandfrom �10� to 70�.

The detector is composed of a central carpet large�74 � 78 m2, made of a single layer of Resistive PlateChambers (RPCs) with �93% of active area, enclosed bya guard ring partially (�20%) instrumented with RPCs upto �100 � 110 m2. The apparatus has a modular structure,the basic data acquisition element being a cluster(5.7 � 7.6 m2), made of 12 RPCs (2.8 � 1.25 m2 each). Eachchamber is read by 80 external strips of 6.75 � 61.8 cm2 (the

http://dx.doi.org/10.1016/j.asr.2011.09.037

mailto:[email protected]

http://dx.doi.org/10.1016/j.asr.2011.09.037

) E°(mδ) cosmα-αW (-4 -3 -2 -1 0 1 2 3 4

)

N°(

mδ-δS -4

-3

-2

-1

0

1

2

3

4

-50

-40

-30

-20

-10

0

Fig. 1. Moon shadow significance map observed by the ARGO-YBJdetector in 3200 h on-source for events with Nstrip P 40. The coordinatesare Right Ascension a and declination d centered on the Moon position(am,dm). The color scale gives the statistical significance.

B. D’Ettorre Piazzoli / Advances in Space Research 51 (2013) 268–279 269

spatial pixels), logically organized in 10 independent pads of55.6 � 61.8 cm2 which represent the time pixels of the detec-tor. The read-out of 18,360 pads and 146,880 strips is thedigital output of the detector (Aielli et al., 2006). The rela-tion between strip and pad multiplicity has been measuredand found in fine agreement with the Monte Carlo (MC)prediction (Aielli et al., 2006). In addition, in order toextend the dynamical range up to PeV energies, each cham-ber is equipped with two large size pads (139 � 123 cm2) tocollect the total charge developed by the particles hitting thedetector. The high granularity of the detector readoutallows a detailed reconstruction of the space-time profileof the shower front and of the charge distribution aroundthe core. The RPCs are operated in streamer mode by usinga gas mixture (Ar 15%, Isobutane 10%, TetraFluoroEthane75%) for high altitude operation. The high voltage set at7.2 kV ensures an overall efficiency of about 96% (Aielliet al., 2009a). The central carpet contains 130 clusters (here-after, ARGO-130) and the full detector is composed of 153clusters for a total active surface of �6700 m2. ARGO-YBJoperates in two independent acquisition modes: the shower

mode and the scaler mode. In shower mode, for each eventthe location and timing of every detected particle isrecorded, allowing the reconstruction of the lateral distribu-tion and the arrival direction (Di Sciascio et al., 2007). Inscaler mode the total counts on each cluster are measuredevery 0.5 s, with limited information on both the space dis-tribution and arrival direction of the detected particles, inorder to lower the energy threshold down to�1 GeV (Aielliet al., 2008). In shower mode a simple, yet powerful, elec-tronic logic has been implemented to build an inclusive trig-ger. This logic is based on a time correlation between thepad signals depending on their relative distance. In thisway, all the shower events giving a number of fired padsNpad P Ntrig in the central carpet in a time window of420 ns generate the trigger. This can work with high effi-ciency down to Ntrig = 20, keeping the rate of random coin-cidences negligible. The time of each fired pad in a windowof 2 ls around the trigger time and its location are recordedand used to reconstruct the position of the shower core andthe arrival direction of the primary particle (Di Sciascioet al., 2007). In order to perform the time calibration ofthe 18,360 pads, a software method has been developed(Aielli et al., 2009b).

The whole system, in smooth data taking since July 2006firstly with ARGO-130, is in stable data taking with the fullapparatus of 153 clusters since November 2007, with thetrigger condition Ntrig = 20. The trigger rate is �3.5 kHzwith a dead time of 4%. All the results presented in thispaper have been obtained by analysing the data collectedwith the digital readout.

2.1. Detector performance

The performance of the detector and the operationstability are continuously monitored by observing theMoon shadow, i.e., the deficit of cosmic rays (CR)

detected in its direction. Indeed, the size of the deficitallows the measurement of the angular resolution andits position allows the evaluation of the absolute point-ing accuracy of the detector. In addition, since chargedparticles are deflected by the geomagnetic field (GMF),the observation of the displacement of the Moon pro-vides a direct calibration of the relation between showersize (the recorded pad/strip multiplicity) and primaryenergy.

ARGO-YBJ observes the Moon shadow with a sensitiv-ity of �9 s.d. per month for events with a multiplicityNpad P 40 and zenith angle h < 50�, corresponding to aproton median energy Ep � 1.8 TeV. With all data fromJuly 2006 to December 2009 (about 3200 h on-source intotal) we observed the CR Moon shadowing effect with asignificance of about 55 s.d. (Fig. 1). The data analysisand a full account of the results are given in (Bartoliet al., 2011). The measured angular resolution is better than0.5� for CR-induced showers with energies E > 5 TeV(Fig. 2), in excellent agreement with the MC evaluation.With the same simulation codes we find that the angularresolution for gamma-rays is smaller by about 30–40%,depending on the multiplicity, due to the better definedtime profile of the showers.

The overall absolute pointing accuracy is �0.2�. Thisdisplacement towards North is independent of the multi-plicity. The most important contribution to this systemat-ics is likely due to a residual effect not completelycorrected by the time calibration procedure. The angularresolution stability is at a level of 10% in the periodNovember 2007–November 2010.

The relation between the observed strip multiplicity andthe CR primary energy is shown in Fig. 3. The absoluteenergy scale uncertainty of ARGO-YBJ is estimated tobe less than 13% in the energy range 1–30 TeV/Z.

multiplicity210 310

)°an

gula

r res

olut

ion

(

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

DataMC

1.25 1.76 2.30 3.55 6.05 12.2 26.0 [TeV]/Z50E

Fig. 2. Measured angular resolution of the ARGO-YBJ detector com-pared to the expectations from MC simulation as a function of the particlemultiplicity, i.e., the number of fired strips. The multiplicity bins areshown by the horizontal bars. The upper scale refers to the rigidity (TeV/Z) associated to the median energy in each multiplicity bin.

270 B. D’Ettorre Piazzoli / Advances in Space Research 51 (2013) 268–279

3. Gamma-Ray Astronomy

With about 3 years of data we observed 4 sources with asignificance greater than 5 s.d.: Crab Nebula, Mrk421,MGRO J1908+06 and MGRO J2031+41. Details on theanalysis procedure (e.g., data selection, background evalu-ation) are given in (Aielli et al., 2010; Bartoli et al., 2011).

3.1. The Crab Nebula

We observed the Crab signal with a significance of 17 s.d.in about 1200 days, without any event selection to reject had-ron induced showers. According to MC simulations, 84% ofthe detected events comes from primary photons of energiesgreater than 300 GeV, while only 8% comes from primariesabove 10 TeV. The lowest energy point of the Crab Nebula

multiplicity210 310

) °an

gle

diffe

renc

e (

-1.8

-1.6

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

Data

MC

MC best fit

10%±MC best fit

1.25 1.76 2.30 3.55 6.05 12.2 26.0 [TeV]/Z50E

Fig. 3. Measured westward displacement of the Moon shadow as afunction of multiplicity (black squares). The data are compared to MCsimulation (red circles). The upper scale refers to the median energy ofrigidity (TeV/Z) in each multiplicity bin. (For interpretation of thereferences in colour in this figure legend, the reader is referred to the webversion of this article.)

spectrum shown in Fig. 4 is at a median energy of620 GeV. The observed flux is consistent with a steadyemission, and the observed differential energy spectrum inthe 0.3–30 TeV range is dN/dE = (3.0 ± 0.3stat) � 10�11�(E/1 TeV)�2.57±0.09stat cm�2 s�1 TeV�1 in agreement withother measurements (Aharonian et al., 2006; Albert et al.,2008a).

3.2. The blazar Markarian 421

Mrk421 is one of the brightest blazars and the closest tous (z = 0.031), characterized by a strong broadband flaringactivity with a variability time scale ranging from minutesto months. A correlation of VHE gamma rays with X-rayshas been observed during single flaring episodes by differ-ent detectors and can be interpreted in terms of the Syn-chrotron Self-Compton (SSC) model. The Mrk421 highvariability makes the long term multiwavelength observa-tion very important to constrain the emission mechanismsmodels.

Mrk421 was the first source detected by the ARGO-YBJexperiment in July 2006 (D’Ettorre Piazzoli et al., 2011)when the detector started recording data with only the cen-tral carpet and was in commissioning phase. ARGO-YBJhas monitored Mrk421 for more than 2 years, studyingthe correlation of the TeV flux with X-ray data. Weobserved this source with a total significance of about 12s.d., averaging over quiet and active periods. ARGO-YBJdetected different TeV flares in correlation with X-rayobservations, as can be seen in Fig. 5 where the ARGO-YBJ cumulative events per day are compared to the cumu-lative events per second of the RXTE/ASM and Swift sat-ellites. To get simultaneous data, 552 days have beenselected for this analysis. The X-ray/TeV correlation isquite evident over more than 2 years. The steepness of

Energy (TeV)

-110 1 10

)-1

TeV

-1 s

-2D

iffer

entia

l flu

x (c

m

-1510

-1410

-1310

-1210

-1110

-1010

-910

-810

ARGO-YBJ 2011

HESS 2006

MAGIC 2008

Fig. 4. The Crab Nebula spectrum measured by ARGO-YBJ comparedwith the results of some other detectors (Aharonian et al., 2006; Albertet al., 2008a). The errors reported are only statistical.

MJD-54466-100 0 100 200 300 400 500 600 700 800

Cum

ulat

ive

even

ts p

er s

econ

d of

RXT

E

0

200

400

600

800

1000

1200

Cum

ulat

ive

even

ts p

er d

ay o

f AR

GO

-YB

J

0

5000

100002008 2009 2010

Fig. 5. Cumulative light curve from Mrk421 measured by ARGO-YBJ(red curve) compared with RXTE/ASM (black curve) and Swfit (bluecurve) X-ray data. The shaded red region indicates the corresponding 1rstatistical error. (For interpretation of the references in colour in thisfigure legend, the reader is referred to the web version of this article.)

Fig. 6. ARGO-YBJ >1 TeV flux (in Crab units) vs. RXTE/ASM X-rayflux. Solid line: quadratic best-fit. Dotted line: linear best-fit.

Flux >1 TeV [Crab units]1 2 3 4 5 6

Diff

eren

tial s

pect

ral i

ndex

-2.8

-2.6

-2.4

-2.2

-2

-1.8

-1.6

Fig. 7. Mrk421 spectral index vs. flux above 1 TeV in Crab units. Thesolid line is the function obtained in (Krennrich et al., 2002).


the curve gives the flux variation, that shows an active per-iod at the beginning of 2008, followed by a quiet phase, andin February 2010.

The degree of correlation and the phase difference (timelag) between the variations in X-ray and TeV bands havebeen evaluated using the Discrete Correlation Function(DCF) (Edelson and Krolik, 1988) that gives the correla-tion coefficient for two light curves as a function of the timelag between them. A quadratic correlation is found (Fig. 6)with a time lag of �0:14þ0:86

�0:85 and �0:94þ1:05�1:07 days for

RXTE/ASM and Swift data, respectively. Therefore, nosignificant time lag is found between X-ray data and TeVARGO-YBJ observations.

The extended data set allows to study the relationbetween the flux and the TeV spectral index. The TeV fluxmeasured by ARGO-YBJ ranges from 0.9 to about 7 Crabunits. As shown in Fig. 7 the TeV spectral index hardenswith increasing flux in agreement with the results obtainedby the Whipple collaboration in 2002 (continuous line)(Krennrich et al., 2002). This conclusion generalizes theARGO-YBJ result obtained with the analysis of the June2008 flare (Aielli et al., 2010). The temporal and the spec-tral analysis strongly support the predictions of the one-zone SSC model. A full account of this analysis can befound in (Bartoli et al., 2011)

3.3. MGRO J1908+06

The gamma ray source MGRO J1908+06 was discov-ered by MILAGRO (Abdo et al., 2007) at a median energyof �20 TeV and recently associated with the Fermi pulsar0FGL J1907.5+0602 (Abdo et al., 2009). The data wereconsistent both with a point source and with an extendedsource of diameter <2.6�. HESS confirmed the discoverywith the detection of the extended source HESSJ1908+063 (Aharonian et al., 2009) at energies above300 GeV, positionally consistent with the MILAGRO

source. The extension of the source was extimatedrext ¼ 0:34�þ0:04

�0:03. The MILAGRO and HESS fluxes are indisagreement at a level of 2–3 s.d., the MILAGRO resultbeing about a factor 3 higher at 10 TeV (Smith, 2009).

ARGO-YBJ observed a TeV emission from MGROJ1908+06 with a significance greater than 5 s.d. in a totalon-source time of 5358 h, Fig. 8. At the ARGO-YBJ site,MGRO J1908+06 culminates at the relatively low zenithangle of 24� and is visible for 5.4 h per day with a zenithangle less than 45�. By assuming a 2D Gaussian sourceshape with r.m.s = rext, we fitted the event distribution asa function of the distance from the center (set to the HESSsource position) and, taking into account the Point SpreadFunction (PSF), we found rext = 0.50 ± 0.35, a value largerbut consistent with the HESS measurement.

The best fit power law spectrum is: dN/dE = 2.2 ±0.4 � 10�13 (E/7 TeV)�2.3±0.3 photons cm�2 s�1 TeV�1

(the errors on the parameters are statistical) (Fig. 9). Thesystematic errors are mainly related to the backgroundevaluation and to the uncertainty in the absolute energyscale. According to our estimate, they globally affect thequoted fluxes for 630%. Given the flat slope, above a

[deg]δ) cos0α - α(-4 -3 -2 -1 0 1 2 3 4

[deg

]0δ

- δ

-4

-3

-2

-1

0

1

2

3

4

-3

-2

-1

0

1

2

3

4

5

Fig. 8. Significance map of the 8� � 8� region around MGROJ1908+06obtained by ARGO-YBJ, for events with Npad P 40. The center of themap represents the position of the HESS source. The open circles show thepositions of the gamma-ray sources observed by Fermi in the same region(Abdo et al., 2010a).

Energy (TeV)1 10 210

)-1

TeV

-1 s

-2D

iffer

entia

l flu

x (c

m

-1510

-1410

-1310

-1210

-1110

ARGO-YBJ

MILAGRO

HESS

Fig. 9. Differential flux from MGRO J1908+06 measured by differentdetectors: ARGO-YBJ (red circles, with best fit given by the continuousline); MILAGRO (the blue dashed line gives the fit and the vertical linesrepresent the 1 s.d. errors for some values of the energy); HESS (blacksquares, with best fit given by the dotted line). The plotted errors arepurely statistical for all the detectors. (For interpretation of the referencesin colour in this figure legend, the reader is referred to the web version ofthis article.)


few TeV the MGRO J1908+06 flux becomes larger thanthe Crab Nebula one. A significant disagreement appearsbetween the ARGO-YBJ and HESS fluxes. In the limit ofthe statistical accuracy of this result, our data supportthe MILAGRO measurement of a flux significantly largerthan that measured by HESS. One possible cause of thisdiscrepancy is that ARGO-YBJ and MILAGRO integratethe signal over a solid angle larger than the HESS one, andlikely detect more of the diffuse lateral tail of the extendedsource. A contribution is expected from the diffuse gamma-ray flux produced by cosmic rays interacting with matter

and radiation in the Galaxy. According to the predictionof the “optimized” GALPROP model (Strong et al.,2000; Porter et al., 2008), confirmed by a MILAGRO mea-surement at 15 TeV (Abdo et al., 2008), the amount of thiscontribution to the measured flux from MGRO J1908+06at a few TeV can be 15–20%, and can not account for theentire observed disagreement. A further interesting possi-bility is the variability of the source, since the HESS dataconsist of 27 h of observations in 2005–2007, before ourmeasurements.

3.4. MGRO J2031+41

The gamma ray source MGRO J2031+41 was detectedby MILAGRO (Abdo et al., 2007) at a median energy of�20 TeV, is spatially consistent with the source TeVJ2032+4130 discovered by the HEGRA collaboration(Aharonian et al., 2002; Aharonian et al., 2005) and likelyassociated with the Fermi pulsar 1FGL J2032.2+4127(Abdo et al., 2010b). The extension measured by MILAG-RO, 3.0� ± 0.9�, is much larger than that initially estimatedby HEGRA (about 0.1�).

A preliminary analysis of the ARGO-YBJ data shows aTeV emission from a position consistent with MGROJ2031+41 and TeV J2032+4130, with a significance of6.4 s.d. The intrinsic extension of this emission results tobe about 0.2�, a value consistent with the estimation byHEGRA (and MAGIC Albert et al., 2008b). Assumingrext = 0.1 the integral flux above 1 TeV is about 30% thatof the Crab Nebula, which is much higher than the fluxof TEV J2032+4130 as determined by HEGRA (5%) andMAGIC (3%).

The reason for the large discrepancy between the fluxesmeasured by Cherenkov telescopes and ARGO-YBJ is stillunclear. Even considering a possible contribution from adiffuse emission spreading in the Cygnus region, it is diffi-cult to explain this discrepancy.

3.5. Gamma ray bursts

Working in scaler mode (Aielli et al., 2008) ARGO-YBJhas performed a search for high energy emission fromGRBs collecting data from November 2004 (correspondingto the Swift satellite launch) to April 2011, with a detectoractive area increasing from �700 to �6700 m2. During thisperiod, a total of 131 GRBs was inside the ARGO-YBJfield of view (i.e. with zenith angle h 6 45�, limited onlyby atmospheric absorption); for 110 of these ARGO-YBJdata were available and they have been investigated bysearching for a significant excess in the counting rates coin-cident with the satellite detection. In order to extract themaximum information from the data, two GRB analyseshave been implemented:

� search for a signal from every single GRB;� search for a signal from the pile-up of all GRBs (stacked

analysis).


In the search for GeV gamma rays in coincidence withthe low energy GRBs detected by satellites, no evidenceof emission was found for any event. The stacked search,both in time and phase, has shown no deviation from thestatistical expectations, therefore excluding any integraleffect.

The fluence upper limits obtained in the 1–100 GeVenergy range depend on the zenith angle, time durationand spectral index, reaching values down to 10�5 erg cm�2.These values greatly depend on the energy range of the cal-culation. If we consider our sensitivity in terms of expectednumber of positive detections, our estimate gives a ratebetween 0.2 and 1 per year (Aielli et al., 2008), which iscomparable with similar evaluations for other experimentsworking in different energy regions (e.g. Albert et al., 2007).Finally, the capability of the detector in shower mode tomeasure the arrival direction and energy of individualshowers above a few hundred GeV allows the ARGO-YBJ experiment to study the GRBs in the whole 1 GeV–1 TeV range. A detailed discussion on the methods andthe first published results can be found in (Aielli et al.,2009c).

4. Cosmic Ray Physics

4.1. Medium and Large scale anisotropies

The intensity of cosmic rays at TeV energies is nearlyisotropic due to the Galactic magnetic fields which ran-domize and scramble their arrival directions. However,extensive observations show that there exist a slightanisotropy of �10�3 (Guillian et al., 2007). In additionthe relative intensity map of the cosmic ray flux exhibitssome localized regions of significant excess and deficit(Abdo et al., 2008). The search for small-scale structuresis usually carried out through two steps. The first step is

Fig. 10. Medium scale anisotropy of CRs at energies �2 TeV in equatorial coogives the statistical significance in standard deviations.

the creation of a reference map showing how the skywould appear if the cosmic ray flux was isotropic. Thenthe reference map is subtracted from the distribution ofthe actual arrival directions to find the deviations fromisotropy. The present analysis has been performed averag-ing the data in Right Ascension (R.A.) over an angularscale of about 45�. This has been accomplished by meansof both the direct integration and time swapping methods(Fleysher et al., 2004), finding no differences in the refer-ence maps within 0.3 s.d. These methods naturally com-pensate for non-uniform exposure to different regions ofsky, including the presence of gaps in the detector operat-ing time.

Fig. 10 shows the sky map for events with N > 25 andzenith angle h < 40� (corresponding to a proton medianenergy �2 TeV), obtained with a 5� smoothing radius.The map clearly shows two large hot spots in the directionof the Galactic anticenter. The two excesses (>12 s.d., cor-responding to a flux increase of �0.1%) are observed byARGO-YBJ around the positions a � 120�, d � 40� anda � 60�, d � �5�, in agreement with a similar detectionreported by the MILAGRO Collaboration (Abdo et al.,2008b). However, the maximum of the second excess isslightly shifted towards lower declinations, probablybecause ARGO-YBJ can observe the southern regions ofthe sky with increased efficiency, being located at a lowerlatitude than MILAGRO. The deficit regions parallel tothe excesses are due to a known effect of the analysis, thatuses also the excess events to evaluate the background, arti-ficially increasing it. The origin of this medium-scaleanisotropy is puzzling. In fact, these regions have beeninterpreted as excesses of hadronic CRs, but TeV CRsare expected to be randomized by the magnetic fields.Understanding these anisotropies should be a high priorityas they could be due to a nearby source of CRs, as sug-gested by some authors (e.g., Malkov et al., 2010).

rdinates (the origin of the R.A. scale is on the right side). The color scale


The anisotropy over larger angular scales can beobtained exploiting the same methods. Averaging over allR.A. directions the large scale sidereal cosmic rayanisotropy can be measured. For this analysis the equize-nith-angle method, which can efficiently compensate forenvironmental variations such as changes in pressure andtemperature, has been implemented (Amenomori et al.,2005). With 3 years of data we carried out a 2D measure-ment to investigate the large scale CR anisotropy beyondthe simple R.A. profiles. The so-called ‘tail-in’ and ‘loss-

cone’ regions, correlated to an enhancement or deficit ofCRs, are clearly visible with a significance of about 20 s.d.(Fig. 11). The study of the large scale anisotropy is a usefultool in probing the magnetic field structure in our interstel-lar neighborhood as well as the distribution of sources. Themedium-scale anisotropy of Fig. 10 is contained in the twohot spots in the large-scale anisotropy tail-in region. A newexcess component with a�0.1% increase of the CR intensityin the Cygnus region is observed with a significance ofabout 10 s.d. To quantify the scale of the anisotropy we fit-ted the 1D R.A. projection of the 2D map with the first twoharmonics for three different energies. The preliminaryresults (E = 0.7 TeV, A1 = (3.6 ± 0.1) � 10�4, /1 =63.4� ± 0.9; E = 1.5 TeV, A1 = (6.8 ± 0.1) � 10�4, /1 =41.0� ± 0.7; E = 3.9 TeV, A1 = (9.0 ± 0.1) � 10�4,/1 = 35.3� ± 0.6) are in good agreement with the findingsof other experiments (Guillian et al., 2007).

4.2. Measurement of the light component spectrum of CRs

Protons and Helium nuclei are the bulk of the cosmicrays at energies below the knee (�3 � 1015 eV). Recent pre-cise measurements carried out by the long duration flightsof the balloon-borne CREAM experiment show that theenergy spectra of this light component from 2.5 TeV to

Fig. 11. Large scale CR anisotropy observed by ARGO-YBJ at energies �2 Tlower plot the statistical significance in standard deviations.

250 TeV are flatter than the spectra measured at lowerenergies (Yoon et al., 2011). Since the evolution of the pro-ton and Helium energy spectra and their subtle differencescan be an indication of different population of cosmic raysources or acceleration sites, measurements performed bydifferent techniques can be useful as cross-check of resultsconcerning absolute fluxes and spectral indices. TheARGO-YBJ experiment is able to overlap direct measure-ments in a wide region below 100 TeV, not accessible byother extensive air shower experiments.

The differential energy spectrum of the primary CR lightcomponent (p+He) measured by ARGO-YBJ (filled trian-gles) in the energy region 5–200 TeV by using a Bayesianunfolding approach is compared with the results of otherexperiments in Fig. 12. We selected showers with zenithangle h < 30� and with reconstructed core position insidea fiducial area 50 � 50 m2 large by applying a selection cri-terion based on the particle density. In the energy rangeinvestigated the contamination of heavier nuclei is foundto be negligible, not exceeding a few percent. TheARGO-YBJ data are fairly consistent with the valuesobtained by adding up the proton and helium fluxes mea-sured by CREAM concerning both the total intensitiesand the spectrum (Ahn et al., 2010). The value of the spec-tral index of the power-law fit representing the ARGO-YBJdata is �2.61 ± 0.04, which should be compared tocp = �2.66 ± 0.02 and cHe = �2.58 ± 0.02 obtained byCREAM. The present analysis does not allow the determi-nation of the individual proton and helium contributionsto the measured flux, however the ARGO-YBJ data aremainly induced by protons since the average energy ofhelium primaries contributing to events with a given multi-plicity is about 1.5–2 times greater than the average protonenergy. This result suggests a possible hardening of the pro-ton spectrum at energies >5 TeV with respect to that

eV. In the upper plot the color scale gives the relative CR intensity, in the

Energy (GeV)10 210 310 410 510 610

1.75

(GeV

)-1

-s-s

r)2

(m2.

75 E×

Flu

x

410

510

p

He

Fig. 12. The differential energy spectrum of the light component (protonand Helium) measured by ARGO-YBJ (filled red triangles) compared withthe proton (open circles) and Helium spectrum (filled circles) measured bythe CREAM experiment (Ahn et al., 2010). The crossed circles representthe sum of the proton and Helium data measured by CREAM (Yoonet al., 2011). The blue dotted line represents the best fit to proton andHelium data quoted by Horandel (Horandel, 2003), and the shaded area isobtained considering the errors on the fit parameters. The dashed linerepresents the JACEE proton and Helium measurements (Asakimoriet al., 1998), the dot-dashed line represents the RUNJOB proton andHelium measurements (Derbina et al., 2005) (the errors on both are notreported). The green cross represents the proton and Helium fluxmeasured by EAS-TOP and MACRO collaborations (Aglietta et al.,2004). The spectra obtained at lower energies by AMS (stars, Aguilaret al., 2002), BESS (squares, Haino et al., 2004) and CAPRICE (invertedtriangles, Boezio et al., 2003) are also shown. (For interpretation of thereferences in colour in this figure legend, the reader is referred to the webversion of this article.)


obtained at lower energies from direct measurements withsatellites and balloon-borne detectors. We emphasize thatfor the first time direct and ground-based measurementsoverlap over a wide energy range thus making possiblethe cross-calibration of the different experimentaltechniques.

Fig. 13. The �p=p flux ratio obtained with the ARGO-YBJ experimentcompared with all the available measurements (Adriani et al., 2010;Stephens, 1985; Achard et al., 2005; Amenomori et al., 2007; Ambrosioet al., 2003) and some theoretical expectations. The solid curve refers to acalculation for a pure secondary production of antiprotons during thecosmic ray propagation in the Galaxy (Donato et al., 2009). The dashedlines refer to a model of primary �p production by antistars (Stephens andGolden, 1987). The confinement of cosmic rays in the Galaxy is assumedto be dependent on the rigidity R as / R�d, and the two curves representthe cases for d = 0.6 (upper) and 0.7 (lower). The dotted line refers to thecontribution of antiprotons from the annihilation of a heavy dark matterparticle (Cirelli et al., 2009). The dot-dashed line shows the calculation forsecondary antiprotons including an additional �p component produced andaccelerated at cosmic ray sources (Blasi and Serpico, 2009).

4.3. Measurement of the �p=p ratio at TeV energies

In order to measure the �p=p ratio at TeV energies weexploit the Earth–Moon system as an ion spectrometer: ifprotons are deflected towards East, antiprotons aredeflected towards West. If the energy is low enough andthe angular resolution is adequate we could distinguish,in principle, two shadows, one shifted towards West dueto the protons and the other shifted towards East due tothe antiprotons. If no event deficit is observed on the anti-matter side an upper limit on the antiproton content can becalculated.

We selected 2 multiplicity bins, 40 < N < 100 andN > 100. In the first interval the statistical significance ofthe Moon shadow is 34 s.d., the measured angular resolu-tion is �1� and the median energy is 1.4 TeV. For N > 100

the significance is 55 s.d., the angular resolution is �0.6�and the median energy is 5 TeV. Taking into account thatat these energies protons are about 70% of the total flux,with all the data up to December 2009 we are able to settwo upper limits to the �p=p ratio at the 90% confidencelevel: 5% at 1.4 TeV and 6% at 5 TeV, assuming for theantiprotons the proton spectral index.

The upper limits obtained by ARGO-YBJ are comparedin Fig. 13 to all the available �p=p measurements and tosome theoretical models for antiproton production. Thesolid curves refer to a direct production model. The dashedlines refer to a model of primary �p production by antigal-axies (Stecker and Wolfendale, 1985). The rigidity-depen-dent confinement of cosmic rays in the Galaxy isassumed to be / R�d, being R the rigidity, and the twodashed curves correspond to the cases of d = 0.6 and 0.7.The dotted line refers to a possible contribution from theannihilation of a heavy Dark Matter particle (Cirelliet al., 2009). We note that in the few-TeV range theARGO-YBJ results are the lowest available, useful to con-strain models for antiproton production in antimatterdomains.

4.4. Measurement of the total p–p cross section

To make this measurement, the ARGO-YBJ dataanalysis is based on the study of the shower flux attenua-tion for different zenith angles, i.e. atmospheric depths,and exploits the detector accuracy in reconstructing the

Fig. 15. Total p–p cross section obtained from the rp-air cross sectionmeasured by ARGO-YBJ compared with results at accelerators and byother CR experiments (the corresponding references are given in Aielliet al., 2009d).


shower properties. For fixed primary energy and showerage, such an attenuation is expressed by the absorptionlength, K, connected to the primary (mainly protons)interaction length in the atmosphere by the relation:K = k � kCR, where k depends on the shower developmentin the atmosphere, on its fluctuations and on the detectorresponse. The actual value of k must be evaluated by MCsimulations, and it might in principle depend on the fea-tures of the adopted hadronic interaction model, even ifseveral studies showed that the dependence is small. Forprimary protons, the interaction length is related to thep-air interaction cross-section by: kp(g/cm2) ’ 2.41 � 104/rp-air(mb). Given a primary energy interval, the frequencyof showers as a function of the zenith angle h, for a fixeddistance Xdm between the detector and the shower maxi-mum, is directly related to the distribution of the depthof the shower maximum itself, P(Xmax), with Xmax = h0sech– Xdm, where h0 is the observation vertical depth (=606 g/cm2 at the Yangbajing altitude). It can be shown that forsufficiently large Xmax values, P(Xmax) tends to have a sim-ple exponential falling behaviour with characteristic lengthK. The exponential tail of that distribution can be accessedby selecting showers with the maximum development notfar from the detection level (i.e. by minimizing Xdm) and,obviously, exploring a zenith angle region as wide as possi-ble. With these objectives, the ARGO-YBJ detector loca-tion (that is small atmospheric depth) and its features(full coverage, angular resolution, fine granularity, etc.),which ensure the capability of reconstructing the detectedshowers in a very detailed way, have been exploited (see(Aielli et al., 2009d) for the analysis details).

The measured p-air production cross section as a func-tion of the primary proton energy is reported in Fig. 14.The results found by other experiments an the predictionsof several hadronic interaction models are also shown.

Fig. 14. Proton-air production cross section measured by ARGO-YBJand by other CR experiments. The predictions of two different calculationsbased on the Glauber theory are also shown. References for experimentaldata and theoretical predictions are given in (Aielli et al., 2009d).

The systematics arising from several sources have beentaken into account and evaluated on the basis of a fullMC simulation. In particular, the contribution comingfrom heavy nuclei contained in the primary cosmic raybeam has been estimated by adding a proper helium frac-tion to the proton primary flux in the MC simulation(the contribution of heavier nuclei is negligible).

Finally, Glauber theory has been used to infer the totalproton-proton cross section rpp from the measured proton-air production cross section rp-air. The results are shown inFig. 15. Also shown in the figure are the measurementsmade at accelerators and by other CR experiments startingfrom rp-air. As can be seen, the ARGO-YBJ data lie in anenergy region scarcely explored by accelerator experimentsand permit to better investigate the behaviour of the p–p

total cross section where it starts to significantly increasewith energy. In particular, our result favours the asymp-totic ln2(s) increase of total hadronic cross sections asobtained in (Amsler et al., 2008 and Block and Halzen,2005) from a global analysis of accelerator data. The useof the analog RPC charge readout will allow to extendthe study to collisions with center-of-mass energies up tothe TeV region.

5. Interplanetary magnetic field measurement by Sun shadow

The same shadowing effect observed looking at cosmicrays in the direction of the Moon can be observed in thedirection of the Sun, but the interpretation of this phenom-enology is less straightforward. In fact, the displacement ofthe shadow from the apparent position of the Sun could beexplained by the joint effects of the GMF and of the Solarand Interplanetary Magnetic Fields (SMF and IMF,respectively), whose configuration considerably changeswith the phases of the solar activity cycle. In this regard,understanding the Moon shadow phenomenology is a

Fig. 16. The Sun shadow measured using all the data taken by theARGO-YBJ experiment from July 2006 to October 2009. The maximumsignificance is 44.6 s.d. The step between contour lines is 1 s.d.

Fig. 17. Comparison between two measurements of IMF using differentmethods. The solid curve represents the field component By near the Earthmeasured by the ARGO-YBJ experiment using CR deflection. In theperiod 1 (plot (a)) a clear bisector pattern is observed. Positive signindicates that the field is pointing to the centre of the Sun. An uncertaintyof 1 s.d. is marked by the shaded area. The solid dots represent themeasurements using the OMNI satellite network. In the plot (b), theresults with the 4-sector structure in the period 2 are displayed.


useful tool to disentangle the effect of different magneticfields on the Sun shadow and to perform a measurementof the IMF in a minimum of the solar activity.

The observation using the ARGO-YBJ experiment wasmade just in such a particularly good time window whenthe solar activity stayed at its minimum for an unexpect-edly long time since 2006. At distances greater than 5 solarradii from the sun centre, the IMF is distributed mainly inthe ecliptic plane. Its z-component perpendicular to theecliptic plane deflects cosmic rays in the east-west direction,therefore it drives the shadow with an extra shift in addi-tion to the GMF effect which constantly moves the sunshadow towards west as observed in the moon shadowmeasurement (see Fig. 3). Its y-component, By, in the eclip-tic plane defined to be perpendicular to the line of sight,deflects cosmic rays and thus drives the sun shadow inthe north-south direction. It has no contamination fromthe GMF effect because the declination angle of theGMF is less than 0.5� at the ARGO-YBJ site.

The Sun shadow measured using all data taken by theARGO-YBJ experiment from July 2006 to October 2009is shown in Fig. 16. A detailed analysis has been carriedout to study the shift of the center of the Sun shadow alongthe North–South direction as a function of the synodic

Fig. 18. Showers imaged by the ARGO-YBJ central carpet with thedigital (left plot: the space-time structure of a low energy shower) and theanalog (right plot: the core of a high energy shower) readouts.


Carrington longitude. The transverse component of theIMF, By, is then estimated with a minimal assumption onthe model describing the IMF.

The By oscillates following a bi-sector or a four-sectorpattern (see Fig. 17) in two periods of data taking (period1: from January 2008 to April 2009; period 2: all the othermonths from July 2006 to October 2009). These patternsdepend on the direction of the IMF periods that causesthe deflection of the particle trajectories. The positive signindicates that the field is pointing to the centre. The soliddots in Fig. 17 represent the measurements by the orbitingdetectors (the corresponding data are available at http://omniweb.gsfc.nasa.gov). A good agreement is found afterapplying to the ARGO-YBJ data a phase shift of 21� cor-responding to 1.6 days ahead. The two measurements areof the same order in amplitude (2.0 ± 0.2 nT) and are con-sistent in the alternating periodical pattern. Since relativis-tic CRs fly from the Sun to the Earth in only 8 min, theyare able to take an “instantaneous” picture of the IMFpattern. On the contrary, the same information is trans-ported by the solar wind at an average speed of about400 km s�1 (Aielli et al., 2011).

According to this analysis, a large detector able to mea-sure the Sun shadow with the required accuracy within oneday or less, could foresee IMF disturbances which willsweep the Earth about 2 days later. Future detectors suchas the LHAASO project (Cao et al., 2010) may be sufficientto carry out this measurement.

6. Conclusions

The ARGO-YBJ detector is exploiting the performanceof the RPCs to image the front of atmospheric showerswith unprecedented resolution and detail, as shown inFig. 18. The digital and analog readout will allow a deepstudy of the CR phenomenology in the wide range fromTeVs to PeVs. The results obtained in the low energy rangeafter 3 years of data taking predict an excellent capabilityto address a wide range of important issues in AstroparticlePhysics.

References

Abdo, A.A., Allen, B., Berley, D., et al. TeV gamma-ray sources from asurvey of the galactic plane with MILAGRO. ApJ 664, L91–L94,2007.

Abdo, A.A., Allen, B., Aune, T., et al. A measurement of the spatialdistribution of diffuse TeV gamma-ray emission from the galacticplane with MILAGRO. ApJL 688, 1078–1083, 2008a.

Abdo, A.A., Allen, B., Aune, T., et al. Discovery of localized regions ofexcess 10-TeV cosmic rays. Phys. Rev. Lett. 101, 221101-1–221101-5,2008b.

Abdo, A.A., Allen, B., Aune, T., et al. MILAGRO observations of multi-TeV emission from galactic sources in the Fermi bright source list. ApJ700, L127–L131, 2009.

Abdo, A.A., Ackermann, M., Ajello, M., et al. Fermi large area telescopefirst source catalog. ApJS 188, 405–436, 2010a.

Abdo, A.A., Ackermann, M., Ajello, M., et al. The first fermi large areatelescope catalog of gamma-ray pulsars. ApJS 187, 460–494, 2010b.

Achard, P., Adriani, O., Aguilar-Benitez, M., et al. Measurement of theshadowing of high-energy cosmic rays by the Moon: A search for TeV-energy antiprotons. Astopart. Phys. 23, 411–434, 2005.

Adriani, O., Barbarino, G.C., Bazilevskaya, G.A., et al. PAMELA resultson the cosmic-ray antiproton flux from 60 MeV to 180 GeV in kineticenergy. Phys. Rev. Lett. 105, 121101-1–121101-5, 2010.

Aglietta, M., Alessandro, B., Antonioli, P., et al. helium and CNO fluxesin the 100 TeV energy region from TeV muons and EAS atmosphericCherenkov light observations of MACRO and EAS-TOP. Astropart.Phys. 21, 223–240, 2004.

Aguilar, M., Alcaraz, J., Allaby, J., et al. The alpha magnetic spectrometer(AMS) on the international space station. Part I: results from the testflight on the space shuttle. Phys. Rep. 366, 331–405, 2002.

Aharonian, F., Akhperjanian, A.G., Beilicke, M., et al. An unidentifiedTeV source in the vicinity of Cygnus OB2. A&A 393, L37–L40, 2002.

Aharonian, F., Akhperjanian, A.G., Beilicke, M., et al. The unidentifiedTeV source (TeV J2032+4130) and surrounding field: Final HEGRAIACT-system results. A&A 431, 197–202, 2005.

Aharonian, F., Akhperjanian, A.G., Bazer-Bachi, A.R., et al. Observa-tions of the Crab Nebula with HESS. A&A 457, 899–915, 2006.

Aharonian, F., Akhperjanian, A.G., Anton, G., et al. Detection of veryhigh energy radiation from HESS J1908+063 confirms the MILAGROunidentified source MGRO J1908+06. A&A 499, 723–728, 2009.

Ahn, H.S., Allison, P., Bagliesi, M.G., et al. Discrepant hardeningobserved in cosmic-ray elemental spectra. ApJ 714, L89–L93, 2010.

Aielli, G., Assiro, R., Bacci, C., et al. Layout and performance of RPCsused in the ARGO-YBJ experiment. NIM A 562, 92–96, 2006.

Aielli, G., Bacci, C., Barone, F., et al. Scaler mode technique for theARGO-YBJ detector. Astropart. Phys. 30, 85–95, 2008.

Aielli, G., Bacci, C., Bartoli, B., et al. Temperature effect on RPCperformance in the ARGO-YBJ experiment. NIM A 608, 246–250,2009a.

Aielli, G., Bacci, C., Bartoli, B., et al. Software timing calibration of theARGO-YBJ detector. Astropart. Phys. 30, 287–292, 2009b.

Aielli, G., Bacci, C., Barone, F., et al. Search for gamma ray bursts withthe ARGO-YBJ detector in scaler mode. ApJ 699, 1281–1287, 2009c.

Aielli, G., Bacci, C., Bartoli, B., et al. Proton-air cross section measure-ment with the ARGO-YBJ cosmic ray experiment. Phys. Rev. D 80,092004-1–092004-14, 2009d.

Aielli, G., Bacci, C., Bartoli, B., et al. Gamma-ray flares from Mrk421 in2008 observed with the ARGO-YBJ detector. ApJ 714, L208–L212,2010.

Aielli, G., Bacci, C., Bartoli, B., et al. Mean interplanetary magnetic fieldmeasurement using the ARGO-YBJ experiment. ApJ 729, 113–116,2011.

Albert, J., Aliu, E., Anderhub, H., et al. MAGIC upper limits on the veryhigh energy emission from gamma-ray bursts. ApJ 667, 358–366, 2007.

Albert, J., Aliu, E., Anderhub, H., et al. VHE c-ray observation of theCrab Nebula and its pulsar with the MAGIC telescope. ApJ 674,1037–1055, 2008a.

Albert, J., Aliu, E., Anderhub, H., et al. MAGIC observations of theunidentified source c-ray TeV J2032+4130. ApJ 675, L25–L28, 2008b.

Ambrosio, M., Antolini, R., Baldini, A., et al. Moon and Sun shadowingeffect in the MACRO detector. Astropart. Phys. 20, 145–156, 2003.

Amenomori, M., Ayabe, S., Chen, D., et al. A northern sky survey forsteady tera-electron volt gamma-ray point sources using the tibet airshower array. ApJ 633, 1005–1012, 2005.

Amenomori, M., Ayabe, S., Bi, X.J., et al. Moon shadow by cosmic raysunder the influence of geomagnetic field and search for antiprotons atmulti-TeV energies. Astropart. Phys. 28, 137–142, 2007.

Amsler, C., Doser, M., Antonelli, M., et al. Review of particle physics.Phys. Lett. B 667, 1–1340, 2008.

Asakimori, K., Burnett, T.H., Cherry, M.L., et al. Cosmic-ray proton andhelium spectra: Results from the JACEE experiment. ApJ 502, 278–283, 1998.

Bartoli, B., Bernardini, P., Bi, X.J., et al. Observation of the cosmic rayMoon shadowing effect with the ARGO-YBJ experiment. Phys. Rev.D 84, 022003-1–022003-15, 2011.

http://omniweb.gsfc.nasa.gov

http://omniweb.gsfc.nasa.gov


Bartoli, B., Bernardini, P., Bi, X.J., et al. Long-term monitoring of theTeV emission from Mrk421 with the ARGO-YBJ experiment. ApJ734, 110–117, 2011.

Blasi, P., Serpico, P.D. High-energy antiprotons from old supernovaremnants. Phys. Rev. Lett. 103, 081103-1–081103-4, 2009.

Block, M.M., Halzen, F. New evidence for the saturation of the Froissartbound. Phys. Rev. D72, 036006-1–036006-10, 2005.

Boezio, M., Bonvicini, V., Schiavon, P. The cosmic-ray proton and heliumspectra measured with the CAPRICE98 balloon experiment. Astro-part. Phys. 19, 583–604, 2003.

Cao, Z., Bi, X.J., Chan, J.F., et al. A future project in tibet: the large highaltitude air shower observatory (LHAASO). Chin. Phys. C 34, 249–252, 2010.

Cirelli, M., Kadastik, M., Raidal, M., Strumia, A. Model-independentimplications of the e±, �p cosmic ray spectra on properties of DarkMatter. Nucl. Phys. B 813, 1–21, 2009.

D’Ettorre Piazzoli, B. Highlights from the ARGO-YBJ Experiment.Presented at the 32nd ICRC, 2011.

Derbina, V.A., Galkin, V.I., Hareyama, M., et al. Cosmic-ray spectra andcomposition in the energy range of 10–1000 TeV per particle obtainedby the RUNJOB experiment. ApJ 628, L41–L44, 2005.

Di Sciascio, G., Rossi, E. Measurement of the angular resolution of theARGO-YBJ detector. Proceedings of the 30th ICRC 4, 123–126, 2007<arXiv:0710.1945> , 2007.

Donato, F., Maurin, D., Brun, P., et al. Constraints on WIMP darkmatter from the high energy PAMELA �p=p data. Phys. Rev. Lett. 102,071301-1–071301-4, 2009.

Edelson, R.A., Krolik, J.H. The discrete correlation function – A newmethod for analyzing unevenly sampled variability data. ApJ 333, 646–659, 1988.

Fleysher, R., Fleysher, L., Nemethy, P., et al. Tests of statisticalsignificance and background estimation in gamma-ray air showerexperiments. ApJ 603, 355–362, 2004.

Guillian, G., Hosaka, J., Ishihara, K., et al. Observation of the anisotropyof 10 TeV primary cosmic ray nuclei flux with the Super-Kamiokande-I detector. Phys. Rev. D 75, 062003-1–062003-17, 2007.

Haino, S., Sanuki, T., Abe, K., et al. Measurements of primary andatmospheric cosmic-ray spectra with the BESS-TeV spectrometer.Phys. Lett. B 594, 35–46, 2004.

Horandel, J.R. On the knee in the energy spectrum of cosmic rays.Astropart. Phys. 19, 193–220, 2003.

Krennrich, F., Bond, I.H., Bradbury, S.M., et al. Discovery of spectralvariability of Markarian 421 at TeV energies. ApJ 575, L9–L13, 2002.

Malkov, M.A., Diamond, P.H., Drury, L.O’C., Sagdeev, R.Z. Probingnearby cosmic-ray accelerators and interstellar medium turbulencewith MILAGRO hot spots. ApJ 721, 750–761, and references therein,2010.

Porter, T.A., Moskalenko, I.V., Strong, A.W., et al. Inverse Comptonorigin of the hard X-ray and soft gamma-ray emission from thegalactic ridge. ApJ 682, 400–407, 2008.

Smith, A.J., A survey of fermi catalog sources using data from theMILAGRO gamma-ray observatory. In: 2009 Fermi Symposium andeConference Proceedings C091122. <arXiv:1001.3695>.

Stecker, F.W., Wolfendale, A.W. Antiparticles in the extragalactic cosmicradiation. Proceedings of the 19th ICRC 2, 354–357, 1985.

Stephens, S.A. Antiproton upper limits from the observed charge ratio ofmuons and their implications for cosmic ray models. A&A 149, 1–6,1985.

Stephens, S.A., Golden, R.L. The role of antiprotons in cosmic rayphysics. Space Sci. Rev. 46, 31–91, 1987.

Strong, A.W., Moskalenko, I.V., Reimer, O. Diffuse continuum gammarays from the galaxy. ApJ 537, 763–784, 2000.

Yoon, Y.S., Ahn, H.S., Allison, P.S., et al. Cosmic-ray proton and heliumspectra from the first CREAM flight. ApJ 728, 122–129, 2011.

physics results with the argo-ybj experiment

Documents