Yue Shi, J. Dobaczewski, S. Frauendorf, W. Nazarewicz, J.C. Pei, F.R. Xu, N. Nikolov
January 4, 2012
Stimulated by recent experimental discoveries, triaxial strongly deformed (TSD) states in Er at ultrahigh spins have been studied by means of the SkyrmeHartreeFock model and the tiltedaxiscranking method. Restricting the rotational axis to one of the principal axes  as done in previous cranking calculations  two welldefined TSD minima in the total Routhian surface are found for a given configuration: one with positive and another with negative triaxial deformation . By allowing the rotational axis to change direction, the higherenergy minimum is shown to be a saddle point. This resolves the longstanding question of the physical interpretation of the two triaxial minima at a very similar quadrupole shape obtained in the principal axis cranking approach. Several TSD configurations have been predicted, including a highly deformed band expected to cross lesser elongated TSD bands at the highest spins. Its transitional quadrupole moment eb is close to the measured value of 11eb; hence, it is a candidate for the structure observed in experiment.
While the majority of nuclei have axially symmetric shapes, evidence for triaxial nuclear deformations has been elusive. The clearest signatures come from the gammaray spectroscopy of rotating nuclei. The deformation of a quantum object, such as molecule or atomic nucleus, enables the system to specify an orientation. The quantized motion of this degree of freedom generates the sequences of rotational levels  the rotational bands [1,2]. If the system is triaxial, the associated rotational bands show specific features that allow for distinguishing it from an axial one. In the case of nuclei, the appearance of the wobbling [1,3,4] and spinchirality [5,2] rotational modes are experimental signatures of triaxiality.
Triaxial shapes are expected to appear more frequently at high spin because of the tendency of aligned high quasiparticles to drive rotating nuclei towards triaxiality due to their spatial density distributions [6,7]. In addition, pairing correlations  which generally favor more symmetric shapes  are quenched at high spins and enhance the high alignment effect [8]. Consequently, with increasing spin, nuclei are predicted to go through nonaxial shapes before they eventually fission (see, e.g., [9]).
Recent experiments [10,11,12,13,14,15,16] have reached ultrahigh spins of about in nuclei around Er. It has been observed that with increasing angular momentum, the rotational bands terminate and nuclei assume weakly deformed oblate shapes, as evidenced by the irregular level spacings. At ultrahigh spins they return to collective rotation characterized by regular rotational bands, consistent with the early prediction [17]. Cranked NilssonStrutinsky (CNS) calculations suggest that the observed bands in Er are based on one of the three triaxial strongly deformed (TSD) minima in the potentialenergy surface (see Ref. [16] and references cited therein). For the lowest minimum TSD1, with positive , the calculated value of the transitional quadrupole moment eb considerably underestimates the observed value of eb [16]. This has led to the suggestion that the observed band in Er may be associated with either the minimum TSD2, which has a similar quadrupole deformation parameter as TSD1 but negative , or with the band TSD3, which has a larger triaxial deformation [16].
Most of the existing highspin calculations in the mass160 region assume that the axis of rotation coincides with one of the principal axes of the triaxial potential, which is commonly referred to as principalaxis cranking (PAC). The CNS calculations use the microscopicmacroscopic method, which combines a shell correction derived from a phenomenological potential with the deformation energy of a rotating liquid drop [18,19,17]. It is common to choose the axis as the rotational axis and let the triaxiality parameter cover the range . In the Lund convention, which we adopt in this Letter, the three sectors , , and represent the same triaxial shapes but represent rotation about the long, medium and short axis, respectively. The TSD1 and TSD2 minima in the CNS calculations [16] correspond to similar values of and which means that their shapes are nearly the same. The opposite sign of means that TSD1 rotates about the short axis and TSD2 about the medium axis. This raises the question of their physical interpretation, e.g., whether the higher of the two minima obtained in PAC is stable with respect to a reorientation of the rotational axis.
In this Letter we address this question by means of the tilted axis cranking (TAC) method [20,21], which considers the general orientation of the axis of rotation with respect to the principal axes of the nuclear quadrupole moment. We investigate the structure of ultrahighspin TSD minima in Er by using two approaches: the shellcorrection tilted axis cranking method (SCTAC) [20], which is based on the phenomenological Nilsson potential, and, for the first time, the threedimensional selfconsistent SkyrmeHartreeFock version of tilted axis cranking (SHFTAC) developed in Ref. [22]. Employing a fully selfconsistent rotating mean field  including the full rotational response due to cranking  is expected to improve the reliability of calculations in the realm of ultrahighspin states.
SHFTAC is based on the symmetryunrestricted solver HFODD (v2.49s) [23], which has been successfully applied to the description of chiral bands in La [22]. In the particlehole channel, we use the Skyrme energy density functionals SkM [24] and SLy4 [25], the latter of which has been supplemented with Landau parameters (SLy4) [26,27]. The total energy of the system is obtained by integrating the total energy density over spatial coordinates. We have used 1,000 deformed harmonic oscillator basis states with MeV and MeV. At ultrahigh spins, pairing is negligible; hence, it has been ignored in SCTAC and SHFTAC. As discussed in earlier SHFPAC (a PAC limit of SHFTAC) calculations [7], quadrupole polarization at high spin  both axial and triaxial  is very well described by unpaired theory.


Figure 2 shows the results of the SCTAC calculations. If the axis of rotation agrees with one of the principal axes (PAC), SCTAC coincides essentially with the CNS of Ref. [18]. (In SCTAC the Strutinsky renormalization is only carried out for the nonrotational part of the Routhian, whereas in CNS the rotational energy is also renormalized.) The equilibrium deformation parameters in SCTAC for rotation about the short axis are , (lower minimum), and for rotation about the medium axis , (higher minimum), which are close to the ones found by means of the CNS for the TSD1 and TSD2 minima, respectively [16].
It is clearly seen in Fig. 2 that the PAC minimum at becomes a saddle if the rotational axis is allowed to tilt. The dashed lines show how the energies change in a smooth way when tilting the rotational axis from short ( ) to medium ( ) while keeping and constant. Thus one cannot associate TSD2 with the band observed in Er. In addition, the transition quadrupole moment for the stable minimum TSD1 is 8eb, which is too small as compared with the experimental value of 11eb (see also the discussion in [16]).
In the case of SCTAC, the tilt angle of the rotational axis is defined relative to one of the principal axes of the deformed potential in a straightforward way. In SHFTAC calculations, it must be introduced by means of a constraint on the orientation of along with the constraints on the orientation of the principal axis of the nucleus defined in terms of the total (mass) quadrupole moment . The conditions on the corresponding Lagrange multipliers have been derived by Kerman and Onishi (KO) [21]. We use the 2D counterpart of relation (3.6) of Ref. [21], which states that and are not parallel ( ) if the Routhian is not at a stationary point (e.g., PAC). By using the Augmented Lagrangian Method of the HFODD code [23], we have checked that the KO conditions are obeyed to a high precision for all angles . The resulting angles do not exceed 0.1, depending on configuration.
configuration  (eb)  
A:  7.6  
B:  7.7  
C:  7.4  
D:  10.7 


In Fig. 4 we allow the rotational axis to tilt () by starting from SHFPAC solutions with b. [A rotation of this shape around the axis ( ) is equivalent to that of b around the axis ( ).] It can be seen that for the configuration A, the minimum that appears in Fig. 3 at b ( ) is unstable with respect to a reorientation of the rotational axis, that is, it represents the saddlepoint. On the other hand, the lower minimum, at , remains stable. This is consistent with SCTAC calculations of Fig. 2. A similar situation is predicted for configurations C and D. At =0.5MeV, the configuration B has a minimum at (), but it becomes unstable at higher rotational frequencies and a minimum at develops at =0.7MeV. This interesting behavior, together with a discussion of wobbling modes in bands AD will be discussed in detail in a forthcoming paper.
As seen in Table 1, while bands B and D have the same paritysignature occupations, configuration D has much larger deformation: b, b ( ).

Band D is our best candidate for the structure observed in experiment: the calculated eb of this configuration well reproduces the experimental value of 11eb [16]. (We note that band D has a larger quadrupole moment and smaller than band TSD3 of Ref. [16].) While the energyspin relations of Fig. 5 are not inconsistent with the spin estimates of Ref. [10], the fact that the TSD1 bands A and C do not seem to be seen experimentally, and that the experimental intensity pattern shows an increase in side feeding all the way to the point of decayout, both suggest that band D should be more favored energetically than predicted. Indeed, we do not expect the relative energies calculated in SHF using current functionals to be precise, as evidenced by appreciable differences between SkM and SLy4 predictions in Fig. 5.
Band  (MeV)  (eb)  (eb)  (eb)  () 
0.50  9.0  1.9  7.9  45.1  
0.60  8.7  2.0  7.6  
0.70  8.4  2.0  7.2  
0.80  8.1  2.1  6.9  
0.40  11.8  2.3  10.5  31.5  
0.50  12.2  2.4  10.8  
0.60  12.1  2.5  10.7  
0.70  12.1  2.5  10.6  
0.80  12.0  2.5  10.5 
In summary, we have performed, for the first time, Tilted Axis Cranking calculations within the selfconsistent SkyrmeHartreeFock model in which the KermanOnishi conditions for triaxial rotation are strictly obeyed. To address the recent puzzling experimental data, we studied the nucleus Er at ultrahigh spins. Restricting the direction of the rotational axis to one of the principal axes of the density distribution yields two TSD minima with similar values but with positive and negative deformations, similar to our SCTAC predictions and the previous CNS calculations [16]. Allowing the rotational axis to tilt away from the principal axes shows, however, that the higherenergy minimum is actually a saddle point; hence, it cannot be associated with a physical state. It is the lowerenergy minimum that represents a TSD band. We have thus clarified a longstanding question pertaining to the nature of positive and negative bands associated with the same intrinsic shape in the PAC approach: the rotation of a welldeformed, slightly triaxial configuration can be either about a short or medium axis, but not about both.
Several TSD configurations differing by proton and neutron occupations and quadrupole moments have been investigated. In the angular momentum range of 5070 , they are predicted to have transition quadrupole moments of 78eb, which are below the measured values of eb [16]. We have identified an excited TSD configuration, band D, with a stable positive minimum, which has a large transition quadrupole moment of eb that agrees well with the experimental value. At spins higher than , this band  different from structures TSD3 and SD of Ref. [16]  lies close to the less deformed TSD bands, and it is expected to become yrast above . The experimental intensity pattern suggests that band D should lie lower in energy than predicted by SkM and SLy4 models used in this study; this opens up an interesting direction for future investigations aiming at developing the spectroscopicquality nuclear energy density functional.
Pertinent and stimulating questions by Mark Riley, and numerous valuable discussions with him, are gratefully acknowledged. This work has been supported by the Natural Science Foundation of China under Grants Nos. 10735010 and 10975006; U.S. Department of Energy under Contract Nos. DEFG0296ER40963 (University of Tennessee) and DEFG0295ER40934 (University of Notre Dame); Academy of Finland and the University of Jyväskylä within the FIDIPRO programme.