- Email: [email protected]
Part I: General description of the experiment
NUCLEAR INSTRUMENTS AND METHODS
32 (1965) 4549 ; (D NORTH-HOLLAND PUBLISHING CO.
SHIELDING STUDIES IN STEEL WITH 10 AND 20 GeV/c PROTONS The next six articles proceeded from experimental studips, in collaboration of the fol-lowing laboratories : ABS (Hannover) and DESY (Hamburg)
H . Bindewald, C.
CERN (Geneva)
J. 1aarli, A. Citron, J. Geibel, K. Goebel, L. Hoffmann, J. Ranft, K H. Reich and A. H. Sullivan.
ORNL (Oak Ridge)
R. L. Childers and C. D. Zerby.
RUTHERFORD LAB. (Chilton)
C. M . Fisher, R. H. Thomas and McEwen .
SLAC (Stanford)
W. R. Nelson and M. Whitehead .
)w,
J. Seetzen and H. Schultz .
PART 1 : GENERAL DESCRIPTION OF THE EXPERIMENT J. GEIBEL and K. H. REICH CERN, Geneva and J. SEETZEN* DESY, Hamburg Received 2 July 1964
The longitudinal and for the lateral development of the nuclear cascade was studied 10 and 19.2 GeV/c protons incident on a steel absorber . The particle flux disti~ibution perpendicular to the beam direction was measured at several depths using nuclear emulsions, ionization chambers and carbon- I I-activation as detectors. The build-up and the subsequent attenuation in the absorber are given along the beam axis. as well as for the total flux obtained by integration over the lateral distributions .
Similar measurements were made on the radiation emitted from a Be-target at different angles betwwn 20' and 40* for incident protons of 10 and 19-2 GeV/c. The fiux density of the emitted radiation was found to follow a power law dependent on the emission angle 0 in the interval of 10" to 90" . Monte Carlo shielding calculations are compared wit] : the results of the cascade experiments .
1. Introducthm The shielding design for the multi-GeV proton synchrotrons at Brookhaven and CERN was based on the knowledge and experience gained in cosmic ray work and with lower energy accelerators. It was found to be adequate in both cases. The state of the art of shielding design at the ti-me wher, the big accelerators became operational was reviewed by S. J. Lindenbaum) . Further studies at the newly available energies were desirable in order to establish a firm expt-,rimental basis for the shielding problems arising with new projects e.g. extracted beams, higher -. ntensitie,; and higher
energies. A series of experiments was performed at the CERN proton synchrotron with .he collaboration of several other laboratories. In thefirst two experiments proton beams of relatively large cross-sections incident on heavy concrete were used 2 ). They gave information concerning the development of the nucleonic cascade in the forward direction.. The experiment reported in th;s paper was designed to study i) the three dimensional development ofthe nucleonic cascade induced in steel by incident pencil beams of 10 and 19.2 GeV/c protons, referred to subsequently as the "'cascade experiment" : ii) the shielding requirements for secondary particles
* Now at: Kemforschungszentrum Karlsruhe, Germany.
45
J. GEIBEL , et
al.
i, Eicam and expcrimental layout. M bending magneti; Q quadrupole magnets ; CI, C2 chamber; T beryllium target .
emer&g under large angle (20*AW) from 4 target ruc!x by 19.2 GeV/c protons, referred to subsequently w; the "side shielding experiment," . The detailed qualitative results ofthe diflerent phases. of tht.- experiment will be given in the following five parts of this paper. In this firstpart the experimental layout is described and, - the results are summarized qualilatively. The fallowing three parts deal with the cascale experim- ent. In part 11 the results of emulsion exposures III with 19.2 GeV/c incident protons- are given, in pairt the same, for 10 GeV/c incident momentum
and in part IV the results obtained with carbon- I I-
activated plastic scintillators and ionisation chambers for bath energies. In part V the, results of the side shielding experiment are summarised. In part VI results of the cascade experiment are compared with MonteCarlo shielding calculations . 2. Expcrimental layout 2 .1 . &iAJA AND MONITORING he internal proton 'beam of the synchrotron hit a beryRium target from which protons were elastically scattered by about 20 milliradians in the horizontal plane. These went through the fringing magnetic field of the machine which produced a vertical focus at a distance of 25 rn from the target. The beam defined in this way was bent into the South Experimental Hall by
Fig. 2.
collimators ; CHI monitor ionintion
three bending magnets M 1, M2 and M3 (fig' 1). Bending through a total angle of 120 milliradians, 40 milliradians per magnet, helped to clean the beam from neutrons. Four quadrupole magnets Q, to Q4 formed an imageof the target inside the 0.5 x 0.5 cm1 aperture of the collimator CI A second collimator C,1 with an aperture ,.-)f 5 x 5 cm, I placed just behind the last quadrupole Q4 confined the divergence ofthe beam pawing through C!, to ± 2.1 milliradians. For the side shielding exposures, a beryllium target representing a tenth of a total interaction length (3 cm) was placed at point T in the proton beam. The beam was monitored by an ionisation chamber CH I, situated in between the collimators C, and C2 and by a coinn.-idence counter telescope located behind col,. limator C I (fig. 2). The performance of the counter telescope was checked with emulsions. The agreement was within 10%. The time variation of the beam intensity which is important for the activation measurements, was monitored by another io,,,aisation chambei CH 2 near the absorber for the cascade experiment . The intensity of the beam for the cascade experiment as defined by the collimators C, and C2 was 2 x 101 protons per machine pulse of 3 x 10" protons at 19.2 GeV/c and 101 protons per pulse at 10 GeV/c. Thisi represented 0.7% of the beam intensity in front
etailed experimental layout.
SHIELDING STUDIES IN STEEL -PART I
of the collimator C, as measured by the ionisation chamber. The total number of protons incident on the absorber was about IWO protons at 19.2 GeV/c and 2.2 x W protons at 10 GeVIc. The neutron contamination of the beam behind the collimator C, was about 1(r/. at 19.2 GeV/c. (See part 11). 3. Absorber a The arrangement'of the absorbers is shown in detail in fig. 2. Different absorber stackswere employed for the cascade and the side shielding experiment, respectively. Both stacks were composed of steel slabs with the dimensions 160 x 100 x 10 cml. Slots of a width of 3 cm were left between certain steel slabs for placing detectors as shown in the figure. The total steel thickness of the (main) stack servicing for the cascade experiment was 3.20 rn (excluding the slots). The beam hit this stack at the center of the first slab atA distance of 2.9 m from the collimator C 1 . The cross section of the beam at impact was about 1.2 x 1 .2 cml. The stack for the side shielding experiment consisted of 4 slabs. The front surface of the first slab was aligned parallel to the beam axis at a distance of 25.5 cm. Secondary particles produced in the target T at angles between 10' and 60' hit this front slab. Heavy concrete shielding around the main stack and upstream of the side shielding stack completed the layout. 4. Detedors The detcctors were placed in front, inside and behind the absorbers. Stacks of a few Ilford G5-cmulsion plates were used to measure track and star densities . Particle Jux densities were measured by carbon-I I activation in plastic scintillators and radiation doses with small ionisation chambers . 5. QuMitative summary of the remft For quantitative results see the following parts of this paper . Qualitatively the reSults of the cascade experiment confirmed and clarified a conclusion suggested by experiments at lower energies : The cascade propagates through the absorber due to a central "'hard core" of high energy strongly interacting particles . This "hard core" spreads slowly with increasing depth at a rate to be expected from elastic nuclear and multiple Coulomb scattering ofthe primary protons and the strong forward peaking of high energy -=ondaries originating in ripheral collisions') . The "hard core" is surrounded by a cloud of particles of lower energy . These are the products of inte,,-actions of hard core particles emitted at larger angles which, in turn, might create a few more
47
generations of secondaries with ~apidly decreasing energies. The build up factors which describe the initial increase in particle flux due to the multiple secondaries production were measured . These factors show the magnitude and energy dependence to be expected from the known multiplicities of secondaries hi single interactions . The same holds for the depths of the transition region, i.e. the depth at which the flux after build up has fallen to the flux of the incident beam. However a comparison of the measured minimum track density and the track density to be expected from the measured star density suggests that a large fraction of the tracks is due to electrons of the electromagnetic cascade following n-decay. Muons originating in the cascade itself do not show up down to the attenuation of 10-1 reached in this experiment . The fraction of neutron induced interactions of more than 100 MeV incident energy increases steadily with increasing absorber depth. At large depths of the absorber the fluxes attenuate nearly exponentially ; the attenuation length for the integrated flu;~ was found to be almost the same at 10 and 19.2 GeV/c incident proton momentum . It was about 30% larger than the nuclear mean free path for inelastic collisions . If the "hard core" were identical with the primary protons, one would expect to find this attenuation length equal to the nuclear mean free path, as it is the case for a few ,--XeV/c incident proton momenta . The side shielding experiment confirmed that shielding walls paral!el to the primary beam direction can be thin compared to shields normal to it. This was expected as a consequence of the strong forward peaking of secondaries from peripheral high energy interactions and the rapid decrease of their energy with increasing production angle. We are grateful to the "Bundesministerium fUr Wissenschaftliche Forschung" for considerable financial aid which made th,! experiment possible. We are indebted to Mr. H. Schultz for invaluable assistance given during the preparation and running of the experiment . The help of Mr. McEwan in designing and setting up the beam was much appreciated . Thanks are due to the Proton Synchrotron Machine Division for their efficient collaboration . Reference,, 1) S. J. Lindenbaurn, Ann. Rev. Nucl . Science, 11 (19A) 21~. 2) A . Citron, L. Holminann, C. Passow, Nucl. Instr, and Meth., 14 (1961) 97. 1) G. Cocconi, L . 1. Koester, D. H . Perkins, Berkeley High Energy Study, UCR L 10022, (1961) 167.
32 (1965) 4549 ; (D NORTH-HOLLAND PUBLISHING CO.
SHIELDING STUDIES IN STEEL WITH 10 AND 20 GeV/c PROTONS The next six articles proceeded from experimental studips, in collaboration of the fol-lowing laboratories : ABS (Hannover) and DESY (Hamburg)
H . Bindewald, C.
CERN (Geneva)
J. 1aarli, A. Citron, J. Geibel, K. Goebel, L. Hoffmann, J. Ranft, K H. Reich and A. H. Sullivan.
ORNL (Oak Ridge)
R. L. Childers and C. D. Zerby.
RUTHERFORD LAB. (Chilton)
C. M . Fisher, R. H. Thomas and McEwen .
SLAC (Stanford)
W. R. Nelson and M. Whitehead .
)w,
J. Seetzen and H. Schultz .
PART 1 : GENERAL DESCRIPTION OF THE EXPERIMENT J. GEIBEL and K. H. REICH CERN, Geneva and J. SEETZEN* DESY, Hamburg Received 2 July 1964
The longitudinal and for the lateral development of the nuclear cascade was studied 10 and 19.2 GeV/c protons incident on a steel absorber . The particle flux disti~ibution perpendicular to the beam direction was measured at several depths using nuclear emulsions, ionization chambers and carbon- I I-activation as detectors. The build-up and the subsequent attenuation in the absorber are given along the beam axis. as well as for the total flux obtained by integration over the lateral distributions .
Similar measurements were made on the radiation emitted from a Be-target at different angles betwwn 20' and 40* for incident protons of 10 and 19-2 GeV/c. The fiux density of the emitted radiation was found to follow a power law dependent on the emission angle 0 in the interval of 10" to 90" . Monte Carlo shielding calculations are compared wit] : the results of the cascade experiments .
1. Introducthm The shielding design for the multi-GeV proton synchrotrons at Brookhaven and CERN was based on the knowledge and experience gained in cosmic ray work and with lower energy accelerators. It was found to be adequate in both cases. The state of the art of shielding design at the ti-me wher, the big accelerators became operational was reviewed by S. J. Lindenbaum) . Further studies at the newly available energies were desirable in order to establish a firm expt-,rimental basis for the shielding problems arising with new projects e.g. extracted beams, higher -. ntensitie,; and higher
energies. A series of experiments was performed at the CERN proton synchrotron with .he collaboration of several other laboratories. In thefirst two experiments proton beams of relatively large cross-sections incident on heavy concrete were used 2 ). They gave information concerning the development of the nucleonic cascade in the forward direction.. The experiment reported in th;s paper was designed to study i) the three dimensional development ofthe nucleonic cascade induced in steel by incident pencil beams of 10 and 19.2 GeV/c protons, referred to subsequently as the "'cascade experiment" : ii) the shielding requirements for secondary particles
* Now at: Kemforschungszentrum Karlsruhe, Germany.
45
J. GEIBEL , et
al.
i, Eicam and expcrimental layout. M bending magneti; Q quadrupole magnets ; CI, C2 chamber; T beryllium target .
emer&g under large angle (20*AW) from 4 target ruc!x by 19.2 GeV/c protons, referred to subsequently w; the "side shielding experiment," . The detailed qualitative results ofthe diflerent phases. of tht.- experiment will be given in the following five parts of this paper. In this firstpart the experimental layout is described and, - the results are summarized qualilatively. The fallowing three parts deal with the cascale experim- ent. In part 11 the results of emulsion exposures III with 19.2 GeV/c incident protons- are given, in pairt the same, for 10 GeV/c incident momentum
and in part IV the results obtained with carbon- I I-
activated plastic scintillators and ionisation chambers for bath energies. In part V the, results of the side shielding experiment are summarised. In part VI results of the cascade experiment are compared with MonteCarlo shielding calculations . 2. Expcrimental layout 2 .1 . &iAJA AND MONITORING he internal proton 'beam of the synchrotron hit a beryRium target from which protons were elastically scattered by about 20 milliradians in the horizontal plane. These went through the fringing magnetic field of the machine which produced a vertical focus at a distance of 25 rn from the target. The beam defined in this way was bent into the South Experimental Hall by
Fig. 2.
collimators ; CHI monitor ionintion
three bending magnets M 1, M2 and M3 (fig' 1). Bending through a total angle of 120 milliradians, 40 milliradians per magnet, helped to clean the beam from neutrons. Four quadrupole magnets Q, to Q4 formed an imageof the target inside the 0.5 x 0.5 cm1 aperture of the collimator CI A second collimator C,1 with an aperture ,.-)f 5 x 5 cm, I placed just behind the last quadrupole Q4 confined the divergence ofthe beam pawing through C!, to ± 2.1 milliradians. For the side shielding exposures, a beryllium target representing a tenth of a total interaction length (3 cm) was placed at point T in the proton beam. The beam was monitored by an ionisation chamber CH I, situated in between the collimators C, and C2 and by a coinn.-idence counter telescope located behind col,. limator C I (fig. 2). The performance of the counter telescope was checked with emulsions. The agreement was within 10%. The time variation of the beam intensity which is important for the activation measurements, was monitored by another io,,,aisation chambei CH 2 near the absorber for the cascade experiment . The intensity of the beam for the cascade experiment as defined by the collimators C, and C2 was 2 x 101 protons per machine pulse of 3 x 10" protons at 19.2 GeV/c and 101 protons per pulse at 10 GeV/c. Thisi represented 0.7% of the beam intensity in front
etailed experimental layout.
SHIELDING STUDIES IN STEEL -PART I
of the collimator C, as measured by the ionisation chamber. The total number of protons incident on the absorber was about IWO protons at 19.2 GeV/c and 2.2 x W protons at 10 GeVIc. The neutron contamination of the beam behind the collimator C, was about 1(r/. at 19.2 GeV/c. (See part 11). 3. Absorber a The arrangement'of the absorbers is shown in detail in fig. 2. Different absorber stackswere employed for the cascade and the side shielding experiment, respectively. Both stacks were composed of steel slabs with the dimensions 160 x 100 x 10 cml. Slots of a width of 3 cm were left between certain steel slabs for placing detectors as shown in the figure. The total steel thickness of the (main) stack servicing for the cascade experiment was 3.20 rn (excluding the slots). The beam hit this stack at the center of the first slab atA distance of 2.9 m from the collimator C 1 . The cross section of the beam at impact was about 1.2 x 1 .2 cml. The stack for the side shielding experiment consisted of 4 slabs. The front surface of the first slab was aligned parallel to the beam axis at a distance of 25.5 cm. Secondary particles produced in the target T at angles between 10' and 60' hit this front slab. Heavy concrete shielding around the main stack and upstream of the side shielding stack completed the layout. 4. Detedors The detcctors were placed in front, inside and behind the absorbers. Stacks of a few Ilford G5-cmulsion plates were used to measure track and star densities . Particle Jux densities were measured by carbon-I I activation in plastic scintillators and radiation doses with small ionisation chambers . 5. QuMitative summary of the remft For quantitative results see the following parts of this paper . Qualitatively the reSults of the cascade experiment confirmed and clarified a conclusion suggested by experiments at lower energies : The cascade propagates through the absorber due to a central "'hard core" of high energy strongly interacting particles . This "hard core" spreads slowly with increasing depth at a rate to be expected from elastic nuclear and multiple Coulomb scattering ofthe primary protons and the strong forward peaking of high energy -=ondaries originating in ripheral collisions') . The "hard core" is surrounded by a cloud of particles of lower energy . These are the products of inte,,-actions of hard core particles emitted at larger angles which, in turn, might create a few more
47
generations of secondaries with ~apidly decreasing energies. The build up factors which describe the initial increase in particle flux due to the multiple secondaries production were measured . These factors show the magnitude and energy dependence to be expected from the known multiplicities of secondaries hi single interactions . The same holds for the depths of the transition region, i.e. the depth at which the flux after build up has fallen to the flux of the incident beam. However a comparison of the measured minimum track density and the track density to be expected from the measured star density suggests that a large fraction of the tracks is due to electrons of the electromagnetic cascade following n-decay. Muons originating in the cascade itself do not show up down to the attenuation of 10-1 reached in this experiment . The fraction of neutron induced interactions of more than 100 MeV incident energy increases steadily with increasing absorber depth. At large depths of the absorber the fluxes attenuate nearly exponentially ; the attenuation length for the integrated flu;~ was found to be almost the same at 10 and 19.2 GeV/c incident proton momentum . It was about 30% larger than the nuclear mean free path for inelastic collisions . If the "hard core" were identical with the primary protons, one would expect to find this attenuation length equal to the nuclear mean free path, as it is the case for a few ,--XeV/c incident proton momenta . The side shielding experiment confirmed that shielding walls paral!el to the primary beam direction can be thin compared to shields normal to it. This was expected as a consequence of the strong forward peaking of secondaries from peripheral high energy interactions and the rapid decrease of their energy with increasing production angle. We are grateful to the "Bundesministerium fUr Wissenschaftliche Forschung" for considerable financial aid which made th,! experiment possible. We are indebted to Mr. H. Schultz for invaluable assistance given during the preparation and running of the experiment . The help of Mr. McEwan in designing and setting up the beam was much appreciated . Thanks are due to the Proton Synchrotron Machine Division for their efficient collaboration . Reference,, 1) S. J. Lindenbaurn, Ann. Rev. Nucl . Science, 11 (19A) 21~. 2) A . Citron, L. Holminann, C. Passow, Nucl. Instr, and Meth., 14 (1961) 97. 1) G. Cocconi, L . 1. Koester, D. H . Perkins, Berkeley High Energy Study, UCR L 10022, (1961) 167.