Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Quantum entanglement in mixed-spin trimer: effects of a magnetic field and heterogeneous g𝑔gitalic_g-factors

Zhirayr Adamyan Laboratory of Theoretical Physics, Yerevan State University, 1 Alex Manoogian Str., 0025 Yerevan, Armenia CANDLE, Synchrotron Research Institute, 31 Acharyan Str., 0040 Yerevan, Armenia    Vadim Ohanyan Laboratory of Theoretical Physics, Yerevan State University, 1 Alex Manoogian Str., 0025 Yerevan, Armenia CANDLE, Synchrotron Research Institute, 31 Acharyan Str., 0040 Yerevan, Armenia
(May 17, 2024)
Abstract

Mixed spin-(1/2,1/2,1) trimer with two different LandΓ© g𝑔gitalic_g-factors and two different exchange couplings is considered. The main feature of the model is non-conserving magnetization. The Hamiltonian of the system is diagonalized analytically. We presented a detailed analysis of the ground state properties, revealing several possible ground state phase diagrams and magnetization profiles. The main focus is on how non-conserving magnetization affects quantum entanglement. We have found that non-conserving magnetization can bring to the continuous dependence of the entanglement quantifying parameter (negativity) on magnetic field within the same eigenstate, while for the case of uniform g𝑔gitalic_g-factors it is a constant. The main result is an essential enhancement of the entanglement in case of uniform couplings for one pair of spins caused by an arbitrary small difference in the values of g𝑔gitalic_g-factors. This enhancement is robust and brings to almost 7-fold increasing of the negativity. We have also found weakening of entanglement for other cases. Thus, non-conserving magnetization offers a broad opportunity to manipulate the entanglement by means of magnetic field.

Molecular magnets, Quantum entanglement, Non-conserving magnetization
pacs:
75.30 Kz, 75.75 -c, 75.10.Jm 03.76 Bg

I Introduction

Quantum entanglement, due to its critical role in quantum communication and information processing, has recently attracted more and more attention [1, 2, 3, 4, 5]. It is an essential ingredient of quantum teleportation [6, 7, 8, 9, 10, 11], quantum computing [12, 13, 14], and quantum cryptography [15, 16]. Moreover, comprehension of quantum entanglement has facilitated a progress in understanding black holes using quantum field theory methods. Entanglement also provides a new perspective for studying quantum phase transitions and collective phenomena in many-body systems and condensed matter physics. Recently, studying entanglement features of quantum spin clusters or molecular magnets [17, 18] has became very popular, as there are firm evidences that molecular magnets can be promising materials for realizing qubits for quantum information technologies [19, 20, 21]. In numerous papers, which have been published recently, various aspects of quantum entanglement in few-body spin systems have been figured out [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39].

There are different methods of quantifying quantum entanglement using special entanglement measures [40]. The entanglement measure, exploited in the present paper, is the so-called ”negativity” [41]. The principal advantage of negativity is its universality, i. e. possibility of its formulation and calculation in terms of the reduced two-particle density matrix independent of the physical nature of the system. Moreover, it can be efficiently exploited not only for the zero temperature entanglement (pure quantum entanglement) but, for the finite T𝑇Titalic_T as well (thermal entanglement).

The field of quantum molecular magnetism is rapidly advancing, with researchers pursuing both fundamental and practical applications of magnetic molecules and materials. This interdisciplinary field draws on physics, chemistry and nanotechnology, with specific applications in molecular spintronics, quantum technologies, metal-organic frameworks, and molecular spin qubits [17, 18, 19, 20, 21]. From formal point of view, molecular magnets can be considered as models of a few interacting spins (magnetic moments) with a variety of peculiar magneto-thermal properties. Such materials are usually referred to as zero-dimensional magnets. Some magnetic materials with structure of molecular crystal may also contain a one-dimensional arrays of interacting spins in the form of coordination polymers. Normally, magnetization processes of zero-dimensional magnets at low enough temperature reflect their finite spectrum in a straightforward way. A typical zero-temperature magnetization curve of the models of few interacting spins is a system of plateaus with jumps between them. Each plateau corresponds to the certain eigenstate with fixed value of magnetization, which is the ground state at given values of magnetic field. However, the magnetization profile can be changed crucially if the model possesses a non-conserving magnetization operator [27, 29, 30, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51]. In this case magnetization is not a good quantum number, thus, its value for a given eigenstate can be non-constant. This, in its turn, leads to a continuous dependance of the magnetization expectation value on the magnetic filed within the same eigenstate [46]. For the spin clusters or molecular magnets with non-conserving magnetization, in virtue of possible continuous dependence of magnetization within a single eigenstate, low-temperature magnetization curves profiles have a lot in common with magnetization processes in real many-body models. It is worth mentioning that for the systems with non-conserved z𝑧zitalic_z-component of total spin, St⁒o⁒tzsuperscriptsubscriptπ‘†π‘‘π‘œπ‘‘π‘§S_{tot}^{z}italic_S start_POSTSUBSCRIPT italic_t italic_o italic_t end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT, magnetization is non-conserving as well. However, even if [β„‹,St⁒o⁒tz]=0β„‹superscriptsubscriptπ‘†π‘‘π‘œπ‘‘π‘§0\left[\mathcal{H},S_{tot}^{z}\right]=0[ caligraphic_H , italic_S start_POSTSUBSCRIPT italic_t italic_o italic_t end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ] = 0 the magnetization can not commute with the system Hamiltonian, provided different spins have different g𝑔gitalic_g-factors [45, 46, 49, 50]. Even for the simplest isotropic model with the Hamiltonian

β„‹=βˆ‘i,jJi⁒j⁒𝐒i⁒𝐒jβˆ’B⁒ℳz,β„³z=βˆ‘igi⁒Sizformulae-sequenceβ„‹subscript𝑖𝑗subscript𝐽𝑖𝑗subscript𝐒𝑖subscript𝐒𝑗𝐡superscriptℳ𝑧superscriptℳ𝑧subscript𝑖subscript𝑔𝑖superscriptsubscript𝑆𝑖𝑧\displaystyle\mathcal{H}=\sum_{i,j}J_{ij}\mathbf{S}_{i}\mathbf{S}_{j}-B% \mathcal{M}^{z},\;\;\mathcal{M}^{z}=\sum_{i}g_{i}S_{i}^{z}caligraphic_H = βˆ‘ start_POSTSUBSCRIPT italic_i , italic_j end_POSTSUBSCRIPT italic_J start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT bold_S start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT bold_S start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT - italic_B caligraphic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT , caligraphic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = βˆ‘ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT italic_g start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT italic_S start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT (1)

one has

[β„‹,β„³z]=iβ’βˆ‘i,jJi⁒j⁒(giβˆ’gj)⁒(Six⁒Sjyβˆ’Siy⁒Six).β„‹superscriptℳ𝑧𝑖subscript𝑖𝑗subscript𝐽𝑖𝑗subscript𝑔𝑖subscript𝑔𝑗superscriptsubscript𝑆𝑖π‘₯superscriptsubscript𝑆𝑗𝑦superscriptsubscript𝑆𝑖𝑦superscriptsubscript𝑆𝑖π‘₯\displaystyle\left[\mathcal{H},\mathcal{M}^{z}\right]=i\sum_{i,j}J_{ij}(g_{i}-% g_{j})\left(S_{i}^{x}S_{j}^{y}-S_{i}^{y}S_{i}^{x}\right).[ caligraphic_H , caligraphic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ] = italic_i βˆ‘ start_POSTSUBSCRIPT italic_i , italic_j end_POSTSUBSCRIPT italic_J start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT ( italic_g start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT - italic_g start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT ) ( italic_S start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT italic_S start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT - italic_S start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT italic_S start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT ) . (2)

Thus, even if a single pair of spins in the lattice has giβ‰ gjsubscript𝑔𝑖subscript𝑔𝑗g_{i}\neq g_{j}italic_g start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT β‰  italic_g start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT, then the magnetization operator does not commute with the Hamiltonian.

In this work, we consider a mixed spin trimer with two spin-1/2 and one spin-1 magnetic ions, with two different exchange couplings and two different but isotropic LandΓ© g𝑔gitalic_g-factors. We consider the case when ion with spin 1 and one of the ions with spin 1/2 have g𝑔gitalic_g-factor equal to g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT while the second spin-1/2 ion has g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT (See Fig. 1). Also, the distribution of exchange couplings is a bit unusual, as we assume that the one of the ions with spin-1/2 interacts with spin-1 and another spin-1/2 with the same exchange constant, J1subscript𝐽1J_{1}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. The main reason for such configuration of the system is its highest possible symmetry, facilitating the analytical finding of the Hamiltonian eigenvalues and eigenstates. The paper deals with the problem of finding the optimal parameters to create entangled states and control them by external magnetic field. The main goal is to figure out how non-conserving magnetization affects entanglement measure like negativity. For all quantities of interest analytical consideration is possible. Our main finding is that inhomogeneous g𝑔gitalic_g-factors for appropriate values of other parameters lead to enhancement of entanglement with respect to the case when all spins are taken to have the same g𝑔gitalic_g-factors.

The paper is organized as follows, in the Second section we formulate the model and present its exact spectrum and eigenstates. The next Third section is devoted to zero-temperature phase diagrams and magneto-thermal properties of the system. In section IV we calculate quantum (T=0𝑇0T=0italic_T = 0) negativity and present the plots of its magnetic field behaviour. The paper ends with Conclusion.

II The Model and its exact solution

Let us consider a model of trimetallic heterogeneous molecular magnet with spins 1/2,1/212121/2,1/21 / 2 , 1 / 2 and 1111 in triangular geometry. We also suppose the exchange coupling between two spin-1/2121/21 / 2 ions, J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT, is a different from the other two, J1subscript𝐽1J_{1}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. We focuss on the effects of non-conserving magnetization, caused by the non-uniformity of g𝑔gitalic_g-factors. Therefore, the model under consideration includes two magnetic ions with g𝑔gitalic_g-factors equal to g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT (one of the spin-1/2 ions and spin-1 ion) and another spin-1/2 ion with g𝑔gitalic_g-factor equal to g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT (See Fig. 1).

Refer to caption
Figure 1: Symbolic picture of triangular trinuclear mixed-spin magnetic molecule with spins 1/2, 1/2 (small balls), and 1 (large ball). The LangΓ¨ g𝑔gitalic_g-factors for one spin-1/2 ion and spin-1 ion are supposed to be the same (g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT), whereas the g𝑔gitalic_g-factor for the second spin-1/2 is different from them, g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT.

The Hamiltonian of the model has the following form:

β„‹=J1⁒(𝒔1⁒𝒔2+𝒔2⁒𝑺3)+J2⁒𝒔1⁒𝑺3βˆ’B⁒(g1⁒s1z+g2⁒s2z+g1⁒S3z)β„‹subscript𝐽1subscript𝒔1subscript𝒔2subscript𝒔2subscript𝑺3subscript𝐽2subscript𝒔1subscript𝑺3𝐡subscript𝑔1superscriptsubscript𝑠1𝑧subscript𝑔2superscriptsubscript𝑠2𝑧subscript𝑔1superscriptsubscript𝑆3𝑧\mathcal{H}=J_{1}\left(\bm{s}_{1}\bm{s}_{2}+\bm{s}_{2}\bm{S}_{3}\right)+J_{2}% \bm{s}_{1}\bm{S}_{3}-B\left(g_{1}s_{1}^{z}+g_{2}s_{2}^{z}+g_{1}S_{3}^{z}\right)caligraphic_H = italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ( bold_italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT bold_italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT + bold_italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ) + italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT bold_italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT - italic_B ( italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ) (3)

where 𝒔asubscriptπ’”π‘Ž\bm{s}_{a}bold_italic_s start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT, a=1,2π‘Ž12a=1,2italic_a = 1 , 2, are spin-1/2 operators, and 𝑺3subscript𝑺3\bm{S}_{3}bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT stands for spin-1 operators. Exchange interaction between all pairs of spins is supposed to be isotropic. Due to presence of magnetic ions with two different g𝑔gitalic_g-factors, the magnetization operator does not commute with the Hamiltonian,

[β„‹,Mz]=i⁒J1⁒gβˆ’β’(s1x⁒s2yβˆ’s1y⁒s2x+s2y⁒S3xβˆ’s2x⁒S3y),β„‹superscript𝑀𝑧𝑖subscript𝐽1subscript𝑔superscriptsubscript𝑠1π‘₯superscriptsubscript𝑠2𝑦superscriptsubscript𝑠1𝑦superscriptsubscript𝑠2π‘₯superscriptsubscript𝑠2𝑦superscriptsubscript𝑆3π‘₯superscriptsubscript𝑠2π‘₯superscriptsubscript𝑆3𝑦\displaystyle[\mathcal{H},M^{z}]=iJ_{1}g_{-}\left(s_{1}^{x}s_{2}^{y}-s_{1}^{y}% s_{2}^{x}+s_{2}^{y}S_{3}^{x}-s_{2}^{x}S_{3}^{y}\right),[ caligraphic_H , italic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ] = italic_i italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT ( italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT - italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT + italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT - italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_y end_POSTSUPERSCRIPT ) ,
Mz=g1⁒s1z+g2⁒s2z+g1⁒S3z,superscript𝑀𝑧subscript𝑔1superscriptsubscript𝑠1𝑧subscript𝑔2superscriptsubscript𝑠2𝑧subscript𝑔1superscriptsubscript𝑆3𝑧\displaystyle M^{z}=g_{1}s_{1}^{z}+g_{2}s_{2}^{z}+g_{1}S_{3}^{z},italic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ,
gβˆ’=g1βˆ’g2.subscript𝑔subscript𝑔1subscript𝑔2\displaystyle g_{-}=g_{1}-g_{2}.italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT - italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT . (4)

However, St⁒o⁒tz=s1z+s2z+S3zsuperscriptsubscriptπ‘†π‘‘π‘œπ‘‘π‘§superscriptsubscript𝑠1𝑧superscriptsubscript𝑠2𝑧superscriptsubscript𝑆3𝑧S_{tot}^{z}=s_{1}^{z}+s_{2}^{z}+S_{3}^{z}italic_S start_POSTSUBSCRIPT italic_t italic_o italic_t end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT conserves. The Hamiltonian (3) can be diagonalized analytically. The twelve eigenvalues are

E1,2subscript𝐸12\displaystyle{E}_{1,2}italic_E start_POSTSUBSCRIPT 1 , 2 end_POSTSUBSCRIPT =\displaystyle== 14⁒(3⁒J1+2⁒J2βˆ“2⁒B⁒(3⁒g1+g2)),14minus-or-plus3subscript𝐽12subscript𝐽22𝐡3subscript𝑔1subscript𝑔2\displaystyle\frac{1}{4}\left(3J_{1}+2J_{2}\mp 2B(3g_{1}+g_{2})\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( 3 italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + 2 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT βˆ“ 2 italic_B ( 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ) ) ,
E3,4subscript𝐸34\displaystyle{E}_{3,4}italic_E start_POSTSUBSCRIPT 3 , 4 end_POSTSUBSCRIPT =\displaystyle== 14⁒(J1βˆ’4⁒J2βˆ“2⁒B⁒(g1+g2)),14minus-or-plussubscript𝐽14subscript𝐽22𝐡subscript𝑔1subscript𝑔2\displaystyle\frac{1}{4}\left(J_{1}-4J_{2}\mp 2B(g_{1}+g_{2})\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT - 4 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT βˆ“ 2 italic_B ( italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ) ) , (5)
E5,6subscript𝐸56\displaystyle{E}_{5,6}italic_E start_POSTSUBSCRIPT 5 , 6 end_POSTSUBSCRIPT =\displaystyle== 14⁒(βˆ’J1βˆ’4⁒J2βˆ“2⁒B2⁒gβˆ’2+J12),14minus-or-plussubscript𝐽14subscript𝐽22superscript𝐡2superscriptsubscript𝑔2superscriptsubscript𝐽12\displaystyle\frac{1}{4}\left(-J_{1}-4J_{2}\mp 2\sqrt{B^{2}g_{-}^{2}+J_{1}^{2}% }\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( - italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT - 4 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT βˆ“ 2 square-root start_ARG italic_B start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG ) ,
E7,8subscript𝐸78\displaystyle{E}_{7,8}italic_E start_POSTSUBSCRIPT 7 , 8 end_POSTSUBSCRIPT =\displaystyle== 14⁒(βˆ’J1+2⁒J2βˆ“2⁒B2⁒gβˆ’2+4⁒J12),14minus-or-plussubscript𝐽12subscript𝐽22superscript𝐡2superscriptsubscript𝑔24superscriptsubscript𝐽12\displaystyle\frac{1}{4}\left(-J_{1}+2J_{2}\mp 2\sqrt{B^{2}g_{-}^{2}+4J_{1}^{2% }}\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( - italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + 2 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT βˆ“ 2 square-root start_ARG italic_B start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG ) ,
E9,10subscript𝐸910\displaystyle{E}_{9,10}italic_E start_POSTSUBSCRIPT 9 , 10 end_POSTSUBSCRIPT =\displaystyle== 14⁒(βˆ’J1+2⁒J2βˆ’4⁒B⁒g1βˆ“2⁒Q+),14minus-or-plussubscript𝐽12subscript𝐽24𝐡subscript𝑔12superscript𝑄\displaystyle\frac{1}{4}\left(-J_{1}+2J_{2}-4Bg_{1}\mp 2Q^{+}\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( - italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + 2 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT - 4 italic_B italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT βˆ“ 2 italic_Q start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT ) ,
E11,12subscript𝐸1112\displaystyle{E}_{11,12}italic_E start_POSTSUBSCRIPT 11 , 12 end_POSTSUBSCRIPT =\displaystyle== 14⁒(βˆ’J1+2⁒J2+4⁒B⁒g1βˆ“2⁒Qβˆ’),14minus-or-plussubscript𝐽12subscript𝐽24𝐡subscript𝑔12superscript𝑄\displaystyle\frac{1}{4}\left(-J_{1}+2J_{2}+4Bg_{1}\mp 2Q^{-}\right),divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( - italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + 2 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT + 4 italic_B italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT βˆ“ 2 italic_Q start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT ) ,
where⁒QΒ±wheresuperscript𝑄plus-or-minus\displaystyle\mbox{where}\;\;\;Q^{\pm}where italic_Q start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT =\displaystyle== (B⁒gβˆ’Β±J1)2+3⁒J12.superscriptplus-or-minus𝐡subscript𝑔subscript𝐽123superscriptsubscript𝐽12\displaystyle\sqrt{(Bg_{-}\pm J_{1})^{2}+3J_{1}^{2}}.square-root start_ARG ( italic_B italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT Β± italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG .

Non-linear dependence of some eigenvalues on the magnetic field is the direct consequence of non-commutativity of the magnetization operator and the Hamiltonian. The corresponding eigenvectors respectively are

|ψ1,2⟩=|Β±12,Β±12,Β±1⟩,ketsubscriptπœ“12ketplus-or-minus12plus-or-minus12plus-or-minus1\displaystyle|\psi_{1,2}\rangle=\left|\pm\frac{1}{2},\pm\frac{1}{2},\pm 1% \right\rangle,| italic_ψ start_POSTSUBSCRIPT 1 , 2 end_POSTSUBSCRIPT ⟩ = | Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± 1 ⟩ , (6)
|ψ3,4⟩=13⁒(Β±2⁒|βˆ“12,Β±12,Β±1βŸ©βˆ“|Β±12,Β±12,0⟩),ketsubscriptπœ“3413minus-or-plusplus-or-minus2ketminus-or-plus12plus-or-minus12plus-or-minus1ketplus-or-minus12plus-or-minus120\displaystyle|\psi_{3,4}\rangle=\frac{1}{\sqrt{3}}\left(\pm\sqrt{2}\left|\mp% \frac{1}{2},\pm\frac{1}{2},\pm 1\right\rangle\mp\left|\pm\frac{1}{2},\pm\frac{% 1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 3 , 4 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 3 end_ARG end_ARG ( Β± square-root start_ARG 2 end_ARG | βˆ“ divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± 1 ⟩ βˆ“ | Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) ,
|ψ5,6⟩=13⁒(1+MΒ±2)⁒(M±⁒(2⁒|12,12,βˆ’1βŸ©βˆ’|βˆ’12,12,0⟩)+2⁒|βˆ’12,βˆ’12,1βŸ©βˆ’|12,βˆ’12,0⟩),ketsubscriptπœ“56131subscriptsuperscript𝑀2plus-or-minussuperscript𝑀plus-or-minus2ket12121ket121202ket12121ket12120\displaystyle|\psi_{5,6}\rangle=\frac{1}{\sqrt{3(1+M^{2}_{\pm})}}\left(M^{\pm}% \left(\sqrt{2}\left|\frac{1}{2},\frac{1}{2},-1\right\rangle-\left|-\frac{1}{2}% ,\frac{1}{2},0\right\rangle\right)+\sqrt{2}\left|-\frac{1}{2},-\frac{1}{2},1% \right\rangle-\left|\frac{1}{2},-\frac{1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 5 , 6 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 3 ( 1 + italic_M start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT ) end_ARG end_ARG ( italic_M start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT ( square-root start_ARG 2 end_ARG | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ - | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) + square-root start_ARG 2 end_ARG | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ - | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) ,
|ψ7,8⟩=13⁒(4+Kβˆ“2)⁒(Kβˆ“β’(|12,12,βˆ’1⟩+2⁒|βˆ’12,12,0⟩)+2⁒|βˆ’12,βˆ’12,1⟩+2⁒2⁒|12,βˆ’12,0⟩),ketsubscriptπœ“78134subscriptsuperscript𝐾2minus-or-plussuperscript𝐾minus-or-plusket121212ket121202ket1212122ket12120\displaystyle|\psi_{7,8}\rangle=\frac{1}{\sqrt{3(4+K^{2}_{\mp})}}\left(K^{\mp}% \left(\left|\frac{1}{2},\frac{1}{2},-1\right\rangle+\sqrt{2}\left|-\frac{1}{2}% ,\frac{1}{2},0\right\rangle\right)+2\left|-\frac{1}{2},-\frac{1}{2},1\right% \rangle+2\sqrt{2}\left|\frac{1}{2},-\frac{1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 7 , 8 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 3 ( 4 + italic_K start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT ) end_ARG end_ARG ( italic_K start_POSTSUPERSCRIPT βˆ“ end_POSTSUPERSCRIPT ( | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ + square-root start_ARG 2 end_ARG | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) + 2 | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ + 2 square-root start_ARG 2 end_ARG | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) ,
|ψ9,10⟩=13+GΒ±2⁒(2⁒|12,12,0βŸ©βˆ’G±⁒|12,βˆ’12,1⟩+|βˆ’12,12,1⟩),ketsubscriptπœ“91013subscriptsuperscript𝐺2plus-or-minus2ket12120superscript𝐺plus-or-minusket12121ket12121\displaystyle|\psi_{9,10}\rangle=\frac{1}{\sqrt{3+G^{2}_{\pm}}}\left(\sqrt{2}% \left|\frac{1}{2},\frac{1}{2},0\right\rangle-G^{\pm}\left|\frac{1}{2},-\frac{1% }{2},1\right\rangle+\left|-\frac{1}{2},\frac{1}{2},1\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 9 , 10 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 3 + italic_G start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT end_ARG end_ARG ( square-root start_ARG 2 end_ARG | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ - italic_G start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ + | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ ) ,
|ψ11,12⟩=13+Uβˆ“2⁒(|12,βˆ’12,βˆ’1⟩+Uβˆ“β’|βˆ’12,12,βˆ’1⟩+2⁒|βˆ’12,βˆ’12,0⟩),ketsubscriptπœ“111213subscriptsuperscriptπ‘ˆ2minus-or-plusket12121superscriptπ‘ˆminus-or-plusket121212ket12120\displaystyle|\psi_{11,12}\rangle=\frac{1}{\sqrt{3+U^{2}_{\mp}}}\left(\left|% \frac{1}{2},-\frac{1}{2},-1\right\rangle+U^{\mp}\left|-\frac{1}{2},\frac{1}{2}% ,-1\right\rangle+\sqrt{2}\left|-\frac{1}{2},-\frac{1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 11 , 12 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 3 + italic_U start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT end_ARG end_ARG ( | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ + italic_U start_POSTSUPERSCRIPT βˆ“ end_POSTSUPERSCRIPT | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ + square-root start_ARG 2 end_ARG | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) ,

where the following notations are adopted:

MΒ±=βˆ’B⁒gβˆ’Β±B2⁒gβˆ’2+J12J1,superscript𝑀plus-or-minusplus-or-minus𝐡subscript𝑔superscript𝐡2superscriptsubscript𝑔2superscriptsubscript𝐽12subscript𝐽1\displaystyle M^{\pm}=\frac{-Bg_{-}\pm\sqrt{B^{2}g_{-}^{2}+J_{1}^{2}}}{J_{1}},italic_M start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT = divide start_ARG - italic_B italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT Β± square-root start_ARG italic_B start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG end_ARG start_ARG italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT end_ARG , (7)
KΒ±=B⁒gβˆ’Β±B2⁒gβˆ’2+4⁒J122⁒J1,superscript𝐾plus-or-minusplus-or-minus𝐡subscript𝑔superscript𝐡2superscriptsubscript𝑔24superscriptsubscript𝐽122subscript𝐽1\displaystyle K^{\pm}=\frac{Bg_{-}\pm\sqrt{B^{2}g_{-}^{2}+4J_{1}^{2}}}{2J_{1}},italic_K start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT = divide start_ARG italic_B italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT Β± square-root start_ARG italic_B start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG end_ARG start_ARG 2 italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT end_ARG ,
GΒ±=B⁒gβˆ’+J1Β±Q+J1,superscript𝐺plus-or-minusplus-or-minus𝐡subscript𝑔subscript𝐽1superscript𝑄subscript𝐽1\displaystyle G^{\pm}=\frac{Bg_{-}+J_{1}\pm Q^{+}}{J_{1}},italic_G start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT = divide start_ARG italic_B italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT + italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT Β± italic_Q start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT end_ARG start_ARG italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT end_ARG ,
UΒ±=B⁒gβˆ’βˆ’J1Β±Qβˆ’J1,superscriptπ‘ˆplus-or-minusplus-or-minus𝐡subscript𝑔subscript𝐽1superscript𝑄subscript𝐽1\displaystyle U^{\pm}=\frac{Bg_{-}-J_{1}\pm Q^{-}}{J_{1}},italic_U start_POSTSUPERSCRIPT Β± end_POSTSUPERSCRIPT = divide start_ARG italic_B italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT - italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT Β± italic_Q start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT end_ARG start_ARG italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT end_ARG ,

and the basis vectors obey the standard notations, |s1z,s2z,S3z⟩ketsuperscriptsubscript𝑠1𝑧superscriptsubscript𝑠2𝑧superscriptsubscript𝑆3𝑧\left|s_{1}^{z},s_{2}^{z},S_{3}^{z}\right\rangle| italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT , italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT , italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT ⟩. As has been mentioned above, they are eigenvectors of St⁒o⁒tzsuperscriptsubscriptπ‘†π‘‘π‘œπ‘‘π‘§S_{tot}^{z}italic_S start_POSTSUBSCRIPT italic_t italic_o italic_t end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT operator as well, having eigenvalues form βˆ’22-2- 2 to 2222. Non-conserving magnetization operator leads to situation where only four eigenvectors of the system, |ψ1βŸ©βˆ’|ψ4⟩ketsubscriptπœ“1ketsubscriptπœ“4\left|\psi_{1}\right\rangle-\left|\psi_{4}\right\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ - | italic_ψ start_POSTSUBSCRIPT 4 end_POSTSUBSCRIPT ⟩, have fixed value of magnetic moment, in contrast to the others. However, the expectation value of the magnetization operator can be evaluated for each of them:

M1,2=Β±3⁒g1+g22,M3,4=Β±g1+g22,formulae-sequencesubscript𝑀12plus-or-minus3subscript𝑔1subscript𝑔22subscript𝑀34plus-or-minussubscript𝑔1subscript𝑔22\displaystyle M_{1,2}=\pm\frac{3g_{1}+g_{2}}{2},\;M_{3,4}=\pm\frac{g_{1}+g_{2}% }{2},italic_M start_POSTSUBSCRIPT 1 , 2 end_POSTSUBSCRIPT = Β± divide start_ARG 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT end_ARG start_ARG 2 end_ARG , italic_M start_POSTSUBSCRIPT 3 , 4 end_POSTSUBSCRIPT = Β± divide start_ARG italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT end_ARG start_ARG 2 end_ARG , (8)
M5,6=βˆ’(MΒ±2βˆ’1)⁒gβˆ’2⁒(MΒ±2+1),M7,8=βˆ’(Kβˆ“2βˆ’4)⁒gβˆ’2⁒(Kβˆ“2+4),formulae-sequencesubscript𝑀56superscriptsubscript𝑀plus-or-minus21subscript𝑔2superscriptsubscript𝑀plus-or-minus21subscript𝑀78superscriptsubscript𝐾minus-or-plus24subscript𝑔2superscriptsubscript𝐾minus-or-plus24\displaystyle M_{5,6}=-\frac{(M_{\pm}^{2}-1)g_{-}}{2(M_{\pm}^{2}+1)},\;M_{7,8}% =-\frac{(K_{\mp}^{2}-4)g_{-}}{2(K_{\mp}^{2}+4)},italic_M start_POSTSUBSCRIPT 5 , 6 end_POSTSUBSCRIPT = - divide start_ARG ( italic_M start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - 1 ) italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT end_ARG start_ARG 2 ( italic_M start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 ) end_ARG , italic_M start_POSTSUBSCRIPT 7 , 8 end_POSTSUBSCRIPT = - divide start_ARG ( italic_K start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - 4 ) italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT end_ARG start_ARG 2 ( italic_K start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 ) end_ARG ,
M9,10=3⁒g++GΒ±2⁒(3⁒g1βˆ’g2)2⁒(GΒ±2+3),subscript𝑀9103subscript𝑔superscriptsubscript𝐺plus-or-minus23subscript𝑔1subscript𝑔22superscriptsubscript𝐺plus-or-minus23\displaystyle M_{9,10}=\frac{3g_{+}+G_{\pm}^{2}(3g_{1}-g_{2})}{2(G_{\pm}^{2}+3% )},italic_M start_POSTSUBSCRIPT 9 , 10 end_POSTSUBSCRIPT = divide start_ARG 3 italic_g start_POSTSUBSCRIPT + end_POSTSUBSCRIPT + italic_G start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT ( 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT - italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ) end_ARG start_ARG 2 ( italic_G start_POSTSUBSCRIPT Β± end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 ) end_ARG ,
M11,12=βˆ’3⁒g++2⁒Uβˆ“2⁒(3⁒g1βˆ’g2)2⁒(2⁒Uβˆ“2+3),subscript𝑀11123subscript𝑔2superscriptsubscriptπ‘ˆminus-or-plus23subscript𝑔1subscript𝑔222superscriptsubscriptπ‘ˆminus-or-plus23\displaystyle M_{11,12}=-\frac{3g_{+}+2U_{\mp}^{2}(3g_{1}-g_{2})}{2(2U_{\mp}^{% 2}+3)},italic_M start_POSTSUBSCRIPT 11 , 12 end_POSTSUBSCRIPT = - divide start_ARG 3 italic_g start_POSTSUBSCRIPT + end_POSTSUBSCRIPT + 2 italic_U start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT ( 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT - italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ) end_ARG start_ARG 2 ( 2 italic_U start_POSTSUBSCRIPT βˆ“ end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 ) end_ARG ,
g+=g1+g2.subscript𝑔subscript𝑔1subscript𝑔2\displaystyle g_{+}=g_{1}+g_{2}.italic_g start_POSTSUBSCRIPT + end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT .

Due to non-commutativity of the magnetization operator and the Hamiltonian of the system majority of the eigenstates have magnetic field and J1subscript𝐽1J_{1}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT dependent expectation values of the magnetization [45, 46]. Interestingly, dependence on J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT does not occur. The way we introduced non-equal exchange interaction in the model is a special one. The part of the Hamiltonian which describes interaction between first and third spins turns to be conserved,

[β„‹,J2⁒𝒔1⁒𝑺3]=0.β„‹subscript𝐽2subscript𝒔1subscript𝑺30\displaystyle[\mathcal{H},J_{2}\bm{s}_{1}\bm{S}_{3}]=0.[ caligraphic_H , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT bold_italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ] = 0 . (9)

Thus, the spectrum of a system with non-zero J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and the spectrum of the system with J2=0subscript𝐽20J_{2}=0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 0 differ from each other by the constant shift proportional to J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. Therefore, magneto-thermal properties of the system in triangular shape (J2β‰ 0subscript𝐽20J_{2}\neq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β‰  0) and in linear geometry (J2=0subscript𝐽20J_{2}=0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 0) are to a great extent identical. Interestingly, not all eigenvectors from Eq.(6) are continuous under the limit restoring conservation of magnetization, g2β†’g1β†’subscript𝑔2subscript𝑔1g_{2}\rightarrow g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β†’ italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. This is a quite common situation in the spin clusters with non-uniform g𝑔gitalic_g-factors [46]. In case of our model six eigenvectors, corresponding to Sz=0,Β±1superscript𝑆𝑧0plus-or-minus1S^{z}=0,\pm 1italic_S start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = 0 , Β± 1 do not change continuously under the limit of uniform g𝑔gitalic_g-factors. For the triangular mixed spin cluster with g1=g2subscript𝑔1subscript𝑔2g_{1}=g_{2}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT the counterparts of eigenvectors |ψ3,4⟩ketsubscriptπœ“34|\psi_{3,4}\rangle| italic_ψ start_POSTSUBSCRIPT 3 , 4 end_POSTSUBSCRIPT ⟩, |ψ6⟩ketsubscriptπœ“6|\psi_{6}\rangle| italic_ψ start_POSTSUBSCRIPT 6 end_POSTSUBSCRIPT ⟩, |ψ7⟩ketsubscriptπœ“7|\psi_{7}\rangle| italic_ψ start_POSTSUBSCRIPT 7 end_POSTSUBSCRIPT ⟩ |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ and |ψ11⟩ketsubscriptπœ“11|\psi_{11}\rangle| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ have the following form:

|ψ3,4⟩0=12⁒(|12,βˆ’12,Β±1⟩+|βˆ’12,12,Β±1βŸ©βˆ’2⁒|Β±12,Β±12,0⟩),subscriptketsubscriptπœ“34012ket1212plus-or-minus1ket1212plus-or-minus12ketplus-or-minus12plus-or-minus120\displaystyle|\psi_{3,4}\rangle_{0}=\frac{1}{2}\left(\left|\frac{1}{2},-\frac{% 1}{2},\pm 1\right\rangle+\left|-\frac{1}{2},\frac{1}{2},\pm 1\right\rangle-% \sqrt{2}\left|\pm\frac{1}{2},\pm\frac{1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 3 , 4 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG 2 end_ARG ( | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± 1 ⟩ + | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± 1 ⟩ - square-root start_ARG 2 end_ARG | Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , Β± divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) , (10)
|ψ6⟩=12⁒(|βˆ’12,βˆ’12,1βŸ©βˆ’|12,12,βˆ’1⟩),ketsubscriptπœ“612ket12121ket12121\displaystyle|\psi_{6}\rangle=\frac{1}{\sqrt{2}}\left(\left|-\frac{1}{2},-% \frac{1}{2},1\right\rangle-\left|\frac{1}{2},\frac{1}{2},-1\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 6 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 2 end_ARG end_ARG ( | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ - | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ ) ,
|ψ7⟩=12⁒(|βˆ’12,12,0βŸ©βˆ’|12,βˆ’12,0⟩),ketsubscriptπœ“712ket12120ket12120\displaystyle|\psi_{7}\rangle=\frac{1}{\sqrt{2}}\left(\left|-\frac{1}{2},\frac% {1}{2},0\right\rangle-\left|\frac{1}{2},-\frac{1}{2},0\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 7 end_POSTSUBSCRIPT ⟩ = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 2 end_ARG end_ARG ( | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ - | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 0 ⟩ ) ,
|ψ9⟩0=12⁒(|βˆ’12,12,1βŸ©βˆ’|12,βˆ’12,1⟩),subscriptketsubscriptπœ“9012ket12121ket12121\displaystyle|\psi_{9}\rangle_{0}=\frac{1}{\sqrt{2}}\left(\left|-\frac{1}{2},% \frac{1}{2},1\right\rangle-\left|\frac{1}{2},-\frac{1}{2},1\right\rangle\right),| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 2 end_ARG end_ARG ( | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ - | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , 1 ⟩ ) ,
|ψ11⟩0=12⁒(|βˆ’12,12,βˆ’1βŸ©βˆ’|12,βˆ’12,βˆ’1⟩).subscriptketsubscriptπœ“11012ket12121ket12121\displaystyle|\psi_{11}\rangle_{0}=\frac{1}{\sqrt{2}}\left(\left|-\frac{1}{2},% \frac{1}{2},-1\right\rangle-\left|\frac{1}{2},-\frac{1}{2},-1\right\rangle% \right).| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG square-root start_ARG 2 end_ARG end_ARG ( | - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ - | divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - divide start_ARG 1 end_ARG start_ARG 2 end_ARG , - 1 ⟩ ) .

The rest of the eigenvectors change continuously under g2β†’g1β†’subscript𝑔2subscript𝑔1g_{2}\rightarrow g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β†’ italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT.

Last but not least, unusual choice of the interaction constants distribution have to be explained. This is quite an extraordinary suggestion that the exchange interaction constant between two spin-1/2 ions are the same as between spin-1 and spin-1/2 ions. It is worth mentioning, that in principle this can be the case, as least within certain numerical range. Thus, speaking about possible experimental realization of the model, one should have in mind numerical proximity of two exchange constants. From the practical point of view, we aimed to have the model which admits analytical solution in a possible simple way. Analysing all possible combinations of g𝑔gitalic_g-factors and coupling we arrive at the model given by the Hamiltonian (3), which due to conservation of St⁒o⁒tzsuperscriptsubscriptπ‘†π‘‘π‘œπ‘‘π‘§S_{tot}^{z}italic_S start_POSTSUBSCRIPT italic_t italic_o italic_t end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT, J2⁒𝒔1⁒𝑺3subscript𝐽2subscript𝒔1subscript𝑺3J_{2}\bm{s}_{1}\bm{S}_{3}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT bold_italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT and discrete symmetry 𝒔1↔𝑺3↔subscript𝒔1subscript𝑺3\bm{s}_{1}\leftrightarrow\bm{S}_{3}bold_italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ↔ bold_italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT have block-diagonal form with four one-dimensional and four two-dimensional blocks. This is the most simple Hamiltonian among all possible (1/2,1/2,1)12121(1/2,1/2,1)( 1 / 2 , 1 / 2 , 1 ) spin mixed clusters with two g𝑔gitalic_g-factors.

III Ground states phase diagrams and magnetization processes

As the model under consideration has many parameters and possesses non-conserving magnetization operator, the variety of possible ground state phase diagrams is quite large. First of all, one should distinguish the case of ferromagnetic, J1<0,J2<0formulae-sequencesubscript𝐽10subscript𝐽20J_{1}<0,J_{2}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0, antiferromagnetic, J1>0,J2>0formulae-sequencesubscript𝐽10subscript𝐽20J_{1}>0,J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 and mixed J1⁒J2<0subscript𝐽1subscript𝐽20J_{1}J_{2}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0 couplings. For the case of zero magnetic field the ground states phase diagram exhibits large degeneracy, except for the region where |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ is a ground state (See Fig. 2). Obviously, the ground states and their properties are highly affected by the g𝑔gitalic_g-factors difference, gβˆ’subscript𝑔g_{-}italic_g start_POSTSUBSCRIPT - end_POSTSUBSCRIPT. Therefore, in this section we present the ground-state phase diagrams in (g2/g1,B⁒g1/J1)subscript𝑔2subscript𝑔1𝐡subscript𝑔1subscript𝐽1(g_{2}/g_{1},Bg_{1}/J_{1})( italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT / italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT , italic_B italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT / italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ) or (g2/g1,B⁒g1/J2)subscript𝑔2subscript𝑔1𝐡subscript𝑔1subscript𝐽2(g_{2}/g_{1},Bg_{1}/J_{2})( italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT / italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT , italic_B italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT / italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ) planes, depending on which coupling constant we set as an energy unit. Without loss of generality one can consider five different types of ground state phase diagrams, given by the following conditions for the coupling constants: purely ferromagnetic, J1=βˆ’1subscript𝐽11J_{1}=-1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1, J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0 (Fig. 3); purely antiferromagnetic with equal couplings, J1=J2=1subscript𝐽1subscript𝐽21J_{1}=J_{2}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 (Fig. 5); purely antiferromagnetic with non-equal couplings, J2>0,J1>0formulae-sequencesubscript𝐽20subscript𝐽10J_{2}>0,J_{1}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 , italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0 (Fig. 7); two cases of mixed coupling, J1=1subscript𝐽11J_{1}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1, J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0 (Fig. 9) and J1<0subscript𝐽10J_{1}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0, J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 (Fig. 11).

Refer to caption
Figure 2: Zero-field ground-states phase diagram.
Refer to caption
Figure 3: Ground states phase diagram for J1=βˆ’1subscript𝐽11J_{1}=-1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1 and arbitrary J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0.
Refer to caption
Figure 4: Magnetization processed for J1=βˆ’1subscript𝐽11J_{1}=-1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1 and arbitrary J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0 for T/|J1|=0.00001𝑇subscript𝐽10.00001T/|J_{1}|=0.00001italic_T / | italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT | = 0.00001 and g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 for different negative values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT: g2=βˆ’5subscript𝑔25g_{2}=-5italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 5 (solid); g2=βˆ’4subscript𝑔24g_{2}=-4italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 4 (dashed) and g2=βˆ’3subscript𝑔23g_{2}=-3italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 3 (dot-dashed). The value of the saturated magnetization, M0subscript𝑀0M_{0}italic_M start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT, is 3⁒g1+|g2|23subscript𝑔1subscript𝑔22\frac{3g_{1}+|g_{2}|}{2}divide start_ARG 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + | italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT | end_ARG start_ARG 2 end_ARG.

III.0.1 Purely ferromagnetic couplings

For the system under consideration three regions of ferromagnetic parameters can be specified: J1=J2<0subscript𝐽1subscript𝐽20J_{1}=J_{2}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0, J1<0,J2<0,J1β‰ J2formulae-sequencesubscript𝐽10formulae-sequencesubscript𝐽20subscript𝐽1subscript𝐽2J_{1}<0,J_{2}<0,J_{1}\neq J_{2}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT β‰  italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and J1<0,J2=0formulae-sequencesubscript𝐽10subscript𝐽20J_{1}<0,J_{2}=0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 0. The former one corresponds to the case of linear spin cluster, which can be described as an limiting case of our model. Interestingly, due to the commutativity of the J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT-term with the Hamiltonian (Eq. (9)), the ground state energies for all those cases differ from each other only by the same constant value 2⁒J22subscript𝐽22J_{2}2 italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT, therefore all ground state phase diagram for ferromagnetic coupling have the same structure, presented in the Fig. 3. To have non-trivial eigenstates in addition to the fully polarized one, |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩, we have to investigate the region of negative g𝑔gitalic_g-factors. Speaking about negative g𝑔gitalic_g-factors one should bear in mind that this situation can be relevant rather for pseudospin [52], mainly for the magnetic ions with large values of orbital moment and strong spin-orbit coupling. However, magnetic materials with negative or mixed g𝑔gitalic_g-factors are not very exotic [48, 50, 53, 54, 55, 56] and can be found among molecular magnets [53] as well as among the low-dimensional quantum magnets [55, 56]. Moreover, theoretical investigations of the quantum and classical spin models with mixed g𝑔gitalic_g-factors reveal many unusual effects, like various partially ordered phases with disordered sublattices [48, 56], or additional features in the spin structure factors [50]. The very distinguishing feature of the phase diagram in Fig. 3 is the zero-field highly degeneracy. Moreover, there is no continuous transition between Bβ‰ 0𝐡0B\neq 0italic_B β‰  0 ground states and zero-field ground-states manifold. Actually, at B=0𝐡0B=0italic_B = 0 the following eigenstates become degenerate: |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩, |ψ2⟩ketsubscriptπœ“2|\psi_{2}\rangle| italic_ψ start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ⟩, |ψ8⟩ketsubscriptπœ“8|\psi_{8}\rangle| italic_ψ start_POSTSUBSCRIPT 8 end_POSTSUBSCRIPT ⟩, |ψ10⟩ketsubscriptπœ“10|\psi_{10}\rangle| italic_ψ start_POSTSUBSCRIPT 10 end_POSTSUBSCRIPT ⟩, |ψ12⟩ketsubscriptπœ“12|\psi_{12}\rangle| italic_ψ start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT ⟩. The same is valid for the special point, g2=βˆ’3subscript𝑔23g_{2}=-3italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 3. Another interesting feature of the magnetization processes in this case which can be seen from the ground state phase diagram is asymptotic saturation. As one of the S=1/2𝑆12S=1/2italic_S = 1 / 2 spins has negative g𝑔gitalic_g-factor in the case under consideration, the value of the total spin Sz=s1z+s2z+S3zsuperscript𝑆𝑧superscriptsubscript𝑠1𝑧superscriptsubscript𝑠2𝑧superscriptsubscript𝑆3𝑧S^{z}=s_{1}^{z}+s_{2}^{z}+S_{3}^{z}italic_S start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = italic_s start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_s start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT + italic_S start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT for the magnetically saturated state must be 1, whereas the saturated magnetization in the units of Bohr magnetons is equal to 12⁒(3⁒g1+|g2|)123subscript𝑔1subscript𝑔2\frac{1}{2}\left(3g_{1}+|g_{2}|\right)divide start_ARG 1 end_ARG start_ARG 2 end_ARG ( 3 italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT + | italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT | ). In this case the eigenstates with Sz=1superscript𝑆𝑧1S^{z}=1italic_S start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = 1, |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩, turns to be a saturated, or quasi-saturated ground state. Due to non-conserved magnetization the high-field ground state, |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩, does not correspond to constant expectation value of the magnetization operator, on the contrary, as it can be seen form Eq. (8), the expectation value has explicit magnetic field dependence, leading to impossibility of reaching the saturation at finite values of magnetic field. This feature is also inherent to other two ground states, |ψ7⟩ketsubscriptπœ“7|\psi_{7}\rangle| italic_ψ start_POSTSUBSCRIPT 7 end_POSTSUBSCRIPT ⟩ and |ψ11⟩ketsubscriptπœ“11|\psi_{11}\rangle| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩, appeared in the phase diagram from Fig. 3. Thus, only two ground states |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ and |ψ2⟩ketsubscriptπœ“2|\psi_{2}\rangle| italic_ψ start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ⟩ for purely ferromagnetic coupling have constant value of magnetization. Non-conserving magnetization, therefore, essentially modifies zero-temperature magnetization curves for finite spin clusters [42, 45, 46, 48]. Instead of the typical step-like structure with plateaus corresponding to eigenstates with constant values of Mzsuperscript𝑀𝑧M^{z}italic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT and jumps between them corresponding to transitions between eigenstates at given values of magnetic field, in some cases magnetization curves can contain continuous part of monotone magnetization change stemming out from the magnetic field dependant expectation values of Mzsuperscript𝑀𝑧M^{z}italic_M start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT for some eigenstates. To illustrate this behaviour we draw few plots of zero-temperature magnetization processes corresponding to phase diagram form Fig. 3 (See Fig. 4).

III.0.2 Purely antiferromagnetic couplings with J1=J2subscript𝐽1subscript𝐽2J_{1}=J_{2}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT

For the case of uniform antiferromagnetic couplings, J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0, four ground states are possible, |ψ1⟩,|ψ3⟩,|ψ5⟩ketsubscriptπœ“1ketsubscriptπœ“3ketsubscriptπœ“5|\psi_{1}\rangle,|\psi_{3}\rangle,|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ , | italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ , | italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩. In addition to that, when g2=g1subscript𝑔2subscript𝑔1g_{2}=g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT there is no continuous transition between the eigenvectors, corresponding to |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ eigenstates. Therefore, along dashed line in the ground states phase diagram in Fig. 5 the correct ground state in given by the degenerate doublet |ψ3⟩0,|ψ9⟩0subscriptketsubscriptπœ“30subscriptketsubscriptπœ“90|\psi_{3}\rangle_{0},|\psi_{9}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT , | italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT from the Eq. (10). Another interesting feature is zero-field ground states which turns to be |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ for all values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and is never degenerate in contrast to the previous case. The saturated state, the ground state for large enough values of magnetic field, is different for positive and negative values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. For the positive values of the g𝑔gitalic_g-factors the ordinary saturated (maximally polarized) eigenstate, |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ is the saturated one, however, region of negative g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT as in a previous case has |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ as high-field ground state. It is also worth mentioning, that only two ground states |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ have constant value of the magnetization. Additional intermediate ground state, |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ appears only when g2>g1subscript𝑔2subscript𝑔1g_{2}>g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. The zero-temperature magnetization curves illustrating the phase diagram are presented in Fig. 6. Here one can see all the magnetization processes features we demonstrated in the ground state phase diagram, quasi-saturation for negative g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT as well as regular saturation for g2<0subscript𝑔20g_{2}<0italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0, intermediate states with constant and continuously growing magnetization and magnetization jumps.

Refer to caption
Figure 5: Ground states phase diagram for J1=J2=1subscript𝐽1subscript𝐽21J_{1}=J_{2}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1.
Refer to caption
Figure 6: Magnetization processed for J1=J2=1subscript𝐽1subscript𝐽21J_{1}=J_{2}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 for T/|J1|=0.00001𝑇subscript𝐽10.00001T/|J_{1}|=0.00001italic_T / | italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT | = 0.00001 and g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 for different values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT: g2=βˆ’2subscript𝑔22g_{2}=-2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 2 (solid); g2=1/2subscript𝑔212g_{2}=1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 / 2 (dashed); g2=1subscript𝑔21g_{2}=1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 (dot-dashed) and g2=2subscript𝑔22g_{2}=2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 2 (dotted).
Refer to caption
Figure 7: Ground-states phase diagram for J1=1/5subscript𝐽115J_{1}=1/5italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 / 5, J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1
Refer to caption
Figure 8: Magnetization processed for J1=1/51subscript𝐽1151J_{1}=1/51italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 / 51, J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1, T/|J1|=0.00001𝑇subscript𝐽10.00001T/|J_{1}|=0.00001italic_T / | italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT | = 0.00001, g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 for different values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT: g2=βˆ’1subscript𝑔21g_{2}=-1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 (solid); g2=1/5subscript𝑔215g_{2}=1/5italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 / 5 (dashed); g2=1subscript𝑔21g_{2}=1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 (dot-dashed) and g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2 (dotted).
Refer to caption
Figure 9: Ground-states phase diagram for J1=βˆ’1/5subscript𝐽115J_{1}=-1/5italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1 / 5, J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1
Refer to caption
Figure 10: Magnetization processed for J1=βˆ’1/5subscript𝐽115J_{1}=-1/5italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1 / 5, J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1, T/|J1|=0.00001𝑇subscript𝐽10.00001T/|J_{1}|=0.00001italic_T / | italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT | = 0.00001, g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 for different values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT: g2=2subscript𝑔22g_{2}=2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 2 (solid); g2=βˆ’1/10subscript𝑔2110g_{2}=-1/10italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 10 (dashed); g2=βˆ’1subscript𝑔21g_{2}=-1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 (dot-dashed) and g2=βˆ’2subscript𝑔22g_{2}=-2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 2 (dotted).
Refer to caption
Figure 11: Ground-states phase diagram for J1=1subscript𝐽11J_{1}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1, and arbitrary J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0

III.0.3 Non-uniform antiferromagnetic couplings, J2>J1>0subscript𝐽2subscript𝐽10J_{2}>J_{1}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0

The case of antiferromagnetic non-uniform coupling with J2>J1>0subscript𝐽2subscript𝐽10J_{2}>J_{1}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0 shares many features of the previous one (See Fig. 7 for J1=1/5subscript𝐽115J_{1}=1/5italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 / 5 and J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1). The set of possible ground states is the same, |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩, |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩, |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩, again the zero-field ground state is non-degenerate and is given by |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩. As in the previous case we plotted the ground states phase diagram for negative and positive values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT with saturated state given by |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ and |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ respectively. However, the topology of diagram is a bit different as well as an additional feature appears here, for a narrow region of positive g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT a direct transition from |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ to |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ is found. The magnetization processes, corresponding to different values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT for this case are presented in Fig. 8.

III.0.4 Mixed case J1<0subscript𝐽10J_{1}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 and J2>0subscript𝐽20J_{2}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0

The ground state phase diagram for the case J1=βˆ’1/5subscript𝐽115J_{1}=-1/5italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - 1 / 5 and J2=1subscript𝐽21J_{2}=1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 is presented in Fig. 9. Here we see additional possible ground state |ψ4⟩ketsubscriptπœ“4|\psi_{4}\rangle| italic_ψ start_POSTSUBSCRIPT 4 end_POSTSUBSCRIPT ⟩ along with all four ground states form the previous case. Interestingly, |ψ4⟩ketsubscriptπœ“4|\psi_{4}\rangle| italic_ψ start_POSTSUBSCRIPT 4 end_POSTSUBSCRIPT ⟩ is the zero-field ground state for negative g2β‰€βˆ’1subscript𝑔21g_{2}\leq-1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ - 1. At the point g2=βˆ’1subscript𝑔21g_{2}=-1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 it is degenerate with the g2β‰₯βˆ’1subscript𝑔21g_{2}\geq-1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β‰₯ - 1 zero-field ground state, |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩, which was the zero-field ground state for all range of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT in previous case. However, the region of |ψ4⟩ketsubscriptπœ“4|\psi_{4}\rangle| italic_ψ start_POSTSUBSCRIPT 4 end_POSTSUBSCRIPT ⟩ ground state in the phase diagram is extremely narrow. As usual, saturated state for positive values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT is fully positively polarized |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ and for the negative values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT the same role plays Sz=1superscript𝑆𝑧1S^{z}=1italic_S start_POSTSUPERSCRIPT italic_z end_POSTSUPERSCRIPT = 1 eigenstate |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩. Typical magnetization curves for this case are presented in Fig. 10.

III.0.5 Mixed case J1>0subscript𝐽10J_{1}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0 and J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0

Another case of mixed ferromagnetic-antiferromagnetic couplings corresponds to J1>0subscript𝐽10J_{1}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 0 and J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0. The corresponding ground-states phase diagram is presented in Fig. 11. Without loss of generality we set here J1=1subscript𝐽11J_{1}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 and J2=βˆ’1subscript𝐽21J_{2}=-1italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1, however, the phase diagram preserve its structure for arbitrary J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0. As was mentioned above, this is due to Eq. (9). Here, the zero-field ground states can be |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩, |ψ7⟩ketsubscriptπœ“7|\psi_{7}\rangle| italic_ψ start_POSTSUBSCRIPT 7 end_POSTSUBSCRIPT ⟩ and |ψ11⟩ketsubscriptπœ“11|\psi_{11}\rangle| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ depending on the value of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. All the three ground states become degenerate at the particular value of g2=5subscript𝑔25g_{2}=5italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 5. Saturated state is |ψ1⟩ketsubscriptπœ“1|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ for arbitrary value of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. The magnetization plots illustrating this phase diagram are presented in Fig. 12.

Refer to caption
Figure 12: Magnetization processed for J1=1subscript𝐽11J_{1}=1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 and arbitrary J2≀0subscript𝐽20J_{2}\leq 0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0, T/J1=0.00001𝑇subscript𝐽10.00001T/J_{1}=0.00001italic_T / italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 0.00001 and g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 for different values of g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT: g2=1subscript𝑔21g_{2}=1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1 (solid); g2=3subscript𝑔23g_{2}=3italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 (dashed); g2=5subscript𝑔25g_{2}=5italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 5 (dot-dashed) and g2=6subscript𝑔26g_{2}=6italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 6 (dotted).

IV Negativity Plots

In this section we present our main results concerning the manipulation of the quantum entanglement of the spins within the cluster by means of magnetic field. Particulary, the effects of non-conserving magnetization related to the non-uniform g𝑔gitalic_g-factors are in the focus of our attention. There are several ways of quantifying the quantum entanglement using various entanglement measures [1, 40]. However, for the mixed-spin clusters negativity [41] is the most convenient one, as it can be easily constructed and calculated for any pair of spins with the aid of reduced density matrix. For the system of three particles three pairwise entanglement measures can be considered. The negativity value, which varies from 0 (no entanglement) to 1/2121/21 / 2 (maximally entangled pair), corresponding to the i-th and j-th particles of the system, N⁒ei⁒j𝑁subscript𝑒𝑖𝑗Ne_{ij}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT, equals the sum of absolute values of negative eigenvalues of partially transposed reduced two-particle density matrix, ρi⁒jTsuperscriptsubscriptπœŒπ‘–π‘—π‘‡\rho_{ij}^{T}italic_ρ start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_T end_POSTSUPERSCRIPT, which is constructed in the following way:

⟨ξi~,ΞΎj|ρi⁒jT|ΞΎi,ΞΎj~⟩=⟨ξi,ΞΎj|ρi⁒j|ΞΎi~,ΞΎj~⟩,quantum-operator-product~subscriptπœ‰π‘–subscriptπœ‰π‘—superscriptsubscriptπœŒπ‘–π‘—π‘‡subscriptπœ‰π‘–~subscriptπœ‰π‘—quantum-operator-productsubscriptπœ‰π‘–subscriptπœ‰π‘—subscriptπœŒπ‘–π‘—~subscriptπœ‰π‘–~subscriptπœ‰π‘—\displaystyle\left\langle\tilde{\xi_{i}},\xi_{j}\right|\rho_{ij}^{T}\left|\xi_% {i},\tilde{\xi_{j}}\right\rangle=\left\langle\xi_{i},\xi_{j}\left|\rho_{ij}% \right|\tilde{\xi_{i}},\tilde{\xi_{j}}\right\rangle,⟨ over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT end_ARG , italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT | italic_ρ start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_T end_POSTSUPERSCRIPT | italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT , over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT end_ARG ⟩ = ⟨ italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT , italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT | italic_ρ start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT | over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT end_ARG , over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT end_ARG ⟩ , (11)
ρi⁒j=βˆ‘ΞΎk⟨ξi,ΞΎj,ΞΎk|ρ|ΞΎi~,ΞΎj~,ΞΎk⟩subscriptπœŒπ‘–π‘—subscriptsubscriptπœ‰π‘˜quantum-operator-productsubscriptπœ‰π‘–subscriptπœ‰π‘—subscriptπœ‰π‘˜πœŒ~subscriptπœ‰π‘–~subscriptπœ‰π‘—subscriptπœ‰π‘˜\displaystyle\rho_{ij}=\sum_{\xi_{k}}\left\langle\xi_{i},\xi_{j},\xi_{k}\left|% \rho\right|\tilde{\xi_{i}},\tilde{\xi_{j}},\xi_{k}\right\rangleitalic_ρ start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT = βˆ‘ start_POSTSUBSCRIPT italic_ΞΎ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT end_POSTSUBSCRIPT ⟨ italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT , italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT , italic_ΞΎ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT | italic_ρ | over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT end_ARG , over~ start_ARG italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT end_ARG , italic_ΞΎ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ⟩

where, |ΞΎi,ΞΎj,ΞΎk⟩ketsubscriptπœ‰π‘–subscriptπœ‰π‘—subscriptπœ‰π‘˜\left|\xi_{i},\xi_{j},\xi_{k}\right\rangle| italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT , italic_ΞΎ start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT , italic_ΞΎ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ⟩ is a standard basis for the states of 𝐬𝟏,𝐬𝟐subscript𝐬1subscript𝐬2\mathbf{s_{1}},\mathbf{s_{2}}bold_s start_POSTSUBSCRIPT bold_1 end_POSTSUBSCRIPT , bold_s start_POSTSUBSCRIPT bold_2 end_POSTSUBSCRIPT (ΞΎi=Β±1/2subscriptπœ‰π‘–plus-or-minus12\xi_{i}=\pm 1/2italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT = Β± 1 / 2) and π’πŸ‘subscript𝐒3\mathbf{S_{3}}bold_S start_POSTSUBSCRIPT bold_3 end_POSTSUBSCRIPT (ΞΎi=βˆ’1,0,1subscriptπœ‰π‘–101\xi_{i}=-1,0,1italic_ΞΎ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT = - 1 , 0 , 1) spins. Then the negativity is obtained according to

N⁒ei⁒j=βˆ‘a|ΞΌa|,𝑁subscript𝑒𝑖𝑗subscriptπ‘Žsubscriptπœ‡π‘ŽNe_{ij}=\sum\limits_{a}|\mu_{a}|,italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT = βˆ‘ start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT | italic_ΞΌ start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT | , (12)

where ΞΌasubscriptπœ‡π‘Ž\mu_{a}italic_ΞΌ start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT – negative eigenvalues of partially transposed reduced two-particle density matrix ρi⁒jTsubscriptsuperscriptπœŒπ‘‡π‘–π‘—\rho^{T}_{ij}italic_ρ start_POSTSUPERSCRIPT italic_T end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT [41]. As the main object of our research is a quantum entanglement, i. e. the zero-temperature or the pure state properties , the density matrix we are working with is defined for each of the twelve eigenstates of the Hamiltonian as a pure-state density matrix:.

ρi=|Ξ¨i⟩⁒⟨Ψi|,i=1,…,12.formulae-sequencesubscriptπœŒπ‘–ketsubscriptΨ𝑖brasubscriptΨ𝑖𝑖1…12\rho_{i}=|\Psi_{i}\rangle\langle\Psi_{i}|,\;\;i=1,...,12.italic_ρ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT = | roman_Ξ¨ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT ⟩ ⟨ roman_Ξ¨ start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT | , italic_i = 1 , … , 12 . (13)

In case of n𝑛nitalic_n degenerate eigenstates one should use

ρi1⁒…⁒in=1nβ’βˆ‘a=1n|Ξ¨ia⟩⁒⟨Ψia|.subscript𝜌subscript𝑖1…subscript𝑖𝑛1𝑛superscriptsubscriptπ‘Ž1𝑛ketsubscriptΞ¨subscriptπ‘–π‘ŽbrasubscriptΞ¨subscriptπ‘–π‘Ž\displaystyle\rho_{i_{1}...i_{n}}=\frac{1}{n}\sum_{a=1}^{n}|\Psi_{i_{a}}% \rangle\langle\Psi_{i_{a}}|.italic_ρ start_POSTSUBSCRIPT italic_i start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT … italic_i start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG italic_n end_ARG βˆ‘ start_POSTSUBSCRIPT italic_a = 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT | roman_Ξ¨ start_POSTSUBSCRIPT italic_i start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT end_POSTSUBSCRIPT ⟩ ⟨ roman_Ξ¨ start_POSTSUBSCRIPT italic_i start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT end_POSTSUBSCRIPT | . (14)
Refer to caption
Figure 13: Density plots of negativity for the case of J1<0,J2≀0formulae-sequencesubscript𝐽10subscript𝐽20J_{1}<0,J_{2}\leq 0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0. Panels (a), (b) and (c) show the values of N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively.

Therefore, in general thirty six negativity functions are relevant for the mixed-spin trimer, however, many of them are trivial, or just vanish. In addition to that, the case of g1=g2subscript𝑔1subscript𝑔2g_{1}=g_{2}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT should be considered separately. For the spin clusters with conserved magnetization eigenvectors do not contain magnetic filed dependent coefficient, thus neither do the negativity. However, all they can be obtained analytically, we list the expressions for non-zero negativity functions in the Appendix. Thus, for the eigenstates form |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ to |ψ12⟩ketsubscriptπœ“12|\psi_{12}\rangle| italic_ψ start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT ⟩ at least for some pairs of spins negativity may have continuous dependence on the magnetic field, in contrast to the constant values for the clusters with conserved magnetization. Interestingly, all negativity functions have no dependents on J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. As was mentioned above, some eigenstates of the system with non-uniform g𝑔gitalic_g-factors do not admit continuous limit g2β†’g1β†’subscript𝑔2subscript𝑔1g_{2}\rightarrow g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β†’ italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT, for those eigenstates negativity was calculates separately. For demonstrating the overall picture of how quantum entanglement changes across the various eigenstates we present density plots of negativity projected into the ground-states phase diagrams (See Figs. 13-17). In Figs. 13, 14 and 17 panel (a)π‘Ž(a)( italic_a ) corresponds to N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, (b)𝑏(b)( italic_b ) to N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and (c)𝑐(c)( italic_c ) to N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT. For two phase diagrams, Figs. 15 and 16 only two panels are presented, as for both cases, J2>0,J1=βˆ’15⁒J2formulae-sequencesubscript𝐽20subscript𝐽115subscript𝐽2J_{2}>0,J_{1}=-\frac{1}{5}J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 , italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - divide start_ARG 1 end_ARG start_ARG 5 end_ARG italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and J2>0,J1=15⁒J2formulae-sequencesubscript𝐽20subscript𝐽115subscript𝐽2J_{2}>0,J_{1}=\frac{1}{5}J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 , italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG 5 end_ARG italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT, N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT vanishes almost everywhere, except for small region within the high-field ground state |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ where ist value does not exceed 0.070.070.070.07, for those plots panel (a)π‘Ž(a)( italic_a ) corresponds to N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and panel (b)𝑏(b)( italic_b ) to N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT. For the case of ferromagnetic couplings only negative g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT can lead to non-trivial magneto-thermal and entanglement properties (See Fig. 13). Here one can see essentially entangled state of the first and the third spins (panel (c)) within the |ψ11⟩ketsubscriptπœ“11|\psi_{11}\rangle| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ eigenstate. To have more clear picture of how magnetic field affects the negativity, or how negativity is changing under the magnetization process we plotted the negativity plots as a function of magnetic field for g2=βˆ’4subscript𝑔24g_{2}=-4italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 4 in Fig. 18. Three plots, corresponding to the entanglement of three pairs of spins, are composed of four parts, each one corresponding to a certain eigenstates form the sequence of transitions taking place during the magnetization process, |ψ2βŸ©β†’|ψ11βŸ©β†’|ψ7βŸ©β†’|ψ9βŸ©β†’ketsubscriptπœ“2ketsubscriptπœ“11β†’ketsubscriptπœ“7β†’ketsubscriptπœ“9|\psi_{2}\rangle\rightarrow|\psi_{11}\rangle\rightarrow|\psi_{7}\rangle% \rightarrow|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ⟩ β†’ | italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ β†’ | italic_ψ start_POSTSUBSCRIPT 7 end_POSTSUBSCRIPT ⟩ β†’ | italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩. Only negativity for |ψ2⟩ketsubscriptπœ“2|\psi_{2}\rangle| italic_ψ start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ⟩ is zero for all pairs of spins, the other eigenstates feature explicit magnetic field dependence of negativity. Negativity reaches its maximum for the first and the third spins within |ψ11⟩ketsubscriptπœ“11|\psi_{11}\rangle| italic_ψ start_POSTSUBSCRIPT 11 end_POSTSUBSCRIPT ⟩ eigenstate and its value is approximately equal to 0.410.410.410.41. Another situation when non-uniform g𝑔gitalic_g-factors bring to essential enhancement of entanglement is presented in Figs. 14 and 19. The most efficiently non-conserving magnetization affects the entanglement of the system spins in case of uniform antiferromagnetic couplings, J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0. As was shown in Fig. 5, possible ground states of the model are |ψ1⟩,|ψ3⟩,|ψ5⟩ketsubscriptπœ“1ketsubscriptπœ“3ketsubscriptπœ“5|\psi_{1}\rangle,|\psi_{3}\rangle,|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩ , | italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ , | italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩. However, at uniform g𝑔gitalic_g-factors, g2=g1subscript𝑔2subscript𝑔1g_{2}=g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT, |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ transform into |ψ3⟩0subscriptketsubscriptπœ“30|\psi_{3}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT and |ψ9⟩0subscriptketsubscriptπœ“90|\psi_{9}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT form Eq. (10). This fact is indicated in Fig. 14 by a yellow line separating |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ at g2=g1subscript𝑔2subscript𝑔1g_{2}=g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. It is obviously seen from density plots of the negativity, that systems with g2β‰ g1subscript𝑔2subscript𝑔1g_{2}\neq g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT β‰  italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT exhibit much higher entanglement that their uniform counterpart, which is manifested in the most straightforward way for N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT within the |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ eigenstate. Magnetic field behaviour of negativity for both uniform and non-uniform g𝑔gitalic_g-factors cases is presented in Fig. 19. As usual, panels (a), (b) and (c) corresponds to N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT. For all panels solid and dot-dashed lines demonstrate the negativity variation under the magnetization process for g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2 and g2=βˆ’1/2subscript𝑔212g_{2}=-1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 2, respectively, whereas dashed line corresponds to the uniform g𝑔gitalic_g-factor case, g2=1subscript𝑔21g_{2}=1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 1. For N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT (panel (a)) no entanglement is observed for g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2, though during the magnetization process the system passes through the following sequence of eigenstates, |ψ5βŸ©β†’|ψ3βŸ©β†’|ψ1βŸ©β†’ketsubscriptπœ“5ketsubscriptπœ“3β†’ketsubscriptπœ“1|\psi_{5}\rangle\rightarrow|\psi_{3}\rangle\rightarrow|\psi_{1}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ β†’ | italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ β†’ | italic_ψ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT ⟩, which have a complicated structure, but it turns out, that the first and the second spins are not entangled there. At the same time uniform g𝑔gitalic_g-factors case (dashed) exhibits non-zero constant negativity for the finite range of magnetic field, corresponding to the phase boundary between |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ and |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩, which in its turn is a degenerate superposition of |ψ3⟩0subscriptketsubscriptπœ“30|\psi_{3}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT and |ψ9⟩0subscriptketsubscriptπœ“90|\psi_{9}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT. The negativity has a constant value here, N⁒e12(3+9)0=18⁒(2βˆ’1)≃0.052𝑁superscriptsubscript𝑒12subscript3901821similar-to-or-equals0.052Ne_{12}^{(3+9)_{0}}=\frac{1}{8}\left(\sqrt{2}-1\right)\simeq 0.052italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 8 end_ARG ( square-root start_ARG 2 end_ARG - 1 ) ≃ 0.052. For the case of g2=βˆ’1/2subscript𝑔212g_{2}=-1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 2 (dot-dashed) one can see a jump from zero to N⁒e129≃0.14similar-to-or-equals𝑁superscriptsubscript𝑒1290.14Ne_{12}^{9}\simeq 0.14italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 9 end_POSTSUPERSCRIPT ≃ 0.14 with the further continuous decrease. Negativity plots quantifying entanglement for the second and the third particle can be seen in the panel (b). Interestingly, in the case of uniform g𝑔gitalic_g-factors (dashed) this pair of spins exhibit the strongest entanglement at low field, N⁒e235=1/3𝑁superscriptsubscript𝑒23513Ne_{23}^{5}=1/3italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = 1 / 3, then it drops to another constant, but much smaller value, N⁒e23(3+9)0=116⁒(17βˆ’3)≃0.07𝑁superscriptsubscript𝑒23subscript390116173similar-to-or-equals0.07Ne_{23}^{(3+9)_{0}}=\frac{1}{16}\left(\sqrt{17}-3\right)\simeq 0.07italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 16 end_ARG ( square-root start_ARG 17 end_ARG - 3 ) ≃ 0.07. The case g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2 depicted by the solid line doesn’t look really promising as only in the vicinity of B=0𝐡0B=0italic_B = 0 the negativity is close to the uniform case, then it is decreasing a little bit, and after transition to |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ it drops down to zero. The case of negative g2=βˆ’1/2subscript𝑔212g_{2}=-1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 2 (dot-dashed) has no advantage over the uniform g𝑔gitalic_g case for low field, actually within the |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ eigenstate, when N⁒e235𝑁superscriptsubscript𝑒235Ne_{23}^{5}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT drops down from 1/3131/31 / 3 to the value slightly bellow 0.2. However, it undergoes a jump at the transition point to the eigenstate |ψ9⟩ketsubscriptπœ“9|\psi_{9}\rangle| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ with the further slow decrease, being, nevertheless, much higher than the negativity of the uniform g𝑔gitalic_g system.

Refer to caption
Figure 14: Density plots of negativity for the case of J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0. Panels (a), (b) and (c) show the values of N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively.

The most significant enhancement of entanglement due to non-uniform g𝑔gitalic_g-factors occurs for the first and the third particles (panel (c)). The negativity curve for uniform g𝑔gitalic_g-factors has the same form as N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT, two plateaus at 1/3131/31 / 3 and 116⁒(17βˆ’3)≃0.07similar-to-or-equals1161730.07\frac{1}{16}\left(\sqrt{17}-3\right)\simeq 0.07divide start_ARG 1 end_ARG start_ARG 16 end_ARG ( square-root start_ARG 17 end_ARG - 3 ) ≃ 0.07 corresponding to |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ and degenerate superposition of |ψ3⟩0subscriptketsubscriptπœ“30|\psi_{3}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT and |ψ9⟩0subscriptketsubscriptπœ“90|\psi_{9}\rangle_{0}| italic_ψ start_POSTSUBSCRIPT 9 end_POSTSUBSCRIPT ⟩ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT respectively. Negativity curve for g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2 (solid) has weak magnetic field dependence within the |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ eigenstate where it demonstrates slow growth starting from 1/3131/31 / 3 value, but transition to the |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ eigenstate is accompanied with a jump to a constant value N⁒e135=3/2≃0.47𝑁superscriptsubscript𝑒13532similar-to-or-equals0.47Ne_{13}^{5}=\sqrt{3}/2\simeq 0.47italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = square-root start_ARG 3 end_ARG / 2 ≃ 0.47 which is very close to maximal possible value, 1/2121/21 / 2. This is the highest value of negativity which can be achieved for the system under consideration. It is worth mentioning here, that the value of negativity, N⁒e133=3/2𝑁superscriptsubscript𝑒13332Ne_{13}^{3}=\sqrt{3}/2italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT = square-root start_ARG 3 end_ARG / 2 is independent of g𝑔gitalic_g-factors values, however, for g2≀1subscript𝑔21g_{2}\leq 1italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 1 |ψ3⟩ketsubscriptπœ“3|\psi_{3}\rangle| italic_ψ start_POSTSUBSCRIPT 3 end_POSTSUBSCRIPT ⟩ does not appear as eigenstate during zero-temperature magnetization process. Thus, for the system with uniform antiferromagnetic exchange couplings, J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0, even an arbitrary small difference between g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT brings to essential increase of negativity N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT. The negativity curve for negative g2=βˆ’1/2subscript𝑔212g_{2}=-1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 2 (dot-dashed) demonstrates an enhancement of entanglement for |ψ5⟩ketsubscriptπœ“5|\psi_{5}\rangle| italic_ψ start_POSTSUBSCRIPT 5 end_POSTSUBSCRIPT ⟩ eigenstate with respect to uniform g𝑔gitalic_g case, one can see monotone increase of N⁒e135𝑁superscriptsubscript𝑒135Ne_{13}^{5}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT up to approximately 0.410.410.410.41, which corresponds to about 23%percent\%% increase with respect to N⁒e135=1/3𝑁superscriptsubscript𝑒13513Ne_{13}^{5}=1/3italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = 1 / 3 for g2=g1subscript𝑔2subscript𝑔1g_{2}=g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT. For the system with mixed couplings, J1⁒J2<0subscript𝐽1subscript𝐽20J_{1}J_{2}<0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0, as it can be seen from Figs. 15-17 properties of the negativity do not exhibit essential changes in case of non-uniform g𝑔gitalic_g-factors, or the eigenstates with enhanced entanglement can appear only for rather large value of the ration g2/g1subscript𝑔2subscript𝑔1g_{2}/g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT / italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT (Fig. 17, panel (c)).

Refer to caption
Figure 15: Density plots of negativity for the case of J2>0subscript𝐽20J_{2}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 and J1=βˆ’15⁒J2subscript𝐽115subscript𝐽2J_{1}=-\frac{1}{5}J_{2}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = - divide start_ARG 1 end_ARG start_ARG 5 end_ARG italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. Panels (a) and (b) show the values of N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively.
Refer to caption
Figure 16: Density plots of negativity for the case of J2>0subscript𝐽20J_{2}>0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0 and J1=15⁒J2subscript𝐽115subscript𝐽2J_{1}=\frac{1}{5}J_{2}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = divide start_ARG 1 end_ARG start_ARG 5 end_ARG italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT. Panels (a) and (b) show the values of N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively.
Refer to caption
Figure 17: Density plots of negativity for the case of and J1>1subscript𝐽11J_{1}>1italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT > 1 and arbitrary J2<0subscript𝐽20J_{2}<0italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT < 0. Panels (a), (b) and (c) show the values of N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively.
Refer to caption
Figure 18: Negativity magnetic field dependence for the case J1<0,J2≀0formulae-sequencesubscript𝐽10subscript𝐽20J_{1}<0,J_{2}\leq 0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT < 0 , italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT ≀ 0, g1=1subscript𝑔11g_{1}=1italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1 and g2=βˆ’4subscript𝑔24g_{2}=-4italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 4. Solid line corresponds to N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, dashed to N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and dot-dashed to N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT
Refer to caption
Refer to caption
Refer to caption
Figure 19: Negativity magnetic field dependence for the case J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0. Panels (a), (b) and (c) show N⁒e12𝑁subscript𝑒12Ne_{12}italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT, N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT and N⁒e13𝑁subscript𝑒13Ne_{13}italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT respectively. In each panel solid line corresponds to g2=3/2subscript𝑔232g_{2}=3/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = 3 / 2, dashed line corresponds to uniform-g𝑔gitalic_g case, g2=g1subscript𝑔2subscript𝑔1g_{2}=g_{1}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT, and dot-dashed line represent g2=βˆ’1/2subscript𝑔212g_{2}=-1/2italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = - 1 / 2 case.

V Conclusion

In the present paper we considered the problem of control and optimization of quantum entanglement in triangular-shaped mixed spin (1/2,1/2,1)12121(1/2,1/2,1)( 1 / 2 , 1 / 2 , 1 ) molecular magnet models by means of magnetic field when g𝑔gitalic_g-factors of spins are non-uniform. More specifically, we chose the model which provided us with the most simple analytical solutions due to high symmetry of the Hamiltonian. The parameters of the model turn out to distribute in the following way: spin-1 ion and one of the two spin-1/2 ion have the same g𝑔gitalic_g-factor, g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT, whereas another spin-1/2 ion has different g𝑔gitalic_g-factor, g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT; two spin-1/2 ion are coupled to spin-1 ion with different exchange constants, J1subscript𝐽1J_{1}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT and J2subscript𝐽2J_{2}italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and the coupling between two spin-1/2 ions is given by the same J1subscript𝐽1J_{1}italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT exchange constant. The presence of non-uniform g𝑔gitalic_g-factors makes the magnetization non-conserved which, in its turn, brings to a series of physical features, like asymptotic saturation, non-constant values of magnetization and negativity within the certain eigenstates. It also has deep impact on the entanglement properties, enhancement or weakening. We illustrated all these features by plotting ground states phase diagrams for five different ranges of exchange couplings and the corresponding magnetization plots. The main results, however, concern the entanglement properties and their modification under the presence of non-conserving magnetization. The most dramatic impact of non-conserving magnetization can be seen in the case of uniform antiferromagnetic couplings, J1=J2>0subscript𝐽1subscript𝐽20J_{1}=J_{2}>0italic_J start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = italic_J start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT > 0, where one can have multiple increase of N⁒e23𝑁subscript𝑒23Ne_{23}italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT. This effect is qualitative and robust, for arbitrary small difference between g2subscript𝑔2g_{2}italic_g start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT and g1subscript𝑔1g_{1}italic_g start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT the value of N⁒e233≃6.7⁒N⁒e23(3+9)0≃0.47similar-to-or-equals𝑁superscriptsubscript𝑒2336.7𝑁superscriptsubscript𝑒23subscript390similar-to-or-equals0.47Ne_{23}^{3}\simeq 6.7Ne_{23}^{(3+9)_{0}}\simeq 0.47italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT ≃ 6.7 italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ≃ 0.47, which is just six percent less than the maximal possible value of negativity. Thus, molecular magnets with non-conserving magnetization offer an opportunity of more efficient manipulation of the quantum entanglement by means of magnetic field. The main difference from the uniform-g𝑔gitalic_g case is a possibility to have continuous dependence of the negativity on the magnetic field within the same ground state as well as to achieve essentially strongly entangled eigenstates. In our research we generated non-conserving magnetization by assuming non-uniformity of the g𝑔gitalic_g-factors, which can be understood in a direct way as presence of 4⁒f4𝑓4f4 italic_f or 5⁒f5𝑓5f5 italic_f ions along with 3⁒d3𝑑3d3 italic_d or/and 4⁒d4𝑑4d4 italic_d ions into the structure of trimetallic complex [57, 58, 59, 60]. However, there is another physical setting which can effectively bring to a problem of quantum spin clusters with the difference in local field acting on different spins in the cluster. They are the models in which small groups of quantum spins are assembled into the regular lattice by alternating with the groups of Ising spins [28, 42, 43, 44, 45, 61, 62, 63, 64, 65]. The later situation is realized in a series of coordination polymer compounds [28, 42, 43, 44, 45].

VI Appendix

Here we present the list of analytical expressions for negativities corresponding to three pairs of spins.

N⁒e123=N⁒e124=N⁒e233=N⁒e234=N⁒e125=0,𝑁superscriptsubscript𝑒123𝑁superscriptsubscript𝑒124𝑁superscriptsubscript𝑒233𝑁superscriptsubscript𝑒234𝑁superscriptsubscript𝑒1250\displaystyle{Ne}_{12}^{3}={Ne}_{12}^{4}={Ne}_{23}^{3}={Ne}_{23}^{4}={Ne}_{12}% ^{5}=0,italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = 0 , (15)
N⁒e133=N⁒e134=23,𝑁superscriptsubscript𝑒133𝑁superscriptsubscript𝑒13423\displaystyle{Ne}_{13}^{3}={Ne}_{13}^{4}=\frac{\sqrt{2}}{3},italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT = divide start_ARG square-root start_ARG 2 end_ARG end_ARG start_ARG 3 end_ARG ,
N⁒e235=16⁒|M+2βˆ’M+⁒M+4+8M+2+1|+16⁒|1βˆ’8⁒M+4+1M+2+1|,𝑁superscriptsubscript𝑒23516superscriptsubscript𝑀2subscript𝑀superscriptsubscript𝑀48superscriptsubscript𝑀211618superscriptsubscript𝑀41superscriptsubscript𝑀21\displaystyle{Ne}_{23}^{5}=\frac{1}{6}\left|\frac{M_{+}^{2}-M_{+}\sqrt{M_{+}^{% 4}+8}}{M_{+}^{2}+1}\right|+\frac{1}{6}\left|\frac{1-\sqrt{8M_{+}^{4}+1}}{M_{+}% ^{2}+1}\right|,italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 6 end_ARG | divide start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT square-root start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 end_ARG end_ARG start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG | + divide start_ARG 1 end_ARG start_ARG 6 end_ARG | divide start_ARG 1 - square-root start_ARG 8 italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG | ,
N⁒e135=16⁒|M+2βˆ’M+4+8M+2+1|+16⁒|1βˆ’8⁒M+4+1M+2+1|,𝑁superscriptsubscript𝑒13516superscriptsubscript𝑀2superscriptsubscript𝑀48superscriptsubscript𝑀211618superscriptsubscript𝑀41superscriptsubscript𝑀21\displaystyle{Ne}_{13}^{5}=\frac{1}{6}\left|\frac{M_{+}^{2}-\sqrt{M_{+}^{4}+8}% }{M_{+}^{2}+1}\right|+\frac{1}{6}\left|\frac{1-\sqrt{8M_{+}^{4}+1}}{M_{+}^{2}+% 1}\right|,italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 6 end_ARG | divide start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 end_ARG end_ARG start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG | + divide start_ARG 1 end_ARG start_ARG 6 end_ARG | divide start_ARG 1 - square-root start_ARG 8 italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_M start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG | ,
N⁒e127=16⁒|Kβˆ’2βˆ’Kβˆ’4+56⁒Kβˆ’2+16+4Kβˆ’2+4|,𝑁superscriptsubscript𝑒12716superscriptsubscript𝐾2superscriptsubscript𝐾456superscriptsubscript𝐾2164superscriptsubscript𝐾24\displaystyle{Ne}_{12}^{7}=\frac{1}{6}\left|\frac{K_{-}^{2}-\sqrt{K_{-}^{4}+56% K_{-}^{2}+16}+4}{K_{-}^{2}+4}\right|,italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 7 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 6 end_ARG | divide start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 56 italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 16 end_ARG + 4 end_ARG start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 end_ARG | ,
N⁒e237=23⁒|2βˆ’2⁒Kβˆ’2+2Kβˆ’2+4|+13⁒|Kβˆ’2βˆ’Kβˆ’4+8⁒Kβˆ’2Kβˆ’2+4|,𝑁superscriptsubscript𝑒2372322superscriptsubscript𝐾22superscriptsubscript𝐾2413superscriptsubscript𝐾2superscriptsubscript𝐾48superscriptsubscript𝐾2superscriptsubscript𝐾24\displaystyle{Ne}_{23}^{7}=\frac{2}{3}\left|\frac{2-\sqrt{2}\sqrt{K_{-}^{2}+2}% }{K_{-}^{2}+4}\right|+\frac{1}{3}\left|\frac{K_{-}^{2}-\sqrt{K_{-}^{4}+8K_{-}^% {2}}}{K_{-}^{2}+4}\right|,italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 7 end_POSTSUPERSCRIPT = divide start_ARG 2 end_ARG start_ARG 3 end_ARG | divide start_ARG 2 - square-root start_ARG 2 end_ARG square-root start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 2 end_ARG end_ARG start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 end_ARG | + divide start_ARG 1 end_ARG start_ARG 3 end_ARG | divide start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG end_ARG start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 end_ARG | ,
N⁒e137=13⁒|4βˆ’2⁒Kβˆ’4+8Kβˆ’2+4|+13⁒|Kβˆ’2βˆ’Kβˆ’4+32Kβˆ’2+4|,𝑁superscriptsubscript𝑒1371342superscriptsubscript𝐾48superscriptsubscript𝐾2413superscriptsubscript𝐾2superscriptsubscript𝐾432superscriptsubscript𝐾24\displaystyle{Ne}_{13}^{7}=\frac{1}{3}\left|\frac{4-\sqrt{2}\sqrt{K_{-}^{4}+8}% }{K_{-}^{2}+4}\right|+\frac{1}{3}\left|\frac{K_{-}^{2}-\sqrt{K_{-}^{4}+32}}{K_% {-}^{2}+4}\right|,italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 7 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 3 end_ARG | divide start_ARG 4 - square-root start_ARG 2 end_ARG square-root start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 end_ARG end_ARG start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 end_ARG | + divide start_ARG 1 end_ARG start_ARG 3 end_ARG | divide start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 32 end_ARG end_ARG start_ARG italic_K start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 4 end_ARG | ,
N⁒e129=|1βˆ’G+2+1G+2+3|,𝑁superscriptsubscript𝑒1291superscriptsubscript𝐺21superscriptsubscript𝐺23\displaystyle{Ne}_{12}^{9}=\left|\frac{1-\sqrt{G_{+}^{2}+1}}{G_{+}^{2}+3}% \right|,italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 9 end_POSTSUPERSCRIPT = | divide start_ARG 1 - square-root start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e239=12⁒|1βˆ’8⁒G+2+1G+2+3|,𝑁superscriptsubscript𝑒2391218superscriptsubscript𝐺21superscriptsubscript𝐺23\displaystyle{Ne}_{23}^{9}=\frac{1}{2}\left|\frac{1-\sqrt{8G_{+}^{2}+1}}{G_{+}% ^{2}+3}\right|,italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 9 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 2 end_ARG | divide start_ARG 1 - square-root start_ARG 8 italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e139=12⁒|G+2βˆ’G+4+8G+2+3|,𝑁superscriptsubscript𝑒13912superscriptsubscript𝐺2superscriptsubscript𝐺48superscriptsubscript𝐺23\displaystyle{Ne}_{13}^{9}=\frac{1}{2}\left|\frac{G_{+}^{2}-\sqrt{G_{+}^{4}+8}% }{G_{+}^{2}+3}\right|,italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 9 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 2 end_ARG | divide start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 end_ARG end_ARG start_ARG italic_G start_POSTSUBSCRIPT + end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e1211=|1βˆ’Uβˆ’2+1Uβˆ’2+3|,𝑁superscriptsubscript𝑒12111superscriptsubscriptπ‘ˆ21superscriptsubscriptπ‘ˆ23\displaystyle{Ne}_{12}^{11}=\left|\frac{1-\sqrt{U_{-}^{2}+1}}{U_{-}^{2}+3}% \right|,italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 11 end_POSTSUPERSCRIPT = | divide start_ARG 1 - square-root start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e2311=12⁒|1βˆ’8⁒Uβˆ’2+1Uβˆ’2+3|,𝑁superscriptsubscript𝑒23111218superscriptsubscriptπ‘ˆ21superscriptsubscriptπ‘ˆ23\displaystyle{Ne}_{23}^{11}=\frac{1}{2}\left|\frac{1-\sqrt{8U_{-}^{2}+1}}{U_{-% }^{2}+3}\right|,italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 11 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 2 end_ARG | divide start_ARG 1 - square-root start_ARG 8 italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 1 end_ARG end_ARG start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e1311=12⁒|Uβˆ’2βˆ’Uβˆ’4+8Uβˆ’2+3|,𝑁superscriptsubscript𝑒131112superscriptsubscriptπ‘ˆ2superscriptsubscriptπ‘ˆ48superscriptsubscriptπ‘ˆ23\displaystyle{Ne}_{13}^{11}=\frac{1}{2}\left|\frac{U_{-}^{2}-\sqrt{U_{-}^{4}+8% }}{U_{-}^{2}+3}\right|,italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 11 end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 2 end_ARG | divide start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - square-root start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + 8 end_ARG end_ARG start_ARG italic_U start_POSTSUBSCRIPT - end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + 3 end_ARG | ,
N⁒e12(3+9)0=18⁒(2βˆ’1),𝑁superscriptsubscript𝑒12subscript3901821\displaystyle{Ne}_{12}^{(3+9)_{0}}=\frac{1}{8}\left(\sqrt{2}-1\right),italic_N italic_e start_POSTSUBSCRIPT 12 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 8 end_ARG ( square-root start_ARG 2 end_ARG - 1 ) ,
N⁒e23(3+9)0=N⁒e13(3+9)0=116⁒(17βˆ’3).𝑁superscriptsubscript𝑒23subscript390𝑁superscriptsubscript𝑒13subscript390116173\displaystyle{Ne}_{23}^{(3+9)_{0}}={Ne}_{13}^{(3+9)_{0}}=\frac{1}{16}\left(% \sqrt{17}-3\right).italic_N italic_e start_POSTSUBSCRIPT 23 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT = italic_N italic_e start_POSTSUBSCRIPT 13 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ( 3 + 9 ) start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT end_POSTSUPERSCRIPT = divide start_ARG 1 end_ARG start_ARG 16 end_ARG ( square-root start_ARG 17 end_ARG - 3 ) .

Expressions for N⁒ei⁒j6𝑁superscriptsubscript𝑒𝑖𝑗6Ne_{ij}^{6}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 6 end_POSTSUPERSCRIPT, N⁒ei⁒j8𝑁superscriptsubscript𝑒𝑖𝑗8Ne_{ij}^{8}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 8 end_POSTSUPERSCRIPT, N⁒ei⁒j10𝑁superscriptsubscript𝑒𝑖𝑗10Ne_{ij}^{10}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 10 end_POSTSUPERSCRIPT and N⁒ei⁒j12𝑁superscriptsubscript𝑒𝑖𝑗12Ne_{ij}^{12}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 12 end_POSTSUPERSCRIPT have the same structure as N⁒ei⁒j5𝑁superscriptsubscript𝑒𝑖𝑗5Ne_{ij}^{5}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT, N⁒ei⁒j7𝑁superscriptsubscript𝑒𝑖𝑗7Ne_{ij}^{7}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 7 end_POSTSUPERSCRIPT, N⁒ei⁒j9𝑁superscriptsubscript𝑒𝑖𝑗9Ne_{ij}^{9}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 9 end_POSTSUPERSCRIPT and N⁒ei⁒j11𝑁superscriptsubscript𝑒𝑖𝑗11Ne_{ij}^{11}italic_N italic_e start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 11 end_POSTSUPERSCRIPT where M+,Kβˆ’,G+superscript𝑀superscript𝐾superscript𝐺M^{+},K^{-},G^{+}italic_M start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT , italic_K start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT , italic_G start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT and Uβˆ’superscriptπ‘ˆU^{-}italic_U start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT are replaced by Mβˆ’,K+,Gβˆ’superscript𝑀superscript𝐾superscript𝐺M^{-},K^{+},G^{-}italic_M start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT , italic_K start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT , italic_G start_POSTSUPERSCRIPT - end_POSTSUPERSCRIPT and U+superscriptπ‘ˆU^{+}italic_U start_POSTSUPERSCRIPT + end_POSTSUPERSCRIPT respectively.

VII Acknowledgements

The authors express their gratitude to Vahag Abgaryan for useful discussions. They also acknowledge partial financial support form ANSEF (Grants No. PS-condmatth-2462 and PS-condmatth-2884) and from CS RA MESCS (Grants No. 21AG-1C047, 21AG-1C006, 20TTAT-QTc004 and 23AA-1C032).

References

  • [1] L. Amico, R. Fazio, A. Osterloh, V. Vedral, Rev. Mod. Phys 80, 517 (2008).
  • [2] O. GΓΌhne, G. TΓ³th, Phys. Rep. 474, 1 (2009).
  • [3] C. H. Bennett, D. P. DiVincenzo, J. A. Smolin, W. K. Wootters, Phys. Rev. A 54, 3824 (1996).
  • [4] C. H. Bennett, G. Brassard, S. Popescu, B. Schumacher, J. A. Smolin, and W. K. Wootters, Phys. Rev. Lett, 76, 722 (1996).
  • [5] A. K. Ekert, Nature (London) 358, 14 (1992).
  • [6] C. H. Bennett, G. Brassard, C. CrΓ©peau, R. Jozsa, A. Peres, and W. K. Wootters, Phys. Rev. Lett. 70, 1895 (1993).
  • [7] D. Bouwmeester, J. W. Pan, K. Mattle, M. Eibl, H. Weinfurter, and A. Zeilinger, Nature (London) 390, 575 (1997).
  • [8] J. Cho and H.-W. Lee, Phys. Rev. Lett. 95, 160501 (2005).
  • [9] X.-M. Jin, J.-G. Ren, B. Yang, Z.-H. Yi, F. Zhou, X.-F. Xu, S.-K. Wang, D. Yang, Y.-F. Hu, S. Jiang, T. Yang, H. Yin, K. Chen, C.-Z.Peng and J.-W. Pan, Nature Photonics 4, 376 (2010).
  • [10] M. Baur, A. Fedorov, L. Steffen, S. Filipp, M. P. da Silva, and A. Wallraff,, Phys. Rev. Lett. 108, 040502 (2012).
  • [11] G. BjΓΆrk, A. Laghaout, and U. L. Andersen, Phys. Rev. A 85, 022316 (2012).
  • [12] P. Kok, W. J. Munro, K. Nemoto, T.C. Ralph, J. P. Dowling, G. J. Milburn, Rev. Mod. Phys. 79, 135 (2007).
  • [13] J. Kim, J.-S. Lee, and S. Lee, C. Cheong, Phys. Rev. A 62, 022312 (2000).
  • [14] A. Brodutch and D. R. Terno, Phys. Rev. A 83, 010301(R) (2011).
  • [15] V. Vedral and E. Kashefi, Phys. Rev. Lett. 89, 037903 (2002).
  • [16] A. Avella, G. Brida, I. P. Degiovanni, M. Genovese, M. Gramegna, and P. Traina, Phys. Rev. A 82, 062309 (2010).
  • [17] E. Coronado, Nature Reviews Materials, 5, 87 (2020).
  • [18] O. Kahn, Molecular Magnetism, Wiley, New York, (1993).
  • [19] M. N. Leuenberger, D. Loss, Nature, 410, 789 (2001).
  • [20] D. Stepanenko, M. Trif, D. Loss, Inorganica Chimica Acta, 361, 3740 (2008).
  • [21] R. Sessoli, ACS Cent. Sci. 1, 473 (2015).
  • [22] N. S. Ananikian, L. N. Ananikyan, L. A. Chakhmakhchyan, and A. N. Kocharian, J. Phys.: Math. Theor. 44, 025001 (2011).
  • [23] V. S. Abgaryan, N. S. Ananikian, L. N. Ananikyan, and A. N. Kocharian, Phys. Ser. 83, 055702 (2011).
  • [24] N. Ananikian, H. Lazaryan, and M. Nalbandyan, Eur. Phys. J. B 85, 223 (2012).
  • [25] J. Strečka, O. Rojas, T. Verkholyak, and M. L. Lyra, Phys. Rev. E 89, 022143 (2014).
  • [26] I. M. Carvalho, J. Torrico, S. M. de Souza, M. Rojas, and O. Rojas, J. Mag. Magn. Mater. 465, 323 (2018).
  • [27] F. Souza, M. L. Lyra, J. Strečka, M. S. S. Pereira, J. Mag. Magn. Mater. 471, 423 (2019).
  • [28] F. Souza, L. M. VerΓ­ssimo, J. Strečka, M. L. Lyra, M. S. S. Pereira, Phys. Rev. B 102, 064414 (2020).
  • [29] Zh. Adamyan, S. Muradyan, V. Ohanyan, J. Contemp. Phys. 55, 292 (2020).
  • [30] H. Čenčarikova, and J. Strečka, Phys. Rev. B 102, 184419 (2020).
  • [31] C. Ekiz, and J. Strečka, Acta Phys. Pol. A 137, 592 (2020).
  • [32] K. KarΔΊovΓ‘, and J. Strečka, Acta Phys. Pol. A 137, 595 (2020).
  • [33] J. Strečka, O. Krupnitska, and J. Richter, EPL 132, 30004 (2020).
  • [34] L. GΓ‘lisovΓ‘, J. Strečka,, T. Verkholyak, S. Havadej, Physica E 125, 114089 (2021).
  • [35] L. GΓ‘lisovΓ‘, M. Kaczor, Entropy 23, 1671 (2021).
  • [36] L. GΓ‘lisovΓ‘, J. Magn. Magn. Mater. 561, 169721 (2022).
  • [37] F. Benabdallah, S. Haddadi, H. Arian Zad, M. R. Pourkarimi, M. Daoud, and N. Ananikian, Sci. Rep. 12, 6406 (2022).
  • [38] H. Arian Zad, A. Zoshki, N. Ananikian, and M. Jaőčur, J. Mag. Magn. Mater. 559, 169533 (2022).
  • [39] Yin-Dan Zheng, Bin Zhou, Physica A 603, 127753 (2022).
  • [40] R. Horodecki, P. Horodecki, M. Horodecki, and K. Horodecki, Rev. Mod. Phys. 81, 865 (2009).
  • [41] G. Vidal, R. F. Werner, Phys. Rev. A, 65, 032314 (2002).
  • [42] J. Strečka, M. Jaőčur, M. Hagiwara, Y. Narumi, K. Kindo, K. Minami, Phys. Rev. B 72, 024459 (2005).
  • [43] D. Visinescu, A. M. Madalan, M. Andruh, C. Duhayon, J.-P. Sutter, L. Ungur, W. Van den Heuvel, and L. F. Chibotaru, Chem. Eur. J. 15, 11808 (2009).
  • [44] W. Van den Heuvel, and L. F. Chibotaru, Phys. Rev. B 82, 174436 (2010).
  • [45] S. Bellucci, V. Ohanyan, O. Rojas, EPL 105, 47012 (2014).
  • [46] V. Ohanyan, O. Rojas, J. Strečka, S. Bellucci, Phys. Rev. B 92, 214423 (2015).
  • [47] J. Torrico, M. Rojas, S. M. de Souza, O. Rojas, Phys. Lett. A 380, 3655 (2016).
  • [48] J. Torrico, V. Ohanyan, O. Rojas, J. Magn. Magn. Mater 454, 85 (2018).
  • [49] A. D. Varizi, and R. C. Drumond, Phys. Rev. E 100, 022104 (2019).
  • [50] T. Krokhmalskii, T. Verkholyak, O. Baran, V. Ohanyan, and O. Derzhko, Phys. Rev. B 102, 144403 (2020).
  • [51] T. Pandey, and G. Y. Chitov, Phys. Rev. B 102, 054436 (2020).
  • [52] L. F. Chibotaru, in Advances in Chemical Physics, edited by S. A. Rice and A. R. Dinner, Vol. 153 (Wiley, New Jersey, 2013), pp. 397–519.
  • [53] M. Atanasov, P. Comba, and C. A. Daul, Inorg. Chem. 47, 2449 (2008).
  • [54] L. F. Chibotaru and L. Ungur, Phys. Rev. Lett. 109, 246403 (2012).
  • [55] Wei-Guo Yin, X. Liu, A. M. Tsvelik, M. P. M. Dean, M. H. Upton, Jungho Kim, D. Casa, A. Said, T. Gog, T. F. Qi, G. Cao, and J.P. Hill, Phys. Rev. Lett. 111, 057202 (2013).
  • [56] Wei-Guo Yin, Ch. R. Roth, and A. M. Tsvelik, Phys. Rev. B 109, 054427 (2024).
  • [57] M.X. Yao, Q. Zheng, K. Qian, Y. Song, S. Gao, J.L. Zuo, Chem. Eur. J. 19, 294 (2013).
  • [58] N. Bridonneau, G. Gontard, V. Marvaud, Dalton Trans. 44, 5170 (2015).
  • [59] M.G. Alexandru, D. Visinescu, M. Andruh, N. Marino, D. Armentano, J. Cano, F. Lloret, M. Julve, Chem. Eur. J. 21, 5429 (2015).
  • [60] J. Kobylarczyk, E. Kuzniak, M. Liberka, S. Chorazy, B. Sieklucka, R. Podgajny, Coord. Chem. Rev. 419, 213394 (2020).
  • [61] V. Ohanyan. Comdems. Matter Phys. 12, 343 (2009).
  • [62] S. Bellucci and V. Ohanyan, Eur. Phys. J. B 75, 531 (2010).
  • [63] V. Ohanyan and A. Honecker, Phys. Rev. B 86, 054412 (2012).
  • [64] S. Bellucci and V. Ohanyan, Eur. Phys. J. B 86, 446 (2013).
  • [65] H. Arian Zad, V. Ohanyan, A. Zoshki, and J. Strečka. Phys. Rev. E 108, 044132 (2023).