What is so exciting about BURSTS?―A historical and phenomenological review

In a preceding publication [1] , some aspects of BURST and Spallation reactions induced by high-energy heavy ions were presented. This contribution will describe some aspects of these reactions in more detail. It is a basic concept of this work [1] [2] that in high-energy heavy ion reactions above a threshold of Ecm/u ~150 MeV two dominant reaction channels are observed: BURST and Spallation. Experimental details have for example been observed with nuclear emulsions and Solid State Nuclear Track Detectors (SSNTD) irradiated with high-energy heavy ions. In emulsion and SSNTD energetic charged particles create tracks which can be studied with optical microscopes. Spallation reactions have been investigated for decades and they are described frequently: At least, one heavy nucleus with charge Z > 1 is observed among the relativistic projectile fragments [1] . A fast nuclear cascade is followed by slow nuclear evaporations from the residual target nucleus.

BURST reactions show exactly ZERO relativistic fragments with Z > 1. Relativistic fragments with Z = 1 are observed as shower prongs in the microscope and their number is called “n_s”. They are due to a complete fragmentation of the projectile or projectile fragment and possibly also of part of the target nucleus into shower prongs. These shower prongs can be due to protons above an energy of 375 MeV or due to charged pions above an energy of 56 MeV. The average number of shower prongs per event in BURSTS is several times larger than in Spallation reactions. Additionally, one observes also tracks from low energy protons with E < 26 MeV, called “black prongs”; their average number per interaction is. One also observes some tracks from medium energy protons with 26 MeV < E < 375 MeV, called “grey prongs”; their average number per interaction is. Finally, one considers “heavy prongs”, n_h, defined as:

BURSTS have been observed for more than 30 years. Obviously, they are due to a very drastic nuclear interaction of high-energy heavy ions, leading to the observation of very many prongs in one specific “star” within the nuclear emulsion. Only charged particles will produce a visible track in the emulsion, however, such an interaction MUST release simultaneously many free neutrons. These neutrons leave no track in the nuclear emulsion but they can induce secondary nuclear interactions yielding all kinds of conventional nuclear reaction products, which can be studied by nuclear chemists with their analytical methods [2] . Some earlier studies of BURSTS using nuclear emulsions and/or Solid State Nuclear Track Detectors (SSNTD) shall be recalled:

・ Khan et al. [3] observed in 1985: “Clusters of light tracks in a stack of CR-39 exposed to 980 MeV/u ²³⁸U”.

・ Ganssauge [4] observed a BURST-interaction produced by 980 MeV/u ²³⁸U in a nuclear emulsion, irradiated at the Bevalac accelerator in Berkeley (California), as shown in Figure 1(a).

・ Tolstov [5] showed a “very drastic” and “peculiar” event. It was produced in a nuclear emulsion irradiated with 72 GeV ²²Ne ions at the Synchrophasotron accelerator in Dubna (Russia) as shown in Figure 1(b). In this reaction, both the projectile and target are completely disintegrating into single nucleons plus pions. No tracks from Z ≥ 2 particles are visible. One should also note the track-free angle of 14 degrees in the forward direction.

BURST-interactions need further experimental investigations and a concept in order to understand this nuclear reaction path better. As we do not understand BURST events, we cannot order experimental results according to model topics but we can only list them. There clearly are unresolved problems in BURST reactions.

1) Fundamental aspects, which could be investigated with nuclear emulsions and SSNTDs are:

・ More accurate angular n_s track-distributions are needed at small angles with respect to the beam direction; they must be measured and should be compared to model calculations. In table 4 of Ref. [1] is shown evidence induced by 72 GeV ²²Ne ions; it shows that 364 events out of the total ensemble of 1011 events (36%) have an open angle in the straight forward direction. Tolstov has observed a similar effect for 19 events (private communication, M.H.).

・ There are no theoretical model calculations known to the authors that can explain the entire BURST-interaction or the special case of the “open gap at zero-degree”. An attempt on BURSTS based on Monte Carlo code MCNPX 2.7 will be presented in this paper.

・ Due to lack of experimental knowledge of the energy carried by a single particle, one cannot attempt at this stage an estimate of the total energy balance in individual BURST interactions.

2) Practical aspects of using BURSTS to solve problems in a specific area of technical applications are:

When one looks into Figure 1(a), one observes a picture of a nearly complete destruction of a uranium nucleus into about 92 tracks, all being from Z = 1 hadrons. Simultaneously, there must have occurred a emission of up to 146 high- energy neutrons. The open question is: can such a process be of use for mankind in any way?

Haiduc et al. [9] have measured further details in 3292 Spallation-events from 72 GeV ²²Ne on nuclear emulsion. They counted the number of black prongs, n_b, and minimum ionizing particles, n_s, for each event and obtained a data-set with a “n_b-versus-n_s” information for all 3292 Spallation events. They reported the actual n_b-value, i.e. the number of black prongs with energy below 26 MeV. They did NOT include grey prongs, n_g, with proton energy [26 MeV < E(n_g) < 375 MeV]. This “n_s-vs.-n_b” data-set is shown for all 3292 Spallation events in Figure 4 in a 3-dimensional representation.

Figure 5 gives another and more detailed picture of the same information; it shows three two-dimensional correlations for different regions of n_s. In the low n_s-region between (0 ≤ n_s ≤ 14) one finds 82.2% of all Spallation events, in the higher n_s-regions correspondingly less. The maximum of events are close to small values of n_b and n_s, or more quantitatively: this maximum is found at 0 ≤ n_b ≤ 5 and 0 ≤ n_s ≤ 15. For Spallation, one observes in this small center

around the origin of the coordinate system approx. 66% of all events.

Haiduc et al. also investigated with the same technique 1011 first generation BURST-events from 72 GeV ²²Ne-interactions in nuclear emulsion and registered for each event the “n_b-vs.-n_s” data-set.

The results are plotted as “n_b-correlated-n_s” data set for BURST events in Figure 6. Figure 7 gives again the more detailed picture of the same information; it shows three two-dimensional correlations for different regions of n_s. In the low n_s-region between (0 ≤ n_s ≤ 14) one finds only 8.3% of all BURST events, in the large n_s-regions (25 ≤ n_s ≤ 60) correspondingly the largest number of 56.0% BURST. These two correlation plots for Spallation and BURST reactions are completely different: For BURSTS one finds in the area 0 ≤ n_b ≤ and 0 ≤ n_s ≤ 15 (around the origin of the coordinate system) only 52 (5%) out of all 1011 events. The other 959 events are found within a wide area of the diagram up to n_b ≤ 20 and n_s ≤ 60.

In short: Spallation reactions yield small values of n_b (low-energy particles) and ns (high-energy particles) whereas in BURST reactions multiplicities of both low- and high-energy particles are significant.

Theoretical studies on the BURST phenomenon are presented in the next section.

Figure 4 shows that the distribution of events in the n_s-vs.-n_b space is essentially centered around n_s = 0 and n_b = 0. This is expected for Spallation reactions where relatively little energy is involved and one does not expect large numbers of particles in the output channel. A more detailed sketch of the same data is shown in Figure 5 where n_s multiplicities are shown as function of n_b for different n_s-regimes. The lower n_s-distribution for n_s £ 15 is strongly peaked around n_b = 0 and it looks like a “Spallation” distribution. In the middle-n_s regime from n_s = 15 to 24 one finds a two-peaked distribution which looks like a mix of Spallation and BURST, whereas in the high-n_s regime one finds most events around larger numbers of n_b. An increasing number of n_s go with an increasing number of n_b.

Figure 6 shows the n_b-vs.-n_s correlation for BURST reactions. Its display is very different from the Spallation distribution in Figure 4 in that most events are distributed to higher values of n_b and n_s, i.e. there is significantly more energy in the system.

Details of the distribution are shown in Figure 7 which is equivalent to Figure 5. In contrast to the Spallation distribution there are only very few events in the small n_s regime whereas the majority of events is found at large numbers of n_s.

It is also interesting to look into another detail of Figure 7 and to study individually the observed n_s-events (i.e. relativistic particles) for BURST interactions having only n_b = 0 (73 events) or n_b = 1 (65 events). The result is shown in Figure 8.

For n_b = 0 one observes a broad distribution of events having n_s = 10 to n_s = 33. In these 73 events one finds a total of 1420 minimum ionizing particles (n_s = 1420) which are associated with zero black prongs (n_b = 0). This happens in 73/1011 = 7.2% of all BURST events.

A very similar situation is found for 65 events having n_b = 1 in which a total of 1388 minimum ionizing particles are measured. In total, the ratio n_b/n_s for these 138 events of 65/2808 = 0.023 is significantly lower than the observed overall ratio [1] of n_b/n_s = 0.28.

Thus, very small numbers of low-energy particles as seen in 138 events (n_b = 0

or 1) are associated with 2808 relativistic particles. In all cases, even at very small n_b, there are at least 10 relativistic particles seen in the exit channel because BURST events are selected. Such distribution of very many high- and very few or even no low-energy particles in the exit channel have not been predicted by the model (see next section). The calculated numbers of low-energy particles per one relativistic particle are typically almost twice as large as experimentally observed for BURST events.

It should be noted that one can clearly see very small numbers of black prongs (n_b = 0 or 1) in an interaction in emulsion and an incorrect assignment of such events can be excluded.

In this section, we report results of an attempt to simulate BURST events using MCNPX 2.7 Monte Carlo code [10] . MCNPX is not designed for this type of calculations, however, due to its flexibilities and availability of many options in the code it is possible to simulate interactions of the relativistic ions with nuclear emulsion event by event.

When using the Monte Carlo method to simulate BURST interactions one has to find the proper and well-defined boundary conditions in order to simulate the observed physical reality. The BURST definition used in this paper and in Ref. [1] does not have such a unique and fixed number of interacting and participating particles. Therefore the authors have to employ a slightly modified definition for BURST interactions in the Monte Carlo calculation with the MCNPX 2.7 code.

Definition: BURSTS are relativistic heavy ion interactions in which the heavy projectile totally fragments into Z = 1 particles and neutrons and no relativistic fragments with Z > 1 are produced

The above-mentioned definition of BURST events implies that:

1) No relativistic projectile-fragment has a charge Z > 1.

2) Charged projectile fragments (Z = 1) can be relativistic and non-relativistic, i.e. they can produce shower and heavy prongs in the emulsion.

3) Target fragments can have any charge less than target nucleus charge (Z < Z_T) and in the emulsion they can result in shower, and heavy prongs.

4) In the case of the Ne + emulsion interactions BURST events (as defined above) must have;

a) Greater than ten Z = 1 prongs, resulting from total fragmentation of the Ne and thus total number of prongs in such events must be n_t ³ 10.

b) The net charge carried by the fragments must be n_q > 10.

c) Number of neutrons released in a BURST interaction must be greater than 12 (number of neutrons in ²²Ne). This condition is an approximation because during the nucleon-nucleon collisions which lead to pion production the type of nucleons may change.

In studies of nuclear reactions, to achieve acceptable statistical uncertainties the Monte Carlo (MC) codes usually run for a large number of histories (incident ions) and the final output of the code is normalized to the number of histories. In high-energy heavy ion interactions, such outputs cannot provide accurate information about the individual interactions. The events recorded in nuclear emulsion are due to individual interactions i.e. one projectile interacts with one target nucleus. Therefore, in order to simulate and understand events recorded in emulsion one must execute the MC-code and recoded interactions, event by event, as is the case in the emulsion experiments.

In this work the MCNPX 2.7 code [10] was used to simulate interactions of 4.1 AGeV/c ²²Ne ions with emulsion. MCNPX transports the incident ion until its energy drops (because of energy loss due to electronic and/or nuclear interactions with the target atoms/nuclei) below a cut-off energy or the ion escapes the target surface after nuclear and/or electronic interactions. An incident ion and its fragments may make more than one nuclear reaction with the target nuclei. In order to simulate the events initiated by the incident ion in emulsion one needs to

・ Isolate individual primary projectile ion interactions.

・ Obtain particle yield and energy spectra of the each type of the product particles in each ion nucleus collision.

The obtained particle yields and spectra must not be contaminated by:

・ The particles produced from interactions of the projectile fragments.

・ The particles produced in interactions of the secondary particles.

Therefore, one wants to record only those particles that are produced in intra-nuclear cascade, pre-equilibrium and evaporation stages of the interactions of the primary ions. These are the particles recorded in the emulsion as prongs of an event. In order to reduce and possibly eliminate the number of secondary particle interactions, the emulsion dimension along the beam line must be reduced. It is shown that for L = 0.5 cm the contribution of secondary particle interaction products is negligible [11] .

It should be noted that in the calculations all charged particles (i.e. protons, deuterons, tritons, ³He, ⁴He and charged pions) are considered to yield measurable prongs regardless of their energies (track length) and total number of prongs present in an event is identical to the number of charged particles in the output channel. This may not be necessarily the case in the experimental measurements [11] where due to track identification and measurement limitations not all tracks may be measureable, in particular in events which are composed of many tracks. Such an underestimation of the track numbers may be more pronounced in the case of the black prongs which have short ranges.

(However, experimentalists are certain that in the measurements the number of tracks can be counted very accurately).

Details of the Monte Carlo simulations of heavy ion interactions with nuclear emulsion using the MCNPX 2.7 code are given elsewhere [11] .

Simulation of BURST events

A total of 10,146 individual inelastic interactions of 4.1 AGeV/c ²²Ne with Ilford-G5 emulsion were recorded. In selecting the BURST events, the following conditions were imposed and only events which satisfied all of the following conditions were selected:

1) Events with Z ≥ 2 shower tracks were removed. The Z = 2 relativistic particles most likely result from projectile fragmentation and appear in 9% of the all events.

2) Events with charge release n_q ≤ 10 were removed. Such events cannot belong to total fragmentation of the projectile (Z = 10).

3) Events with n_q > 57 (total charge available in input channel) were removed. Such events are those in which interactions of the secondary particles were contaminating the results. There were only 9 out 10,146 event that had n_q > 57 (i.e. less than 0.09%), suggesting that contamination of the prong numbers by particles from interactions of secondaries was insignificant.

4) Events with total track number n_t < 10 were removed. In the total fragmentation of the Ne projectile one expects to find at least 10 prongs from the released charged particles plus those resulting from target fragmentation.

5) Events in which the number of released neutrons was less than 12 were removed.

After imposing all above mentioned five selection criteria 3292 events remained for which distributions of the numbers of shower, grey and black prongs per event were produced. It was found that the mean number of shower prongs per event is ~18 as compared to experimental value of ~26 (see Ref. [1] ). Such a disagreement is an indication that the above mentioned five filtering steps were not sufficient for an adequate separation of the BURST from Spallation events. This is mainly due to the fact that in the selection criteria the fragmentation of the target nucleus is not taken into account.

In order to match the mean number of shower prongs per event in simulated events with that of the measurements two different options were investigated:

This paper presents additional information which shall help to come closer to the understanding of “unresolved problems” in heavy ion interactions in thick targets [16] above an energetic threshold of E_CM/u ≈ 150 MeV [2] and to understand BURSTS [1] in the same context.

・ BURST reactions are a major nuclear reaction channel when high-energy projectiles interact with a target. This is especially true for the second generation of interactions in general and in particular when relativistic secondary Z = 2 (helium) nuclei make another interaction.

・ From the fact that the number of low-energy charged particles (n_b) emitted in a BURST reaction is significantly larger than n_b in Spallation reactions one can deduce that there is significantly more energy involved in BURSTS than in Spallation.

The same arguments is supported by the finding that in most Spallation reactions a small number of relativistic fragments is emitted (0 ≤ n_s ≤ 14) whereas the number of relativistic fragments is BURST reactions in most cases is large (n_s > 24).

・ The 3-dimensional probability distribution of low-energy fragment (n_b) vs. relativistic high-energy fragment multiplicity (n_s) shows that Spallation reactions yield small numbers of fragments whereas in most BURST reactions a larger number of fragments is emitted.

・ In BURST reactions one does not find a single event where none or only one low-energy particle (n_b) is emitted together with less than 10 relativistic particles (n_s).

The Monte Carlo calculations were performed using MCNPX 2.7 code. The current form of the code does not allow full definition of the BURST events to be implemented in the simulations. This arises mainly from the fact that one cannot separate the projectile and target fragment from each other which is essential for the BURST definition. Therefore an event filtering method with some assumptions was used to select BURST events from 10,146 recorded inelastic events from the interaction of 72 GeV ²²Ne ions with nuclear emulsion (as described in Sections 4.1 and 4.2).

The method given in Section 4.1 yields mixed results when compared with two sets of experimental findings as shown in Table 3 and Table 4.

The BURST event selection criteria given in Section 4.2 select a special type of the events in which BURST events are accompanied by total destruction of AgBr in emulsion. For this type of interactions, the distributions of multiplicities of the special BURST events are not given in this paper, because they look very similar to the TD events as presented in Ref. [11] .

The calculated results show that Spallation reactions in the interaction of 72 GeV ²²Ne in nuclear emulsion are understood as calculated Spallation results match the experiments well.

One cannot exclude that the differences mentioned above come from peculiarities that happen in the second and following generations of interactions.

From the findings in Ref. [2] and observation of the highly increased rate of the BURST events in the second generation (Figure 2), it is clear that relativistic secondary particles interact more efficiently with the next target nucleus than a primary particle does. It looks as if the secondary particle is bigger, at least for a short while, after the first interaction. Such observation is definitively not in agreement with standard models of physics and is not expected to be predicted by the Monte Carlo simulations; it needs further studies to understand these experimental findings.

At this level of limited knowledge, it is too early to discuss any technological application using the “enhanced destruction capabilities” of secondary fragments [17] .

More detailed simulations of the BURST events (as defined in this paper) are required, and we invite other authors to perform such simulations using different codes. In such future model calculations, it would be important to provide simulation results and answers for:

・ Angular distribution of the prongs with respect to the beam axis. Also examine if model calculations reproduce events in which in the exit channel no charged particles are emitted within a forward cone covering an angle of 14˚.

・ An explanation why relativistic fragments (especially He-ions) from the first generation of ion-nucleus interactions are more efficient in inducing BURST events than normal primary relativistic ions.

Hashemi-Nezhad, R., Brandt, R., Ditlov, V., Haiduc, M., Firu, E., Neagu, A.T., Ganssauge, E. and Westmeier, W. (2017) Further Studies of BURSTS and Spallation in High-Energy Heavy Ion Reactions. World Journal of Nuclear Sci- ence and Technology, 7, 35-57. http://dx.doi.org/10.4236/wjnst.2017.71004