betaFeedback

WorksPhysics and Astronomy

Multi-Nucleon Transfer Reactions and the Creation and the Evolution of the Compound Nucleus

Abstract

We present the first implementation of a novel extension of the Generator Coordinate Method (GCM), dubbed the enhanced GCM (eGCM), which is applied to the grazing Multi-Nucleon Transfer (MNT) reaction 48^{48}Ca+208^{208}Pb near the Coulomb barrier. eGCM incorporates major qualitative differences with either Time-Dependent Hartree-Fock (TDHF) or GCM frameworks used until now for nuclear reactions. We demonstrate that the eGCM framework describes for the first time in a fully quantum microscopic framework the emergence and the time-evolution of Niels Bohr's 1936 conjectured Compound Nucleus (CN). The thermalization time extracted in eGCM is at least two orders of magnitude larger than the Eigenstate Thermalization Hypothesis (ETH) would predict.

Ninety years ago Niels Bohr introduced, based on qualitative arguments and some inferences from experimental results, the concept of a compound nucleus (CN) [1–4], when the individualities of the two initial reaction partners are “irretrievably lost,” a concept which has been widely used in nuclear physics ever since. No microscopic derivation based on the many-body Schrödinger equation was ever convincingly introduced and currently the CN is simply a theoretical conjecture, however, amply used in phenomenological approaches and consistent with many experimental results [3–13]. In the editorial [2] to the Bohr [1] paper, the CN is pictorially described as a superposition of a very large number of “simpler” many-body states, which basically form a continuum and the best guess anybody currently has is that the coefficients in this expansion are complex random numbers, in agreement with the Wigner-Dyson surmise in Random Matrix Theory [14], which has been extensively studied in the case of the nuclear shell-model [15, 16], the relation between classical and quantum chaos in quantum systems with a relatively small number of degrees of freedom [17–20] and for nuclear reactions between heavy-ion nuclei or fission [21–23], processes which are strongly non-equilibrium time-dependent phenomena. The loss of memory conjectured by Niels Bohr [1, 2] is related classical Boltzmann equation [24], Poincaré recurrence time [25], Eigenstate Thermalization Hypothesis (ETH) [19, 26–34] and quantum scars [35, 36], and strong interference between events with extremely large spatial and/or temporal separations or heavy quantum objects [37–43], see End Matter and Ref. [44]. The multi-nucleon transfer reactions might appear to be in a somewhat better position theoretically, as a theoretical framework exists, the Time-Dependent Hartree-Fock (TDHF) method, with only the nagging questions of whether TDHF has negligible quantum mean field fluctuations and is accurate enough.

One of the most fundamental questions in physics is how many elements exist and what is the mechanism of their creation, in the laboratory and in Nature, in stars, in the collisions of neutron stars and black holes and in supernova collapses [45–48]. Elements from Z=1Z=1 (Hydrogen) to Z=92Z=92 (Uranium) were all discovered on Earth until 1939, with a very small number of exceptions, as those inferred from the Oklo natural reactor in Gabon. The number of isotopes which exist naturally on Earth is about 339, including 251 primordial isotopes and 88 radioactive isotopes. Among the rest of the predicted 7000-8000 possible isotopes [49, 50], only about 3,000-3,300 have been created so far in many laboratories throughout the World [51, 52]. While most of these predicted isotopes are not naturally stable, their even rather short-time existence as an intermediate stage in the creation of other elements and isotopes is crucial in order to predict the abundances in the Solar system in particular. The overwhelming number of nuclei and isotopes not found naturally on Earth have been created in multi-nucleon transfer (MNT) heavy-ion reactions and the quest to identify the rest of them is an ongoing process. Unfortunately, there’s still no microscopic framework accurate enough to replace the many existing phenomenological models, which are based on various assumptions and are overwhelmingly classical in their character. These approaches have been discussed in recent overviews and in Refs. [11, 53–58]. In most cases the typical phenomenological equation for production of a heavy reaction product used [11, 53, 54] is

σEVR=∑J=0Jmax⁡σcap(ECoM,J)PCN(E∗,J)Wsur(E∗,J),(1)\sigma_{\mathrm{EVR}}=\sum_{J=0}^{J_{\max}}\sigma_{\mathrm{cap}}(E_{\mathrm{CoM}},J)P_{\mathrm{CN}}(E^*,J)W_{\mathrm{sur}}(E^*,J), \tag*{(1)}

where the CN formation probability PCNP_{\mathrm{CN}} is phenomenological with wide variations, and which is the only quantity discussed in this work.

The most sophisticated microscopic model existing in the literature is the time-dependent mean field theory in either its TDHF incarnation or in its extension, the time-dependent Density Functional Theory (TDDFT) [61, 62, 63, 64] [59, 60], an approach which is based on the adequacy of using mean field, which is an expectation of a one-body operator and thus classical in character. This classical character of the mean field approximation can be corrected in a configuration interaction (CI) framework, of which the generator coordinate method (GCM) was believed for a long time to be a good solution [65, 66, 67, 68], with some reservations [69], although this claim was recently challenged with the introduction of the enhanced GCM (eGCM) framework [44].

The GCM [66] is basically a CI-method based on linearly independent non-orthogonal system of ground state static (generalized) Slater determinants with fixed shapes, which were assumed to describe an adiabatically evolving nucleus from the neutron capture to scission in induced fission [66, 70, 71]. In the case of heavy-ion reactions however the reaction fragments acquire significant excitation energy, not accounted for in an adiabatic evolution as Reinhard et al. [67] noticed, which also happens in induced fission [72, 73, 74, 75]. The corresponding many-body wave functions for the three approaches are defined as

ΨGCM(ξ1…ξA)=∫Qf(Q)Φ(ξ1…ξA∣Q),(2)\Psi_{\mathrm{GCM}}(\xi_1 \ldots\xi_A) = \int_Q f(Q)\Phi(\xi_1 \ldots\xi_A \mid Q), \tag*{(2)}
ΨGCMR(ξ1…ξA,t)=∫Qf(Q,t)Φ(ξ1…ξA∣Q,t),(3)\Psi_{\mathrm{GCMR}}(\xi_1 \ldots\xi_A,t) = \int_Q f(Q,t)\Phi(\xi_1 \ldots\xi_A \mid Q,t), \tag*{(3)}
ΨeGCM(ξ1…ξA)=∫Qf(Q,τ)Φ(ξ1…ξA∣Q,τ).(4)\Psi_{\mathrm{eGCM}}(\xi_1 \ldots\xi_A) = \int_Q f(Q,\tau)\Phi(\xi_1 \ldots\xi_A \mid Q,\tau). \tag*{(4)}

The difference between ΨGCM\Psi_{\mathrm{GCM}} and ΨGCMR\Psi_{\mathrm{GCMR}} is the time dependence of the Slater determinants in the latter [67]. eGCM [44] on the other hand is a “bouillabaisse” of linearly independent non-orthogonal time-dependent (generalized) Slater determinants, generated with a large set of initial conditions, and mixed at all times τ\tau along their TDDFT trajectories, and thus in the continuum limit a path integral [76] along the TDHF trajectories [44].

We will demonstrate here that eGCM is a many-body framework which leads to many-body wave functions of extreme complexity, often quantified by an entanglement entropy [77, 78, 79, 80], describing the collision of heavy ions in particular, which has the expected spatial symmetry from quantum mechanics. We establish that a very large number of relevant many-body configurations were ignored in all previous incarnations of GCM. This is an argument in line with what one would expect from the Fermi golden rule, namely that the number of relevant final states is controlled not only by the magnitude of average transition matrix element at each time step, but more importantly, by the number of such accessible final states.

In any scattering of a plane wave of an unpolarized projectile on an unpolarized target the many-body wave function has an azimuthal symmetry around the axis of the impinging beam, which is faithfully reproduced in eGCM as implemented here. The components of this plane wave corresponding to relatively large impact parameters interact with the target only due to the role of the long range Coulomb repulsion between the reaction partners and nucleon transfer is suppressed. Only those impact parameters where the two reaction fragments come into the range of the nucleon-nucleon

interaction might significantly contribute to nucleon transfer. One has to distinguish at this point between grazing collisions and collisions at small impact parameters of the two reaction fragments, which lead to a large variety of “CN-like states” with life times varying by many orders of magnitude, see review by Hinde et al. [11] and the End Matter. These various long lived states will not be described in detail at this stage into the eGCM framework, but may collectively be referred to as the emerging CN. For collision energies below the Coulomb barrier, the nucleon transfer is expected to have an essentially tunneling under the barrier character and thus be significantly suppressed. Simenel et al. [62] discuss in their recent work the role of the contact time in 40…56Ca+176Yb{}^{40\ldots56}\mathrm{Ca}+{}^{176}\mathrm{Yb} collisions at energies just above the Coulomb barrier and demonstrate that the formation of reaction fragments heavier than the target (which is usually the heavier nucleus) occurs only for relatively short contact times between the reaction fragments. These are the kind of reactions we will discuss in this initial implementation of eGCM to nuclear reactions. We will describe the collision 48Ca+208Pb{}^{48}\mathrm{Ca}+{}^{208}\mathrm{Pb} at a center-of-mass energy 235 MeV, which is slightly above the Coulomb barrier between the two initial partners, in both a time-independent and a time-dependent frameworks, as these points of view have different physical interpretations. This reaction has been studied experimentally and theoretically [81] [11], where the phenomenon of quasi-fission is dominant. At long contact times the 256No102154?{}^{256}\mathrm{No}_{102}{}^{154}\mathrm{?} non-equilibrated nucleus is formed during central collisions of the two initial reaction fragments, experiencing both quasi-fission and fission stages. One can thus separate to a rather good approximation the time evolution of such a collision roughly into a fast and a slow phase, characterized by the contact time between the two reaction partners.

As a first step we perform TDDFT simulations (no pairing) of the collision 48Ca+208Pb{}^{48}\mathrm{Ca}+{}^{208}\mathrm{Pb} at energies slightly above the Coulomb barrier with the impact parameters within a ring in the impact parameter space shown in Figure 1 and we stored the TDDFT single-particle wave functions φk(α)\varphi_k(\alpha), where k=1,…,256k=1,\ldots,256 and α=(bx,by,τ)\alpha=(b_x,b_y,\tau), at various times τ\tau along each TDDFT trajectory labelled by the impact parameter b\mathbf{b}. The TDDFT simulations are performed with the energy density functional SeaLL1 [50] in a simulation box 64364^3 with a lattice constant 1 fm and with the code LISE [82]. The storage of the TDDFT single-particle wave functions used in our study required about 80 TB of memory, and the construction of the eGCM Hamiltonian matrix was performed using 48,000 GPUs on the Frontier supercomputer at Oak Ridge National Lab. The number of times τ\tau the TDDFT wave functions stored for use in GCM calculation was Nτ=385N_\tau=385 with a step Δτ=6.4\Delta\tau=6.4 fm/c. We have performed as well some checks of the eGCM framework with Δτ=3.2\Delta\tau=3.2 fm/c. The total number of impact parameters in the red ring in Figure 1 was Nb=152N_b=152 and the size of the GCM matrix was 39,630, which is arguably the largest basis set used in a nuclear GCM calculation, and particularly in the continuum. For particle number projection in the final state, we used 56,000 GPUs.

Image not extracted

Figure 1. In the impact parameter plane b=(bx,by)\mathbf{b}=(b_x,b_y) only the impact parameter points with centers inside the red ring, with ∣b∣≈5|\mathbf{b}|\approx5–66 fm, were considered. The insets show representative density profiles from the TDHF simulations. On the left are trajectories with b=(5,0)\mathbf{b}=(5,0) and on the right with b=(6,0)\mathbf{b}=(6,0). In the upper frames t=420t=420 fm/cc and the lower frames are shown at their final time. The white circles in the insets indicate the region in which particle number projection is performed at late times.

Using the eigenvalues and the eigenvectors of the norm overlap matrix between various Slater determinants, all linearly independent since νk>0\nu_k>0, see Figure 4, we introduce two new sets of orthogonal linear combinations of Slater determinants ∣Φˉk⟩|\bar{\Phi}_k\rangle and ∣Φˉ(α)⟩|\bar{\Phi}(\alpha)\rangle, in the formulation of the static eGCM equation, Eqs. (8), (10)

Image not extracted

Figure 4. The eGCM energies (blue line) for all trajectories used in this analysis versus the diagonal matrix elements ⟨Φk∣H^∣Φk⟩\langle\Phi_k|\hat{H}|\Phi_k\rangle (red dashed line). The inset displays the eigenvalues νk\nu_k of the norm overlaps, see Eq. (5).

∫dα′⟨Φ(α)∣Φ(α′)⟩gk(α′)=νkgk(α),min⁡(νk)>0.0011,(5)\int d\alpha'\langle\Phi(\alpha)|\Phi(\alpha')\rangle g_k(\alpha')=\nu_k g_k(\alpha),\quad\min(\nu_k)>0.0011, \tag*{(5)}
∫dα gk∗(α)gl(α)=δkl,∫dα gk(α)gk∗(β)=δ(α−β),(6)\int d\alpha\,g_k^*(\alpha)g_l(\alpha)=\delta_{kl},\qquad\int d\alpha\,g_k(\alpha)g_k^*(\beta)=\delta(\alpha-\beta), \tag*{(6)}
∣Φˉk⟩=∫dα νk−1/2gk(α)∣Φ(α)⟩,νk>0,⟨Φˉn∣Φˉm⟩=δnm(7)|\bar{\Phi}_k\rangle=\int d\alpha\,\nu_k^{-1/2}g_k(\alpha)|\Phi(\alpha)\rangle,\quad\nu_k>0,\quad\langle\bar{\Phi}_n|\bar{\Phi}_m\rangle=\delta_{nm} \tag*{(7)}
⟨Φˉk∣H^∣Ψn⟩=∑l⟨Φˉk∣H^∣Φˉl⟩hln=hknEn,(8)\langle\bar{\Phi}_k|\hat{H}|\Psi_n\rangle=\sum_l\langle\bar{\Phi}_k|\hat{H}|\bar{\Phi}_l\rangle h_{ln}=h_{kn}E_n, \tag*{(8)}
∣Φˉ(α)⟩=∑kgk∗(α)∣Φˉk⟩,⟨Φˉ(α)∣Φˉ(β)⟩=δαβ,(9)|\bar{\Phi}(\alpha)\rangle=\sum_k g_k^*(\alpha)|\bar{\Phi}_k\rangle,\quad\langle\bar{\Phi}(\alpha)|\bar{\Phi}(\beta)\rangle=\delta_{\alpha\beta}, \tag*{(9)}
⟨Φˉ(α)∣H^∣Ψn⟩=∑β⟨Φˉ(α)∣H^∣Φˉ(β)⟩hβn=hαnEn,(10)\langle\bar{\Phi}(\alpha)|\hat{H}|\Psi_n\rangle=\sum_\beta\langle\bar{\Phi}(\alpha)|\hat{H}|\bar{\Phi}(\beta)\rangle h_{\beta n}=h_{\alpha n}E_n, \tag*{(10)}

where α=(b,τ)\alpha=(b,\tau), with τ\tau being the additional and crucial generator coordinate that distinguishes eGCM from all previous GCM-type approaches in the literature. The eGCM formulations in Eqs. (8), (10) are each equivalent to a CI solution of the many-body problem in the continuum in the present case. The relation between the many-body wave functions sets ∣Φˉk⟩|\bar{\Phi}_k\rangle and ∣Φˉ(α)⟩|\bar{\Phi}(\alpha)\rangle is similar to the relation between the Bloch and Wannier wave functions in crystals [83] and a similar relation between localized (sinc⁡(x)\operatorname{sinc}(x)) and delocalized wave functions in the discrete variable representation method (DVR) [84].

If the eigenvectors and the eigenvalues of Eq. (8) are determined one can easily construct the solution of the time-dependent eGCM (TDeGCM) equation with specified initial conditions

iℏ∂t∣Ψ(t)⟩=H^∣Ψ(t)⟩,⟨Ψ(t)∣Ψ(t)⟩=1,(11)i\hbar\partial_t|\Psi(t)\rangle=\hat{H}|\Psi(t)\rangle,\quad\langle\Psi(t)|\Psi(t)\rangle=1, \tag*{(11)}
∣Ψ(t)⟩=∑ncn(t)∣Ψn⟩,cn(t)=cn(0)exp⁡(−iEntℏ),(12)|\Psi(t)\rangle=\sum_n c_n(t)|\Psi_n\rangle,\quad c_n(t)=c_n(0)\exp\left(-\frac{iE_nt}{\hbar}\right), \tag*{(12)}
cn(0)=⟨Ψn∣Ψ(0)⟩=N∫db∑kgk∗(b,0)νkn1/2hkn,(13)c_n(0)=\langle\Psi_n|\Psi(0)\rangle=\mathcal{N}\int db\sum_k g_k^*(b,0)\nu_{kn}^{1/2}h_{kn}, \tag*{(13)}

and N\mathcal{N} is the normalization. The initial many-body wave function is a uniform superposition over all initial states in the ring of impact parameters illustrated in Figure 1, and so the many-body wave function has by construction the expected azimuthal symmetry. In the TDeGCM with initial conditions as we described above the total energy Etot=⟨Ψ(t)∣H^∣Ψ(t)⟩E_{\mathrm{tot}}=\langle\Psi(t)|\hat{H}|\Psi(t)\rangle is conserved. Since ∣Ψ(t)⟩|\Psi(t)\rangle is a superposition of eigenstates the initial state has a mean energy Etot=−1817.5E_{\mathrm{tot}}=-1817.5 MeV and a standard deviation ΔE=27.9\Delta E=27.9 MeV. Since ΔEΔT<ℏ\Delta E\Delta T<\hbar and initial relative velocity of the two nuclei vrel=0.113cv_{\mathrm{rel}}=0.113c is significantly smaller than the Fermi velocity, the discretization Δτ\Delta\tau and Δb\Delta b (see Figure 1) appear to be quite satisfactory [85]. Eq. (12) demonstrates that once there is a full solution of any static GCM implementation, there is no need for a GCM time-dependent code [86, 87, 88, 89]. One can introduce then the inverse participation ratios (IPR), see End Matter, for the time-dependent many-body wave function ∣Ψ(t)⟩|\Psi(t)\rangle

∣Ψ(t)⟩=∑kak(t)∣Φˉk⟩=∫dα∣d(α∣t)∣Φˉ(α)⟩,(14)|\Psi(t)\rangle=\sum_k a_k(t)|\bar{\Phi}_k\rangle=\int d\alpha|d(\alpha|t)|\bar{\Phi}(\alpha)\rangle, \tag*{(14)}

using the either ak(t)a_k(t) or d(α∣t)d(\alpha|t) expansion coefficients. The probabilities to find NN and ZZ neutrons and protons in the heavy reaction fragment is given by [90] [59, 60]

PH(N,Z)=∫∫dηNdηZ(2π)2⟨Ψ(t)∣eiηN(N^N−N)+iηZ(N^Z−Z)∣Ψ(t)⟩,P_H(N,Z)=\int\int\frac{d\eta_Nd\eta_Z}{(2\pi)^2}\langle\Psi(t)|e^{i\eta_N(\hat{N}_N-N)+i\eta_Z(\hat{N}_Z-Z)}|\Psi(t)\rangle,
N^N,Z=∫dξ ΘN,Z(r)ψ†(ξ)ψ(ξ),\hat{N}_{N,Z}=\int d\xi\,\Theta_{N,Z}(r)\psi^\dagger(\xi)\psi(\xi),

where N^N,Z\hat{N}_{N,Z} is the neutron/proton number operator, ξ\xi stand for space-spin-isospin coordinates, ∣Ψ(t)⟩|\Psi(t)\rangle is the final many-body wave function and ΘH(r)\Theta_H(r) is non-vanishing in the heavy reaction fragment region. In the case of the extension of GCM due to Reinhard et al. [67], denoted here GCMR_R, both the norm overlap and the Hamiltonian overlap matrices have a block-diagonal structure with ⟨Φ(b,τ)∣Φ(b′,τ′)⟩∝δτ,τ′\langle\Phi(b,\tau)|\Phi(b',\tau')\rangle\propto\delta_{\tau,\tau'} and ⟨Φ(b,τ)∣H^∣Φ(b′,τ′)⟩∝δτ,τ′\langle\Phi(b,\tau)|\hat{H}|\Phi(b',\tau')\rangle\propto\delta_{\tau,\tau'}, where the number of impact parameters in the present case is typically Nb=152\mathcal{N}_b=152, unlike in the case of eGCM where the dimension of the corresponding matrices is 39,630 in our simulations (arguably the largest GCM simulation ever reported in the literature) [44]. The kinetic energies of all the included TDDFT trajectories are 235 MeV, and the total TDHF energy is shown with the thin solid line in Figure 4 in End Matter. The static eGCM EnE_n spectrum shows an enormous degree of almost degeneracy around the TDHF energy. Within eGCM the basis set is constructed from Nb=152N_b = 152 full TDHF trajectories. For all these impact parameters the reaction fragments separate at the end of the TDHF simulation.

The inverse participation ratios (IPR) for the norm and energy overlaps provide more insights into the structure and the physical interpretations of the eGCM versus the GCMR_R wave functions. Figure 5 in End Matter alone demonstrates the overwhelming superiority of the eGCM over GCMR_R. In case of nuclear pairing correlations for example the spreading of nucleon pairs over about ten degenerate orbitals leads to the what many would call a “phase transition” in a finite system, as seen for example in the magic 100Sn{}^{100}\mathrm{Sn} versus the neutron superfluid 120Sn{}^{120}\mathrm{Sn}. The eGCM Hamiltonian and norm overlaps IPRs are more than two orders of magnitude larger than the maximum value in the case of GCMR_R, which is Nb=152N_b = 152 and which was hoped to lead to a satisfactory description of nuclear reactions [67, 91, 92, 93, 94, 95, 96] [57].

Image not extracted

Figure 5. The IPR for the Norm (thin lines) and the Hamiltonian (thick lines) eGCM overlaps. The maximum possible IPR for the GCMR_R overlaps is shown with a dashed line.

Once the incident projectile and target come into contact, the IPRs start rising dramatically, see Figure 2. Reaction fragments originating from different impact parameters bb and b′b' arrive or leave this interaction region at slightly different times τ\tau and τ′\tau' and that leads to mixing between trajectories after the reaction fragments separate, in the gray region in the Figure 2.

Image not extracted

Figure 2. The IPRs are evaluated for the wave function ∣Ψ(t)⟩\lvert\Psi(t)\rangle, see Eq. (14), using either the expansion coefficients ak(t)a_k(t) (blue line, continued red after tfin=2,482t_{f_{\rm in}} = 2,482 fm/c), and coefficients d(α∣t)d(\alpha|t) (green line, continued purple after tfint_{f_{\rm in}}). The vertical black dashed line indicates the time when all the nuclei first make contact in TDHF. The gray banded region indicates the range of times during which the nuclei separate in TDHF. The inset displays the coefficients ∣cn(t)∣2\lvert c_n(t)\rvert^2, as well as the sorted coefficients ∣d(α∣tfin)∣2\lvert d(\alpha|t_{f_{\rm in}})\rvert^2, ∣ak(tfin)∣2×1.75\lvert a_k(t_{f_{\rm in}})\rvert^2 \times1.75, and ∣cn(tfin)∣2\lvert c_n(t_{f_{\rm in}})\rvert^2 with black, orange, and red lines respectively.

A very unexpected result emerging from the eGCM treatment of MNT reactions is the stark difference with the results of TDHF simulations. In all our TDHF simulations, the two nuclei came into contact in about tc=162t_c = 162 fm/c and all of them separated well before we stopped our simulations at tfin=2,482t_{f_{\rm in}} = 2,482 fm/c. The trajectories with b=5b = 5 fm separate at tsep=2,262t_{\rm sep} = 2,262 fm/c and those with b=6b = 6 fm separate at tsep=775t_{\rm sep} = 775 fm/c, meaning that the nucleus spends 85% and 25% of the total simulation time in the transitional “CN-like” system configuration for b=5b = 5 fm and b=6b = 6 fm respectively. In eGCM however we observed the emergence of a CN 102256No154{}^{256}_{102}\mathrm{No}_{154}, with an excitation energy Eex≈78E_{\rm ex} \approx78 MeV, with Z=102Z = 102 protons and N=154N = 154 neutrons with a probability PH(154,102)=0.34P_H(154,102) = 0.34, which is totally unexpected, as it could never have been found in any previous TDHF MNT simulations, see Figure 3. The CN does not break apart even when simulated for very long times, a result totally at odds with all TDHF trajectories included in our basis. The most surprising outcome however is that PH(N,Z)=∑N>127∑Z>83P(N,Z)≈0.85P_H(N,Z)=\sum_{N>127}\sum_{Z>83}P(N,Z)\approx0.85. We also find that ∑N≤126∑Z>83P(N,Z)≈0.012\sum_{N\leq126}\sum_{Z>83}P(N,Z)\approx0.012 and ∑N>127∑Z≤82P(N,Z)≈0.006\sum_{N>127}\sum_{Z\leq82}P(N,Z)\approx0.006, indicated that MNT favors transfer in either directions of both protons and neutrons, while PH(N,Z)=∑N≤126∑Z≤82P(N,Z)≈0.13P_H(N,Z)=\sum_{N\leq126}\sum_{Z\leq82}P(N,Z)\approx0.13.

Image not extracted

Figure 3. Final probabilities for the heavy fragment PH(N)=∑ZPH(N,Z)P_H(N) = \sum_Z P_H(N,Z) and PH(Z)=∑NPH(N,Z)P_H(Z) = \sum_N P_H(N,Z), showcasing the dramatic differences between TDHF and eGCM. The black and green lines with additional dots are the TDHF probabilities PH(N)P_H(N), and the black and green lines without addition dots refer to PH(Z)P_H(Z). By contrast, the eGCM distributions PH(N)P_H(N) and PH(Z)P_H(Z) are shown in red-with-dots and blue, respectively. The difference between the TDHF and eGCM particle probabilities are as stark as Isaac Newton’s and Thomas Young’s predictions would be for the two-slit experiment.

We used the eGCM energy spectrum, see Eqs. (8),(10), to establish if the properties of our CN are consistent with the Wigner-Dyson surmise in Random Matrix Theory [14–16], and thus in agreement with Bohr [1] CN hypothesis, see End Matter. Our eGCM spectrum has a quite low average level spacing ≈4\approx4 keV, and the normalized distribution of level spacings in eGCM has a standard deviation of 0.5273, in close agreement with the value 0.5227 which one obtains in the Gaussian Orthogonal Ensemble (GOE), see also End Matter. This demonstrates that eGCM is capable of describing a very complex CN formation. The proton and neutron numbers, average and standard deviations, for the heavy fragment are 97.6 ±\pm 7.3 and 147.7 ±\pm 10.2, which demonstrates that even in grazing collisions a significant number of neutrons and protons are transferred predominantly to the heavy target, in contrast to TDHF, see extensive recent published TDHF results [62]. We have also evaluated the total density of the emerged CN at our simulation and in End Matter we present density profiles. The final values of the CN rms radius ⟨r2⟩\langle r^2\rangle = 7.48 fm, quadrupole and octpole moments ⟨Q20⟩\langle Q_{20}\rangle = 29.23 b and ⟨Q30⟩\langle Q_{30}\rangle = 5.99 b3/2^{3/2}, and nucleon density profiles in End Matter at tt = 2,482 fm/c demonstrate that full thermalization and “memory loss” have not been achieved. The CN thermalization time exceeds by orders of magnitude τETH≈ℏΔE≈7\tau_{ETH} \approx\frac{\hbar}{\Delta E} \approx7 fm/c [31], in agreement with thermalization in other strongly interacting quantum many-body systems Refs. [97, 98] [33]. The analysis of experimental results [81] [11], is beyond the scope of this first implementation of eGCM.

eGCM is clearly superior to the tools used so far in the literature, as it emphasizes the major role of the quantum interference and of the entanglement among the components of the eGCM many-body wave functions, not encountered so far in nuclear physics. The quantum interference is manifested in the high probability to form a CN with a very long stability and very high ZZ and NN, P(102256No154)=0.34P({}^{256}_{102}\mathrm{No}_{154}) = 0.34, is the quantum expression of Poincaré’s theorem [25] of why this emerged CN has an enormous life-time. In eGCM the role of nucleon-nucleon collisions, in particular neutron-proton collisions is significantly enhanced. In this work we have evaluated only the incipient stages of the CN emergence and evolution.

Acknowledgments —We thank P. G. Reinhard, I. Stetcu, I. Abdurrahman, G. A. Miller, L. G. Sobotka, T. Kawano, M. J. Savage and K. Sekizawa for discussions and comments. AB acknowledges the funding from the Department of Energy Office of Science, Grant No. DE-FG02-97ER41014. This material is additionally based upon work supported by the Department of Energy, National Nuclear Security Administration, under Award Number DE-NA0004150, the Center for Excellence in Nuclear Training And University-based Research (CENTAUR). This research used resources of the Oak Ridge Leadership Computing Facility, which is a U.S. DOE Office of Science User Facility supported under Contract No. DE-AC05-00OR22725.

Data and codes availability —The data that support the findings of this article are not publicly available upon publication, as it is not technically feasible and/or the cost of preparing, depositing, and hosting the data would be prohibitive within the terms of this research project. Reduced data sets are available from the authors upon reasonable request. Codes, which will become public in time, can be provided without expectations of any guidance.

References

End Matter

Complexity of the Compound Nucleus state. An important aspect of Bohr’s CN conjecture is the loss of memory of the initial state [1, 2, 34]. The loss of memory is similar to Boltzmann’s [24] and Poincaré’s statements about the evolution of complex mechanical systems, in spite of the fact that they are governed by time-reversal invariant equations. Schrödinger’s equation for nuclear systems is time reversal invariant, and thus the statement that memory is lost should be interpreted with care. Naively, one can assume that within TDDFT framework chaos can arise, since the equations are, by construction, non-linear partial differential equations. However, as shown in Refs. [82, 99], the Lyapunov exponents in TDDFT are extremely small and practically negligible. Within TDDFT simulations for nuclear or cold atom many-body systems performed with numerical double precision there is no chaotic behavior. In the classical limit, Poincaré theorem [25] relates the loss of memory of the initial conditions to the long Poincare recurrence time, which by transference is expected to apply to a CN. To the authors’ knowledge so far no microscopic quantum mechanical framework for the formation and the evolution of a CN was ever suggested. In the editorial [2], published simultaneously with Niels Bohr original paper [1], a classical model for the CN was suggested, shown in Fig. 1 in [2], the collision of a billiard ball with an ensemble of billiard balls at rest in a bowl resting on a table. The impinging ball quickly looses its initial kinetic energy and the ensemble of billiard balls originally in the bowl can return to rest only if that energy is returned eventually to a single billiard ball, which can then “escape” from the bowl. This time however, according to the Poincaré’s theorem [25], is unimaginable long, in agreement with the main assumption of the classical Boltzmann equation [24], which implies irreversibility, in “mathematical” disagreement with the time-reversal invariance of the classical equations of motion. In the limit of quantum many-body quantum dynamics, this topic is vast and related, in part, to the Eigenstate Thermalization Hypothesis (ETH) [97, 98] [19, 26, 27, 28, 29, 30, 31, 32, 33] and the present work, quantum scars [35, 36], classical chaos versus quantum chaos within the Gaussian Orthogonal Ensemble [14–18, 20], where many similarities and disagreements arise when compared to the classical limit. The emerging CN starts with a relatively simple many-body wave function, see Eq. (20) below, and evolves into a many-body wave function characterized by a complexity which is significantly and quite rapidly increasing in time as demonstrated in the present work, see Figure 2. In eGCM, unlike in all previous implementations of GCM, a strong interference between different events originating from different either static DFT or TDDFT configurations with large spatial and/or temporal separations are allowed. This type of interference is well known in astronomy [37, 38], in nuclear physics [39, 40], and in two-slit experiments with very low intensity beams, when the particles following different paths hit the screen at very different times [41–43], or between heavy objects [100] as can also occur in heavy-ion collisions, as described here, or fission [5–13], where CN life times are extracted to be of order of attoseconds.

Static and time-dependent eGCM and GOE properties of the eGCM spectrum. Since all the eigenvalues of the overlap matrix Eq. (5) are non-vanishing (νk≥0.0011263\nu_k \ge0.0011263) the entire set of static eGCM Slater determinants ∣Φ(α)⟩|\Phi(\alpha)\rangle are linear independent and the sets of the Slater determinants ∣Φ(α)⟩|\Phi(\alpha)\rangle and of the orthogonal many-body wave functions ∣Φˉk⟩|\bar{\Phi}_k\rangle and ∣Φˉ(α)⟩|\bar{\Phi}(\alpha)\rangle all have the same size, namely 39,630. All many-body wave functions defined in this work depend on 3×256=7683 \times256 = 768 spatial coordinates and 256 spin coordinates. The static eGCM eigenfunctions ∣Ψn⟩|\Psi_n\rangle are defined in Eqs. (8), (10) are

∣Ψn⟩=∑αhαn∣Φˉα⟩=∑βhβn∣Φˉ(β)⟩.(15)|\Psi_n\rangle= \sum_{\alpha} h_{\alpha n}|\bar{\Phi}_{\alpha}\rangle= \sum_{\beta} h_{\beta n}|\bar{\Phi}(\beta)\rangle. \tag*{(15)}

In GCM one needs to introduce an energy density functional E(nα,α′,τα,α′,…)\mathcal{E}(n^{\alpha,\alpha'},\tau^{\alpha,\alpha'},\ldots) depending on the GCM number, kinetic and densities [101, 102],

Mˉkl(α,α′)=⟨ψk(α)∣ψl(α′)⟩,(16)\bar{M}_{kl}(\alpha,\alpha') = \langle\psi_k(\alpha)|\psi_l(\alpha')\rangle, \tag*{(16)}
⟨Φ(α)∣Φ(α′)⟩=det⁡Mˉ(α,α′).(17)\langle\Phi(\alpha)|\Phi(\alpha')\rangle= \det\bar{M}(\alpha,\alpha') . \tag*{(17)}
nα,α′(r,σ∣r′,σ′)=⟨Ψ(α)∣ψ†(r′,σ′)Ψ(r,σ)∣Ψ(α′)⟩⟨Ψ(α)∣Ψ(α′)⟩=∑klϕk(α′∣r,σ)ρklα,α′ϕl∗(α∣r′,σ′),(18)\begin{aligned} n^{\alpha,\alpha'}(\mathbf{r},\sigma|\mathbf{r}',\sigma') = \frac{\langle\Psi(\alpha)|\psi^\dagger(\mathbf{r}',\sigma')\Psi(\mathbf{r},\sigma)|\Psi(\alpha')\rangle}{\langle\Psi(\alpha)|\Psi(\alpha')\rangle} \\ &= \sum_{kl}\phi_k(\alpha'|\mathbf{r},\sigma)\rho_{kl}^{\alpha,\alpha'}\phi_l^*(\alpha|\mathbf{r}',\sigma') , \tag*{(18)} \end{aligned}
ρklα,α′=⟨Ψ(α)∣ck†cl∣Ψ(α′)⟩⟨Ψ(α)∣Ψ(α′)⟩=Mˉkl−1(α,α′),(19)\rho_{kl}^{\alpha,\alpha'} = \frac{\langle\Psi(\alpha)|c_k^\dagger c_l|\Psi(\alpha')\rangle}{\langle\Psi(\alpha)|\Psi(\alpha')\rangle} = \bar{M}_{kl}^{-1}(\alpha,\alpha') , \tag*{(19)}

where ψ†(r′,σ′)\psi^\dagger(\mathbf{r}',\sigma') and ψ(r,σ)\psi(\mathbf{r},\sigma) are creation and annihilation field operators and ϕk(α∣r,σ)\phi_k(\alpha|\mathbf{r},\sigma) are single-particle wave functions defining ∣Φ(α)⟩|\Phi(\alpha)\rangle. The GCM energy is [101, 102]

E=∑α,α′⟨Φ(α)∣Φ(α′)⟩×E = \sum_{\alpha,\alpha'}\langle\Phi(\alpha)|\Phi(\alpha')\rangle\times
∫r,σ,σ′E(nα,α′(r,σ∣r′,σ′),τα,α′(r,σ∣r′,σ′),…),\int_{\mathbf{r},\sigma,\sigma'} \mathcal{E}\left(n^{\alpha,\alpha'}(\mathbf{r},\sigma|\mathbf{r}',\sigma'),\tau^{\alpha,\alpha'}(\mathbf{r},\sigma|\mathbf{r}',\sigma'),\ldots\right),

where τα,α′\tau^{\alpha,\alpha'} is the kinetic energy density and ellipses stand for other needed densities. The notation for spin and isospin degrees of freedom is compounded here as σ\sigma.

A very useful measure of the complexity of the eGCM many-body wave functions, which reflects the degree of mixing between different TDDFT trajectories, are the inverse participation ratios (IPR) for the norm and Hamiltonian overlaps of many-body wave functions in various representations introduced in Eqs. (5), (7), (8), (12), (14),

IPR⁡N(k)=[∫Aα∣gk(α)∣4]−1,IPR⁡N(α)=[∫k∣gk(α)∣4]−1,\operatorname{IPR}_{N}(k)=\left[\int_{\mathcal{A}_{\alpha}}\left|g_k(\alpha)\right|^4\right]^{-1},\qquad\operatorname{IPR}_{N}(\alpha)=\left[\int_k\left|g_k(\alpha)\right|^4\right]^{-1},
IPR⁡H(n)=[∫k∣hk,n∣4]−1,IPR⁡H(α)=[∫n∣hα,n∣4]−1,\operatorname{IPR}_{H}(n)=\left[\int_k\left|h_{k,n}\right|^4\right]^{-1},\qquad\operatorname{IPR}_{H}(\alpha)=\left[\int_n\left|h_{\alpha,n}\right|^4\right]^{-1},

see Figure 5. All these many-body wave functions introduced in Eqs. (5), (7), (8), (12), (14) are related by unitary transformations between the many-body basis states, and some are more economical in practice than others and in some instances characterized by a corresponding intrinsic entanglement entropy [77, 78, 79, 80].

Our calculations show that the norm and Hamiltonian overlaps are not ever close in shape to a Gaussian and the Gaussian overlap approximation cannot be used in eGCM, see also Ref. [103]. Related issues concerning the agreement between exact TDGCM and TDGCM+GOA have been discussed in the literature, which show significant discrepancies between the two approaches [69].

The static eGCM spectrum in Figure 4 is “almost degenerate and linear” over most of the energy interval, with the comparison of level spacing distribution shown in Figure 6. These are the states which define the structure of the initial eGCM wave function Ψ(0)\Psi(0), see Eqs. (12), (13), when the colliding 48Ca{}^{48}\mathrm{Ca} and 208Pb{}^{208}\mathrm{Pb} are initially infinitely separated

Image not extracted

Figure 6. The GOE level spacing distribution P(s)P(s) compared to the corresponding unfolded eGCM spacing distribution. Ns(E)N_s(E) stands for the unfolded cumulative density of states. The eGCM spectrum has also a typical GOE oscillatory behavior lim⁡n→∞∣n−∑k=1nsk∣=O(1)\lim_{n\to\infty}\left|n-\sum_{k=1}^{n}s_k\right|=O(1) [14, 18].

∣Ψ(0)⟩=A(eiK⋅R∣Φ48Ca⟩⊗∣Φ208Pb⟩)=∑ncn(0)∣Ψn⟩,(20)|\Psi(0)\rangle=\mathcal{A}\left(e^{i\mathbf{K}\cdot\mathbf{R}}|\Phi_{{}^{48}\mathrm{Ca}}\rangle\otimes|\Phi_{{}^{208}\mathrm{Pb}}\rangle\right)=\sum_n c_n(0)|\Psi_n\rangle, \tag*{(20)}

where A\mathcal{A} stands for anti-symmetrization and appropriate normalization, R\mathbf{R} is the separation along the collision direction between the two nuclei, KK the relative momentum of their initial center-of-mass energy at infinity, and with cn(0)c_n(0) defined in Eq. (13). The coefficients cnc_n for eigenvalues EnE_n outside the energy interval (250, 36,500) MeV are negligible (≤10−16\le10^{-16}). As we discussed in the text, we considered only grazing collisions with impact parameters b∈[5,6]b \in[5,6] fm, which are typically considered in MNT reactions [61, 62, 63, 64] [59, 60]. For smaller impact parameters the two colliding nuclei do not separate in the time interval considered here in TDHF MNT reactions.

The nucleon density profiles in Figure 7 and the values ⟨r2⟩\langle r^2\rangle, Q20Q_{20}, Q30Q_{30} (see main text) appear to indicate that full CN thermalization and “memory loss” has not been achieved at t=2,482t = 2,482 fm/c.

Image not extracted

Figure 7. The total nucleon density profiles of the emerged CN with 2D density profiles in the insets. The emerged CN average atomic number is ⟨A⟩≈245.3\langle A\rangle\approx245.3.

References

  1. [61]sion processes in the ^{64}Ni + ^{238}U reaction,” Phys. Rev. C 93, 054616 (2016).
  2. [62]C. Simenel, A.S. Umar, K. Godbey, and P. McGlynn, “Shell effects in quasi-fission for calcium induced reactions forming thorium isotopes,” Phys. Lett. B 871, 139955 (2025).arxiv.org/abs/2507.00450
  3. [63]C. Simenel, “Nuclear quantum many-body dynamics (2nd edition),” Eur. Phys. J A 61, 181 (2025).
  4. [64]Z. Wu, X.-X. Sun, and L. Guo, “Impact of octupole correlation on the inverse quasifission in ^{160}Gd + ^{186}W collisions,” Phys. Rev. C 113, 024618 (2026).DOI
  5. [65]R. E. Peierls and J. Yoccoz, “The collective model of nuclear motion,” Proc. Phys. Soc. A 70, 381 (1957).
  6. [66]J. J. Griffin and J. A. Wheeler, “Collective Motions in Nuclei by the Method of Generator Coordinates,” Phys. Rev. 108, 311 (1957).DOI
  7. [67]P.-G. Reinhard, R.Y. Cusson, and K. Goeke, “Time evolution of coherent ground-state correlations and the TDHF approach,” Nucl. Phys. A 398, 141 (1983).DOI
  8. [68]P. J. Reinhard and K. Goeke, “The generator coordinate method and quantised collective motion in nuclear systems,” Rep. Prog. Phys. 50, 1 (1987).DOI
  9. [69]M. Verrière, N. Dubray, N. Schunck, D. Regnier, and P. Dossantos-Uzarralde, “Fission description: First steps towards a full resolution of the time-dependent Hill-Wheeler equation,” EPJ Web Conf. 146, 04034 (2017).DOI
  10. [70]N. Bohr and J. A. Wheeler, “The Mechanism of Nuclear Fission,” Phys. Rev. 56, 426 (1939).DOI
  11. [71]D. L. Hill and J. A. Wheeler, “Nuclear Constitution and the Interpretation of Fission Phenomena,” Phys. Rev. 89, 1102 (1953).DOI
  12. [72]A. Bulgac, P. Magierski, K. J. Roche, and I. Stetcu, “Induced Fission of ^{240}Pu within a Real-Time Microscopic Framework,” Phys. Rev. Lett. 116, 122504 (2016).DOI
  13. [73]A. Bulgac, S. Jin, K. J. Roche, N. Schunck, and I. Stetcu, “Fission dynamics of ^{240}Pu from saddle to scission and beyond,” Phys. Rev. C 100, 034615 (2019).DOI
  14. [74]A. Bulgac, S. Jin, and I. Stetcu, “Nuclear Fission Dynamics: Past, Present, Needs, and Future,” Frontiers in Physics 8, 63 (2020).DOI
  15. [75]M. Bender et al., “Future of nuclear fission theory,” J. Phys. G: Nucl. Part. Phys. 47, 113002 (2020).DOI
  16. [76]R. P. Feynman, “Space-Time Approach to Non-Relativistic Quantum Mechanics,” Rev. Mod. Phys. 20, 367 (1948).
  17. [77]A. Bulgac, “Pure quantum extension of the semiclassical Boltzmann-Uehling-Uhlenbeck equation,” Phys. Rev. C 105, L026101 (2022).DOI
  18. [78]A. Bulgac, M. Kafker, and I. Abdurrahman, “Measures of complexity and entanglement in many-fermion systems,” Phys. Rev. C 107, 044318 (2023).DOI
  19. [79]A. Bulgac, M. Kafker, I. Abdurrahman, and I. Stetcu, “Non-Markovian character and irreversibility of real-time quantum many-body dynamics,” Phys. Rev. C 109, 064617 (2024).DOI
  20. [80]C. E. P. Robin and M. J. Savage, “Quantum Complexity and New Directions in Nuclear Physics and High-Energy Physics Phenomenology,” (2026), arXiv:2604.26376 [quant-ph].arxiv.org/abs/2604.26376
  21. [81]E.V. Prokhorova and others, “The fusion–fission and quasi-fission processes in the reaction ^{48}Ca + ^{208}Pb at energies near the Coulomb barrier,” Nucl. Phys. A 802, 45 (2008).
  22. [82]S. Jin, K. J. Roche, I. Stetcu, I. Abdurrahman, and A. Bulgac, “The LISE package: solvers for static and time-dependent superfluid local density approximation equations in three dimensions,” Comp. Phys. Comm. 269, 108130 (2021).DOI
  23. [83]N. W. Aschroft and N. D. Mermin, Solid State Physics (Saunders College, 1976).
  24. [84]A. Bulgac and M. M. Forbes, “Use of the discrete variable representation basis in nuclear physics,” Phys. Rev. C 87, 051301(R) (2013).
  25. [85]C.E. Shannon, “Communication in the Presence of Noise,” Proceedings of the IRE 37, 10 (1949).
  26. [86]D. Regnier, N. Dubray, M. Verrière, and N. Schunck, “FELIX-2.0: New version of the finite element solver for the time dependent generator coordinate method with the Gaussian overlap approximation,” Comp. Phys. Comm. 225, 180 (2018).
  27. [87]D. Regnier, N. Dubray, N. Schunck, and M. Verrière, “Fission fragment charge and mass distributions in ^{239}Pu(n, f) in the adiabatic nuclear energy density functional theory,” Phys. Rev. C 93, 054611 (2016).DOI
  28. [88]M. Verriere and D. Regnier, “The Time-Dependent Generator Coordinate Method in Nuclear Physics,” Front. Phys. 8, 233 (2020).DOI
  29. [89]M. Verriere, N. Schunck, and D. Regnier, “Microscopic calculation of fission product yields with particle-number projection,” Phys. Rev. C 103, 054602 (2021).DOI
  30. [90]A. Bulgac, “Projection of good quantum numbers for reaction fragments,” Phys. Rev. C 100, 034612 (2019).DOI
  31. [91]D. Regnier and D. Lacroix, “Microscopic description of pair transfer between two superfluid Fermi systems. II. Quantum mixing of time-dependent Hartree-Fock-Bogolyubov trajectories,” Phys. Rev. C 99, 064615 (2019).DOI
  32. [92]N. Hasegawa, K. Hagino, and Y. Tanimura, “Time-dependent generator coordinate method for many-particle tunneling,” Phys. Lett. B 808, 135693 (2020).arxiv.org/abs/2006.06944
  33. [93]P. Marevic, D. Regnier, and D. Lacroix, “Quantum fluctuations induce collective multiphonons in finite fermi liquids,” Phys. Rev. C 108, 014620 (2023).DOI
  34. [94]P. Marevic, D. Regnier, and D. Lacroix, “Multiconfigurational time-dependent density functional theory for atomic nuclei: technical and numerical aspects,” Eur. Phys. A 60, 10 (2024).DOI
  35. [95]B. Li, D. Vretenar, T. Nikšić, P. W. Zhao, and J. Meng, “Generalized time-dependent generator coordinate method for small- and large-amplitude collective motion,” Phys. Rev. C 108, 014321 (2023).DOI
  36. [96]B. Li, D. Vretenar, T. Nikšić, J. Zhao, P. W. Zhao, and J. Meng, “Generalized time-dependent generator coordinate method for induced fission dynamics,” Front. Phys. 19, 44201 (2024).DOI
  37. [97]A. Bulgac, M. M. Forbes, and W. Wlazlowski, “Towards quantum turbulence in cold atomic fermionic superfluids,” J. Phys. B: At. Mol. Opt. Phys. 50, 014001 (2016).DOI
  38. [98]A. Bulgac, “Time-Dependent Density Functional Theory for Fermionic Superfluids: from Cold Atomic gases, to Nuclei and Neutron Star Crust,” Physica Status Solidi B 256, 1800592 (2019).DOI
  39. [99]A. Bulgac, I. Abdurrahman, and G. Wlazłowski, “Sensitivity of time-dependent density functional theory to initial conditions,” Phys. Rev. C 105, 044601 (2022).DOI
  40. [100]M. Arndt, O. Nairz, J. Vos-Andreae, C. Keller,

Paper details

vlnc:W_txz7uvtn
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 11

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents