This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Manipulating topological charges via engineering zeros of wave functions

Xiao-Lin Li School of Physics, Northwest University, Xi’an 710127, China    Ming Gong CAS Key Laboratory of Quantum Information, University of Science and Technology of China, Hefei 230026, China Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei 230026, China Hefei National Laboratory, University of Science and Technology of China, Hefei 230026, China    Yu-Hao Wang School of Physics, Northwest University, Xi’an 710127, China    Li-Chen Zhao [email protected] School of Physics, Northwest University, Xi’an 710127, China Shaanxi Key Laboratory for Theoretical Physics Frontiers, Xi’an 710127, China NSFC-SPTP Peng Huanwu Center for Fundamental Theory, Xi’an 710127 and Hefei 230026, China
Abstract

Topological charges are typically manipulated by managing their energy bands in quantum systems. In this work, we propose a new approach to manipulate the topological charges of systems by engineering density zeros of localized wave excitations in them. We demonstrate via numerical simulation and analytical analysis that the winding number of a toroidal Bose condensate can be well manipulated by engineering the relative velocities between the dark solitons and their backgrounds. The crossing of relative velocities through zero makes a change in winding number by inducing density zeros during acceleration, with the direction of crossing determining whether charge increases or decreases. Possibilities of observing such winding number manipulation are discussed for current experimental settings. This idea may also be to higher dimensions. These results will inspire new pathways in designing topological materials using quantum simulation platforms.

The stability of topological defects is generally guaranteed by the conservation of their topological charges MerminRMP1979 ; ChaikinCambridge1995 ; KlemanSpringer2003 . In liquid crystals, their stability is determined by the homotopy group of loops surrounding the singularities in the real space StephenRMP1974 ; BallSpringer2017 ; FumeronLCR2023 ; FumeronEPJST2023 . Meanwhile, in recent two decades, the great efforts have been devoted to studying topological defects in the Bloch bands in the momentum space using similar mathematical tools CooperRMP2019 ; ZhangAP2019 ; HasanRMP2010 ; ShenSpringer2017 ; AsbothSpringer2016 . In this way, the topological phase transition can only occur when the conduction and valence bands are closed and reopened at some critical parameters, and in the topological phase, the bulk-edge correspondence ensures the emergence of edge states HasanRMP2010 ; ShenSpringer2017 ; AsbothSpringer2016 . The search for various topological phases and their potential application in topological quantum computation remains one of the most important themes in condensed matter physics and quantum simulation HasanRMP2010 ; ShenSpringer2017 ; AsbothSpringer2016 ; NayakRMP2008 ; BulutaRPP2011 ; BlochNP2012 ; GrossSince2017 ; TerhalRMP2015 ; TokuraNRP2019 . In experiments, the topology of bands can well explain the quantum Hall effect HeNSR2014 ; KlitzingNRP2020 ; ChangRMP2023 , Thouless pumping CitroNRP2023 ; NakajimaNP2016 ; LohseNP2016 ; JurgensenNature2021 , quantized Hall drifts AidelsburgerNP2015 ; ZhuSince2023 , and other topological phenomena in terms of the Berry curvature Berry1984 ; XiaoRMP2010 .

The robustness of topological phases implies that it is hard to engineer the topological charges without destroying the geometry of the manifolds ShenSpringer2017 ; AsbothSpringer2016 ; XiaoRMP2010 ; RachelRPP2018 ; DingNRP2022 . This general conclusion also means that we can engineer the topological charges by controlling the singularities in the manifolds. These singularities correspond to the touches of conduction and valence bands of the Bloch bands or the zeros of wave functions ShenSpringer2017 ; AsbothSpringer2016 ; XiaoRMP2010 . As we have shown in Refs. ZhaoPRE2021 ; WangPS2024 , the zeros of wave functions carries the nonzero topological charges in real space, which can be transferred to the system under consideration if they are carefully designed. As a central idea of this work, we expect that engineering the zeros of wave functions could provide an alternative approach for manipulating topological charges.

This Letter reports the manipulation of topological charges in a toroidal Bose condensate by engineering the zeros of wave functions. The major idea is that the acceleration motion of dark solitons usually involves the density zeros ZhaoPRA2020 ; MengJPB2024 , which can be used to manipulate the topology of the condensate. To this end, we consider one or more dark solitons imprinted on a ring-shaped condensate and demonstrate that the topological charge, characterized by its winding number, can be manipulated by controlling the relative velocities between the solitons and their background density currents through accelerating the solitons. We propose two protocols for accelerating dark solitons: loading (I) impurity atoms or (II) bright solitons in dark solitons and applying weak forces on them. These protocols can address the challenges of directly accelerating dark solitons due to their backgrounds. Each crossing of the zero relative velocities (corresponding to the emergence of density zeros) induces a abrupt change in the winding number, with the direction of crossing determining whether winding number increases or decreases. We further show that accelerating (III) multiple dark solitons provides diverse techniques for tailoring topological charges. This idea may be used to engineer the topology of the Bloch bands, leading to the control of the associated edge states.

Refer to caption
Figure 1: (a) The scheme for one dark soliton with coupling impurity atoms driven by a weak force FF in a toroidal Bose condensate. (b) and (c) The relative velocity vrv_{\rm r} and winding number QwQ_{\rm w}. In (b), blue dots and red line correspond to the results of numerical simulation and quasiparticle theory, respectively; green circles point the emergences of density zero, at which topological phase transitions occur, corresponding to the abrupt changes in QwQ_{\rm w} in (c). (d) The density and phase/π/\pi of dark soliton at various moments, where blue circles mark the locations of dark soliton. For this plot, we employ F=(1)t/5000.05sin(2πt/500)F=-(-1)^{\lfloor t/500\rfloor}0.05\sin(2\pi t/500), with parameters gds=gcp=1g_{\rm ds}=g_{\rm cp}=1, gim=0g_{\rm im}=0, |ψgs|2=1|\psi_{\rm gs}|^{2}=1, ε=1/10\varepsilon=1/\sqrt{10}, vr|t=00.01v_{\rm r}|_{t=0}\approx 0.01 (with vs|t=0=0.02v_{\rm s}|_{t=0}=0.02 and vb|t=00.01v_{\rm b}|_{t=0}\approx 0.01), ω=1\omega=1 and R=50R=50.

Physical Model: We use a binary toroidal Bose condensate as platform to demonstrate our idea. After rescaling the atomic mass and Planck’s constant to be 11, the dimensionless mean-field energy of system in polar coordinates can be expressed as DalfovoRMP1999 ; PitaevskiiOUPress2016

E=[\displaystyle E=\int\bigg{[} 12(|ψds|2+|ψlw|2)FRθ|ψlw|2\displaystyle\frac{1}{2}\left(|\nabla\psi_{\rm ds}|^{2}+|\nabla\psi_{\rm lw}|^{2}\right)-FR\theta|\psi_{\rm lw}|^{2}
+12ω2(rR)2(|ψds|2+|ψlw|2)\displaystyle+\frac{1}{2}\omega^{2}(r-R)^{2}(|\psi_{\rm ds}|^{2}+|\psi_{\rm lw}|^{2})
+gds2|ψds|4+glw2|ψlw|4+gcp|ψds|2|ψlw|2]rdrdθ\displaystyle+\frac{g_{\rm ds}}{2}|\psi_{\rm ds}|^{4}+\frac{g_{\rm lw}}{2}|\psi_{\rm lw}|^{4}+g_{\rm cp}|\psi_{\rm ds}|^{2}|\psi_{\rm lw}|^{2}\bigg{]}rdrd\theta (1)

where ψds\psi_{\rm ds} and ψlw\psi_{\rm lw} represent the dark solitons and localized waves components, respectively. The localized waves (either impurity atoms or bright solitons) are trapped by dark solitons through the nonlinear interaction gcp|ψds|2|ψlw|2g_{\rm cp}|\psi_{\rm ds}|^{2}|\psi_{\rm lw}|^{2}. The dimensionless paraments gdsg_{\rm ds}, glwg_{\rm lw} and gcpg_{\rm cp} represent the corresponding interaction strengths among atoms. To avoid snaking instability CarreteroNonlinearity2008 ; FrantzeskakisJPA2010 , the trapping frequency ω\omega must be sufficiently large along the radial direction rr, making the condensate appearing as a quasi-one-dimensional ring along the angular direction θ\theta. The forces FF acting on the localized waves can accelerate the dark solitons without modifying their background density, which provides a good protocol for controlling the motions of dark solitons ZhaoPRA2020 ; MengJPB2024 . Hereafter, we set the radius of the toroidal condensate to R=50R=50. In the following, we present three different strategies for manipulating the topological charges of the condensate.

(I): We consider a dark soliton coupled with some impurity atoms, where the impurity atoms are driven by a weak force FF, as shown in Fig. 1 (a). For convenience, we replace the localized wave ψlw\psi_{\rm lw} with ψim\psi_{\rm im} to account for the impurity atoms. The initial wave function of these two components can be expressed as MengNJP2023 ; MengJPB2024

ψds=[ivrc+1vr2c2tanh(c2vr2Rθ)]eivbRθψgs,\displaystyle\psi_{\rm ds}=\bigg{[}i\frac{v_{\rm r}}{c}+\sqrt{1-\frac{v_{\rm r}^{2}}{c^{2}}}\tanh\big{(}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta\big{)}\bigg{]}e^{iv_{\rm b}R\theta}\psi_{\rm gs}, (2a)
ψim=εsech[c2vr2Rθ]eivsRθψgs.\displaystyle\psi_{\rm im}=\varepsilon\cdot{\rm sech}\big{[}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta\big{]}e^{iv_{\rm s}R\theta}\psi_{\rm gs}. (2b)

In the above states, ψgs\psi_{\rm gs} is the toroidal ground state of the condensate, which can be obtained through the imaginary-time evolution method YangSIAM2010 ; BaoKRM2013 ; BaoJCP2003 ; ChiofaloPRE2000 ; LehtovaaraJCP2007 by solving the two-dimensional Gross-Pitaevskii (GP) equations, assuming that two components admit the same density. The parameters vsv_{\rm s}, vbv_{\rm b} and c=gds|ψgs|2c=\sqrt{g_{\rm ds}|\psi_{\rm gs}|^{2}} represent the soliton velocity, background density velocity, and sound speed, respectively. Therefore, vrvsvbv_{\rm r}\equiv v_{\rm s}-v_{\rm b} signifies the relative velocity between the soliton and background density. Seeing Eq. (2), we can easily check that when R|θ|R|\theta| is sufficiently large, |ψds|2=|ψgs|2|\psi_{\rm ds}|^{2}=|\psi_{\rm gs}|^{2} while |ψim|2=0|\psi_{\rm im}|^{2}=0.

The phase jump of dark soliton is given by Δϕ=(2πiz)1𝑑z\Delta\phi=\int{(2\pi iz)^{-1}dz} with z=ψdseivbRθ/ψgsz=\psi_{\rm ds}e^{-iv_{\rm b}R\theta}/\psi_{\rm gs}, and Δϕ[π,π]\Delta\phi\in[-\pi,\pi]. Notably, when vr=0v_{\rm r}=0, ψds=0\psi_{\rm ds}=0 occurs at θ=0\theta=0, yielding the zero of dark soliton, which is essential for its phase jump. In this way, the topological phase transition can only occur when the relative velocity vr=0v_{\rm r}=0. During the evolution of GP equations, the profiles of two components can still be described by Eq. (2), with vsv_{\rm s} and vbv_{\rm b} being time-dependent, thus vr=0v_{\rm r}=0 can still be employed to identify the phase transition and the associated phase jump of dark soliton. Furthermore, the single-valued nature of the order parameter requires that the phase variations along the ring should satisfy BrandJPB2001 ; MateoPRA2015 ; ShamailovPRA2019

ϕ+2πRvb=2πQw\displaystyle\bigtriangleup\phi+2\pi Rv_{\rm b}=2\pi Q_{\rm w} (3)

where 2πRvb2\pi Rv_{\rm b} represents the phase shift induced by the background density current; see Eq. (2a). The integer QwQ_{\rm w} signifies the winding number of toroidal condensate, which can be defined as the number of times the phase winds through 2π2\pi in a closed loop counterclockwise around the ring. In recent experiments RyuPRL2007 ; RamanathanPRL2011 ; WrightPRL2013 ; RyuPRL2013 ; EckelNature2014 , the winding number can only be manipulated by a rotating weak link in the absence of dark solitons, which is by changing vbv_{\rm b}; however, an easier way to achieve this is by changing vrv_{\rm r}, see below.

We present our numerical results employing the Fourier pseudo-spectral method YangSIAM2010 ; BaoKRM2013 ; BaoJCP2003 with the Strang splitting scheme to solve the GP equations in Fig. 1. The initial parameters for this our simulation are Qw|t=0=0Q_{w}|_{t=0}=0 and vr|t=00.01v_{\rm r}|_{t=0}\approx 0.01, and the weak force is designed as F=(1)t/5000.05sin(2πt/500)F=-(-1)^{\lfloor t/500\rfloor}0.05\sin(2\pi t/500), where \lfloor\ast\rfloor denotes the floor function. Since dark soliton admits a negative inertial mass, its acceleration is directed opposite to the direction of weak force MengNJP2023 ; FrantzeskakisJPA2010 ; MengJPB2024 ; AycockPNAS2017 ; HurstPRA2017 . The density and phase of dark soliton at various moments are shown in Fig. 1 (d), and the detailed dynamical evolutions of the dark soliton and impurity atoms are presented in Sec. SI of Supplemental Material (SM), which includes a video representation. The positions of the dips in the dark soliton are monitored to determine vsv_{\rm s} while the phase gradient of the background density enables the determination of vbv_{\rm b}, allowing for the identification of vrv_{\rm r}. The winding number QwQ_{\rm w} can be obtained by integrating the phase gradient of the condensate along the ring. The evolution of vrv_{\rm r} and QwQ_{\rm w} are depicted in Fig. 1 (b) and (c), respectively. Our results demonstrate that the density zero occurs at vr=0v_{\rm r}=0, precisely corresponding to the abrupt jump points of QwQ_{\rm w}. Moreover, the crossing direction of the zero relative velocity determines whether the winding number increases or decreases. As previously mentioned, the profiles of the two components can still be described by Eq. (2) during evolution, which enables us to theoretically describe the evolution of the soliton by quasiparticle theory using the Lagrangian variational method (see Sec. SII of SM for details). As depicted in Fig. 1 (b), the results given by quasiparticle theory agree with those of numerical simulation.

Refer to caption
Figure 2: The schematic diagram of manipulating winding number QwQ_{\rm w} in a toroidal condensate via engineering the density zero of dark soliton.

In Refs. ZhaoPRE2021 ; WangPS2024 , we have shown that the wave function admits a phase jump of ±π\pm\pi at the density zero, with the sign determined by the approach direction. For one dark soliton, we find that

limvr0Δϕ=±π.\displaystyle\lim_{v_{\rm r}\rightarrow 0^{\mp}}\Delta\phi=\pm\pi. (4)

Therefore, as the zero relative velocity is approached from different directions, the phase jump direction varies. For instance, when vrv_{\rm r} changes from a positive value to zero, a phase jump of π-\pi occurs; and when vrv_{\rm r} changes from zero to a negative value, a phase jump of π\pi occurs; then a total phase jump of 2π2\pi occurs, as demonstrated in Fig. 2 (a); and vice versa. In one-dimensional settings with open boundaries, this 2π2\pi phase jump occurs in accelerating dark soliton but is not observable. Linking the two ends of a one-dimensional dark soliton to form a ring configuration, we uncover a significant finding. As derived from Eq. (3), a ±2π\pm 2\pi phase jump induced by the crossing of zero relative velocity brings a change of ±1\pm 1 in the winding number QwQ_{\rm w}, as demonstrated in Fig. 2 (b). Explicitly, each crossing of zero relative velocity induces a change in the winding number, with the direction of the crossing determining whether the winding number increases or decreases.

The density zeros of wave functions have deep origins in physics. These singularities induce the emergence of nonintegrable phase factors in certain parameter spaces, with the integration encircling these singularities typically yielding a nonzero topological phase. This idea can even be traced back to Dirac’s work on monopoles DiracPRSLA1931 , wherein the relation between nonintegrable phases and the integer topological magnetic charges. Furthermore, Nye and Berry pointed out that the wave dislocations in wave systems correspond to the zeros of wave functions, and the phase singularities play essential roles in many striking phenomena in wave physics NyePRSLA1974 ; BerryLSA2023 . The dynamics of zeros are also closely related with the topological properties of physical systems DuanPRE1999 ; DuanPRE199960 , such as quantifying the spatiotemporal behavior of spiral wave by tracking phase singularities ZhangCPL2007 ; LiPRE2018 ; HePRE2021 . These results, together with the results presented in this work, suggest that the zeros of wave functions have the profound implications for physics.

Refer to caption
Figure 3: (a) The scheme for one spin soliton driven by a weak force FF in a toroidal Bose condensate. (b) and (c) The relative velocity vrv_{\rm r} and winding number QwQ_{\rm w}. In (b), blue dots and red line correspond to the results of numerical simulation and quasiparticle theory, respectively; green circles point the emergences of density zero, at which topological phase transitions occur, corresponding to the abrupt changes in QwQ_{\rm w} in (c). (d) The density and phase/π/\pi of dark soliton at various moments, where blue circles mark the locations of dark soliton. For this plot, we employ F=0.005F=0.005, with parameters gds=3g_{\rm ds}=3, gcp=2g_{\rm cp}=2, gbs=1g_{\rm bs}=1, vr|t=00.01v_{\rm r}|_{t=0}\approx-0.01 (with vs|t=0=0.02v_{\rm s}|_{t=0}=-0.02 and vb|t=00.01v_{\rm b}|_{t=0}\approx-0.01), ω=3\omega=3 and R=50R=50.

(II): The above protocol can only realize two states, characterized by the winding number of Qw|t=0±1Q_{\rm w}|_{t=0}\pm 1. This limitation prompts us to consider a more general question: can we manipulate the winding number to a arbitrary integer number? In following, we propose a feasible protocol to progressively increase or decrease winding number based on the oscillation of the spin soliton driven by a weak constant force FF ZhaoPRA2020 ; MengPRA2022 ; MengCNSNS2022 , as shown in Fig. 3 (a). Here, we replace the localized wave ψlw\psi_{\rm lw} with the bright soliton ψbs\psi_{\rm bs} and set gds+gbs=2gcpg_{\rm ds}+g_{\rm bs}=2g_{\rm cp}. The initial wave function of these two components can be expressed as ZhaoPRA2020 ; MengPRA2022

ψds={ivrcs+1vr2cs2tanh[cs2vr2Rθ]}eivbRθψgs,\displaystyle\psi_{\rm ds}=\bigg{\{}i\frac{v_{\rm r}}{c_{\rm s}}+\sqrt{1-\frac{v_{\rm r}^{2}}{c_{\rm s}^{2}}}\tanh\big{[}\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\theta\big{]}\bigg{\}}e^{iv_{\rm b}R\theta}\psi_{\rm gs}, (5a)
ψbs=1vr2cs2sech[cs2vr2Rθ]eivsRθψgs,\displaystyle\psi_{\rm bs}=\sqrt{1-\frac{v_{\rm r}^{2}}{c_{\rm s}^{2}}}{\rm sech}\big{[}\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\theta\big{]}e^{iv_{\rm s}R\theta}\psi_{\rm gs}, (5b)

where cs=gcpglwc_{\rm s}=\sqrt{g_{\rm cp}-g_{\rm lw}}, which is related to the velocity limit of spin soliton MengPRA2022 . The basic characteristic of spin soliton admits a uniform total density |ψds|2+|ψbs|2=|ψgs|2|\psi_{\rm ds}|^{2}+|\psi_{\rm bs}|^{2}=|\psi_{\rm gs}|^{2}, as shown in Fig. 3 (a). Applying a weak constant force F=0.005F=0.005 on bright soliton and setting initial parameters as Qw|t=0=0Q_{w}|_{t=0}=0 and vr|t=00.01v_{\rm r}|_{t=0}\approx-0.01, we perform numerical simulation and results are shown in Fig. 3. The density and phase of dark soliton at various moments are shown in Fig. 3 (d), and the detailed dynamical evolutions of the dark and bright solitons are presented in Sec. SI of SM, which includes a video representation. We find that that the relative velocity vrv_{\rm r} exhibits a periodic oscillation, which corresponds to the ac oscillation induced by periodic transition between negative and positive inertial mass ZhaoPRA2020 ; MengPRA2022 ; MengCNSNS2022 . Notably, density zero only emerges when the relative velocity vrv_{\rm r} crosses zero from positive to negative values, as pointed by the green circles in in Fig. 3 (b), where the spin soliton admits the negative inertial mass. There is no density zero for inverse crossing direction, which is due to different properties of spin solitons with different inertial masses ZhaoPRA2020 ; MengPRA2022 ; MengCNSNS2022 . Therefore, the dark soliton does admit the density zero at zero relative velocity in the same manner, which leads to a step-like increment in the winding number QwQ_{\rm w}, as presented in Fig. 3 (c). Also, as depicted in Fig. 3 (b), the results given by quasiparticle theory (see Sec. SII of SM for details) agree with those of numerical simulation. However, it should be emphasized that the system may become unstable and eventually break when the winding number QwQ_{\rm w} reaches very high values with the increment of background density velocity vbv_{\rm b}. For the parameters used in Fig. 3, the maximum observed value for QwQ_{\rm w} is approximately 3030.

Refer to caption
Figure 4: (a) The scheme for two well-separated dark solitons with coupling impurity atoms driven by two weak force F1,2F_{1,2} in a toroidal Bose condensate. (b) The winding number QwQ_{\rm w} with different period ratio: T1/T2=1T_{1}/T_{2}=1 and 3/23/2. For this plot, we employ F1,2=(1)t/T1,20.03sin(2πt/T1,2)F_{1,2}=-(-1)^{\lfloor t/T_{1,2}\rfloor}0.03\sin(2\pi t/T_{1,2}), with parameters T1=300T_{1}=300, gds=gcp=1g_{\rm ds}=g_{\rm cp}=1, gim=0g_{\rm im}=0, |ψgs|2=1|\psi_{\rm gs}|^{2}=1, ε=1/10\varepsilon=1/\sqrt{10}, vr|t=00.01v_{\rm r}|_{t=0}\approx 0.01 (with vs|t=0=0.02v_{\rm s}|_{t=0}=0.02 and vb|t=00.01v_{\rm b}|_{t=0}\approx 0.01), ω=1\omega=1 and R=50R=50.

(III): We investigate two well-separated dark solitons with coupling impurity atoms driven by two weak force F1,2F_{1,2}, as shown in Fig. 4 (a), which can be used as a platform to demonstrate the arbitrary tunability of winding number. The initial states for two dark solitons and coupled impurities are described by ψtds={ivr/c+1vr2/c2tanh[c2vr2Rθ1]}{ivr/c+1vr2/c2tanh[c2vr2Rθ2]}eivbRθψgs\psi_{\rm tds}=\big{\{}iv_{\rm r}/c+\sqrt{1-v_{\rm r}^{2}/c^{2}}\tanh\big{[}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta_{1}\big{]}\big{\}}\big{\{}iv_{\rm r}/c+\sqrt{1-v_{\rm r}^{2}/c^{2}}\tanh\big{[}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta_{2}\big{]}\big{\}}e^{iv_{\rm b}R\theta}\psi_{\rm gs} and ψtim=ε{sech[c2vr2Rθ1]eivsRθ1+sech[c2vr2Rθ2]eivsRθ2}ψgs\psi_{\rm tim}=\varepsilon\big{\{}{\rm sech}\big{[}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta_{1}\big{]}e^{iv_{\rm s}R\theta_{1}}+{\rm sech}\big{[}\sqrt{c^{2}-v_{\rm r}^{2}}R\theta_{2}\big{]}e^{iv_{\rm s}R\theta_{2}}\big{\}}\psi_{\rm gs} with θ1,2=θ±π/2\theta_{1,2}=\theta\pm\pi/2. Each soliton-impurity is driven by a weak force F1,2=(1)t/T1,20.03sin(2πt/T1,2)F_{1,2}=-(-1)^{\lfloor t/T_{1,2}\rfloor}0.03\sin(2\pi t/T_{1,2}). We present different evolution of winding number by adjusting period ratio T1/T2T_{1}/T_{2} in Fig. 4 (b). When T1/T2=1T_{1}/T_{2}=1, the winding number QwQ_{\rm w} jumps two times of the case of one dark soliton (see in in Fig. 1 (c)); when T1/T21T_{1}/T_{2}\neq 1, the winding number QwQ_{\rm w} evolves in a near random way due to the different frequencies and directions of the density zero crossings. More manipulations of winding number and cases involving multiple spin solitons are provided in Sec. SIII of SM.

Experimental Observation: Finally, we discuss the possibilities to observe the winding number manipulations in real experiments. Let’s consider a Rb87{}^{87}{\rm Rb} Bose condensate with two internal states HallPRL1998 ; MertesPRL2007 ; BeckerNP2008 to observe the results presented in Fig. 1. The related scattering lengthes are about 100aB100a_{\rm B} (aBa_{\rm B} being the Bohr radius) HallPRL1998 ; MertesPRL2007 . The condensate can be loaded in a toroidal optical dipole trap RyuPRL2007 ; RamanathanPRL2011 ; WrightPRL2013 ; RyuPRL2013 ; EckelNature2014 with radial and vertical trapping frequencies of 2π×(100,200)Hz2\pi\times(100,200){\rm\ Hz} and a radius of 54μm54{\rm\ \mu m}. The initial states can be prepared well by imprinting a dark soliton with 2×1042\times 10^{4} atoms (the background density of 1.3×1013cm31.3\times 10^{13}{\rm\ cm^{-3}}) in state |F=2,mf=0|F=2,m_{f}=0\rangle and filling the density dip with 1313 atoms in state |F=1,mf=1|F=1,m_{f}=-1\rangle FrantzeskakisJPA2010 ; BeckerNP2008 . The effective radial length is 1.4ξ1.4\ \xi with healing length of ξ0.76μm\xi\approx 0.76{\rm\ \mu m}. A weak uniform gradient magnetic field can be applied along the angle coordinate as the effective force to drive atoms LuoSR2016 . By setting a sinusoidal modulating gradient magnetic field LuoSR2016 ; AnnalaPRR2024 ; RayNature2014 ; RayScience2015 with peak gradient of 0.12Gcm10.12{\rm\ G\cdot cm}^{-1} and period of 0.8s0.8{\rm\ s}, one can observe 2 times winding number jumps within the time duration of 2s2{\rm\ s}. The predictions in this work could be realized through the state-of-the-art experimental techniques.

Conclusion and Discussion: To conclude, we demonstrate that the topological charge of a toroidal Bose condensate can be well manipulated by engineering its density zeros. When the relative velocities between the dark solitons and their background density currents are zeros, the density zeros occur. In our schemes, loading and driving dark solitons open up the possibility of manipulating winding number, which provides a new approach for quantum computation, quantum simulation, and other quantum-based applications NayakRMP2008 ; BulutaRPP2011 ; BlochNP2012 ; GrossSince2017 ; TerhalRMP2015 ; TokuraNRP2019 ; AmicoRMP2022 ; PoloQT2024 .

Further, vortices, as fundamental topological excitations in high-dimensional systems, also admit the presence of density zeros and their charges determine the topology of systems PitaevskiiOUPress2016 ; FetterRMP2009 ; MatthewsPRL1999 ; LeanhardtPRL2002 . The higher-order vortices exhibit distinctive properties of density zeros compared with fundamental ones ShinPRL2004 ; IsoshimaPRL2007 ; LanPRL2023 . We expect that the charges of various vortices can be well manipulated through engineering the density zeros. These results could inspire new pathways in the design of topological systems by engineering the zeros of localized waves.

This work is supported by the National Natural Science Foundation of China (Contracts No. 12375005, No. 12235007 and No. 12247103) and the Major Basic Research Program of Natural Science of Shaanxi Province (Grant No. 2018KJXX-094).

SM A: Dynamical Evolutions of Solitons

In order to visualize the dynamics of soliton in Figs. 1 and 3 in main text more clearly, we provide the dynamical evolutions of soliton driven by a weak force FF in a toroidal Bose condensate at various moments (see the videos for more details).

.1 Densities and Phases Evolutions of Soliton-Impurity System

We provide the density and phase evolution of a dark soliton with impurity atoms driven by a weak force FF, as shown in Fig. 5. We set the weak force as F=(1)t/5000.05sin(2πt/500)F=-(-1)^{\lfloor t/500\rfloor}0.05\sin(2\pi t/500) and soliton relative velocity as vr|t=00.01v_{\rm r}|_{t=0}\approx 0.01 (with soliton velocity vs|t=0=0.02v_{\rm s}|_{t=0}=0.02 and background density velocity vb|t=00.01v_{\rm b}|_{t=0}\approx 0.01), where the parameters are the same as those in Fig. 1 in main text. When a weak force applies on the impurity atoms trapped in the dark soliton, the direction of acceleration of dark soliton is in the opposite direction of force FF due to its negative mass effect MengNJP2023 ; MengJPB2024 . Therefore, the dark soliton will initially move in a counterclockwise direction and tend towards zero relative velocity. Subsequently, dark soliton will move in a clockwise direction and reverse its relative velocity direction due to the reversal of direction of force FF. This process will be repeated accordingly, as indicated by the green circles markers in Fig. 5. Explicitly, each crossing of zero relative velocity induces a change in the winding number, with the direction of the crossing determining whether the winding number QwQ_{\rm w} increases or decreases.

Refer to caption
Figure 5: The density and phase of a dark soliton with impurity atoms driven by a weak force FF in a toroidal Bose condensate at various moments. The green circles mark the locations of dark soliton and impurity atoms. The parameters are the same as those in Fig. 1 in main text.
Refer to caption
Figure 6: The density and phase of a spin soliton driven by a weak force FF in a toroidal Bose condensate at various moments. The green circles mark the locations of spin soliton. The parameters are the same as those in Fig. 3 in main text.

.2 Densities and Phases Evolutions of Spin Soliton

We provide the density and phase evolution of a spin soliton driven by a weak force FF, as shown in Fig. 6. We set the weak force as F=0.005F=0.005 and soliton relative velocity as vr|t=00.01v_{\rm r}|_{t=0}\approx-0.01 (with soliton velocity vs|t=0=0.02v_{\rm s}|_{t=0}=-0.02 and background density velocity vb|t=00.01v_{\rm b}|_{t=0}\approx-0.01), where the parameters are the same as those in Fig. 3 in main text. When a weak constant force applies on the bright soliton trapped in the dark soliton, the spin soliton exhibits a periodic oscillation due to the negative-positive mass transition ZhaoPRA2020 ; MengPRA2022 . Thus the spin soliton will initially move in a clockwise direction and then move in a counterclockwise direction. This process will be repeated accordingly, as indicated by the green circles markers in Fig. 6. Notably, the density zero only emerges when the relative velocity vrv_{\rm r} crosses zero from positive to negative values, where spin soliton admits the negative inertial mass ZhaoPRA2020 ; MengPRA2022 . Therefore, the continuous crossing of zero relative velocity in the same manner induces a step-like increment in the winding number QwQ_{\rm w}.

SM B: Derivation of Quasiparticle Description

The profiles of solitons in two components admit certain particle properties during evolution, which provide possibilities to theoretically describe the evolution of solitons by quasiparticle theory using the Lagrangian variational method.

.3 Quasiparticle Description of Soliton-Impurity System

Refer to caption
Figure 7: The evolution of soliton velocity vsv_{\rm s}, background density velocity vbv_{\rm b}, and relative velocity vrv_{\rm r} for soliton-impurity driven by a weak force FF in a toroidal Bose condensate. The dots and lines correspond to the results of numerical simulation and quasiparticle theory, respectively. The parameters are the same as those in Fig. 1 in main text.

Considering a dark soliton coupled with some impurity atoms along the ring, the dynamics of quasi-one-dimensional coupled system can be described by

iψdst=\displaystyle i\frac{\partial\psi_{\rm ds}}{\partial t}= [122R2θ2+gds|ψds|2+gcp|ψim|2]ψds,\displaystyle\left[-\frac{1}{2}\frac{\partial^{2}}{R^{2}\partial\theta^{2}}+g_{\rm ds}|\psi_{\rm ds}|^{2}+g_{\rm cp}|\psi_{\rm im}|^{2}\right]\psi_{\rm ds}, (6a)
iψimt=\displaystyle i\frac{\partial\psi_{\rm im}}{\partial t}= [122R2θ2+gcp|ψds|2FRθ]ψim.\displaystyle\left[-\frac{1}{2}\frac{\partial^{2}}{R^{2}\partial\theta^{2}}+g_{\rm cp}|\psi_{\rm ds}|^{2}-FR\theta\right]\psi_{\rm im}. (6b)

We try to use the Lagrangian variational method to evaluate the dynamics of solitons by introducing the following trial wavefunctions (taking c=1c=1) MengNJP2023 ; MengJPB2024

ψds(t)=\displaystyle\psi_{\rm ds}(t)= {i1p(t)2+p(t)tanh[Rθθc(t)w(t)]}eivb(t)Rθ,\displaystyle\left\{i\sqrt{1-p(t)^{2}}+p(t)\tanh\left[R\frac{\theta-\theta_{\rm c}(t)}{w(t)}\right]\right\}e^{iv_{\rm b}(t)R\theta}, (7a)
ψim(t)=\displaystyle\psi_{\rm im}(t)= ε(t)sech[Rθθc(t)w(t)]ei{γ1(t)+γ2(t)R[θθc(t)]},\displaystyle\varepsilon(t){\rm sech}\left[R\frac{\theta-\theta_{\rm c}(t)}{w(t)}\right]e^{i\left\{\gamma_{1}(t)+\gamma_{2}(t)R[\theta-\theta_{\rm c}(t)]\right\}}, (7b)

where p(t)p(t) accounts for the depth of dark soliton, w(t)w(t) for its width, and θc(t)\theta_{\rm c}(t) for its position angle. Considering the single-valued order parameter whenever a closed loop is followed, we need introduce the continuity condition Δϕ(t)+2πRvb(t)=2πQw(t)\Delta\phi(t)+2\pi Rv_{\rm b}(t)=2\pi Q_{\rm w}(t). Then the Lagrangian of the system can be written as

L(t)=\displaystyle L(t)= ππ{i2[(ψdstψdsψdstψds)(11|ψds|2)+(ψimtψimψimtψim)]\displaystyle\int_{-\pi}^{\pi}\bigg{\{}\frac{i}{2}\bigg{[}\bigg{(}\psi_{\rm ds}^{*}\frac{\partial}{\partial t}\psi_{\rm ds}-\psi_{\rm ds}\frac{\partial}{\partial t}\psi_{\rm ds}^{*}\bigg{)}\bigg{(}1-\frac{1}{|\psi_{\rm ds}|^{2}}\bigg{)}+\bigg{(}\psi_{\rm im}^{*}\frac{\partial}{\partial t}\psi_{\rm im}-\psi_{\rm im}\frac{\partial}{\partial t}\psi_{\rm im}^{*}\bigg{)}\bigg{]}
[12(|ψdsRθ|2+|ψimRθ|2)+gds2(|ψds|21)2+gcp(|ψds|21)|ψim|2FRθ|ψim|2]}Rdθ\displaystyle\ \ \ \ \ \ \ \ -\bigg{[}\frac{1}{2}\bigg{(}\bigg{|}\frac{\partial\psi_{\rm ds}}{R\partial\theta}\bigg{|}^{2}+\bigg{|}\frac{\partial\psi_{\rm im}}{R\partial\theta}\bigg{|}^{2}\bigg{)}+\frac{g_{\rm ds}}{2}\big{(}|\psi_{\rm ds}|^{2}-1\big{)}^{2}+g_{\rm cp}\big{(}|\psi_{\rm ds}|^{2}-1\big{)}|\psi_{\rm im}|^{2}-FR\theta|\psi_{\rm im}|^{2}\bigg{]}\bigg{\}}Rd\theta
=\displaystyle= 2{arcsin[p(t)]p(t)1p(t)2}Rθ˙c(t)+2ε(t)2w(t)[γ˙1(t)+γ2(t)Rθ˙c(t)12γ2(t)2]ε(t)23w(t)2p(t)23w(t)\displaystyle\ 2\{\arcsin\big{[}p(t)\big{]}-p(t)\sqrt{1-p(t)^{2}}\big{\}}R\dot{\theta}_{\rm c}(t)+2\varepsilon(t)^{2}w(t)\bigg{[}-\dot{\gamma}_{1}(t)+\gamma_{2}(t)R\dot{\theta}_{\rm c}(t)-\frac{1}{2}\gamma_{2}(t)^{2}\bigg{]}-\frac{\varepsilon(t)^{2}}{3w(t)}-\frac{2p(t)^{2}}{3w(t)}
23gdsp(t)4w(t)+43gcpε(t)2p(t)2w(t)+Φ[p(t),w(t),t]+2ε(t)2w(t)FRθc(t),\displaystyle-\frac{2}{3}g_{\rm ds}p(t)^{4}w(t)+\frac{4}{3}g_{\rm cp}\varepsilon(t)^{2}p(t)^{2}w(t)+\Phi\big{[}p(t),w(t),t\big{]}+2\varepsilon(t)^{2}w(t)FR\theta_{\rm c}(t), (8)

where Φ[p(t),w(t),t]=p(t)1p(t)2[2πQw(t)Δϕ(t)πR]+[p(t)2w(t)πR][2πQw(t)Δϕ(t)2πR]2\Phi\big{[}p(t),w(t),t\big{]}=p(t)\sqrt{1-p(t)^{2}}\big{[}\frac{2\pi Q_{\rm w}(t)-\Delta\phi(t)}{\pi R}\big{]}+\big{[}p(t)^{2}w(t)-\pi R\big{]}\big{[}\frac{2\pi Q_{\rm w}(t)-\Delta\phi(t)}{2\pi R}\big{]}^{2}. It is now straightforward to apply the Euler-Lagrangian equation ddt[L(t)α˙(t)]=[L(t)α(t)]\frac{d}{dt}\big{[}\frac{\partial L(t)}{\partial\dot{\alpha}(t)}\big{]}=\big{[}\frac{\partial L(t)}{\partial\alpha(t)}\big{]}, where α(t)={p(t),w(t),θc(t),γ1(t),γ2(t)}\alpha(t)=\{p(t),\ w(t),\ \theta_{\rm c}(t),\ \gamma_{1}(t),\ \gamma_{2}(t)\}. Solving them, the most essential relations for soliton motion can be derived as follows

ddt{arcsin[p(t)]p(t)1p(t)2}=ε(t)2w(t)F,\displaystyle\frac{d}{dt}\big{\{}\arcsin\big{[}p(t)\big{]}-p(t)\sqrt{1-p(t)^{2}}\big{\}}=\varepsilon(t)^{2}w(t)F, (9)

and

Φ[p(t),w(t),t]p(t)=4p(t)2Rθ˙c(t)1p(t)243[p(t)w(t)+2gdsp(t)3w(t)],\displaystyle\frac{\partial\Phi[p(t),w(t),t]}{\partial p(t)}=\frac{4p(t)^{2}R\dot{\theta}_{\rm c}(t)}{\sqrt{1-p(t)^{2}}}-\frac{4}{3}\bigg{[}\frac{p(t)}{w(t)}+2g_{\rm ds}p(t)^{3}w(t)\bigg{]}, (10)

where p(t)w(t)=1p(t)w(t)=1 and 2ε(t)2w(t)=Nim2\varepsilon(t)^{2}w(t)=N_{\rm im} denotes the particle number of impurity atoms. In order to describe the evolution of solitons in Fig. 1 of main text, we firstly numerically calculate p(t)p(t) from Eq. (9); then, we can obtain the soliton velocity vs(t)=Rθ˙c(t)v_{\rm s}(t)=R\dot{\theta}_{\rm c}(t) from Eq. (10). The background density velocity can be calculated through vb(t)=[2πQw(t)Δϕ(t)]/2πRv_{\rm b}(t)=[2\pi Q_{\rm w}(t)-\Delta\phi(t)]/2\pi R. Thus, we can directly get the relative velocity vr=vsvbv_{\rm r}=v_{\rm s}-v_{\rm b}, which is presented as the red lines in Fig. 1 (b) of main text. As shown in Fig. 7, results of quasiparticle theory are in agree with those of numerical simulation.

.4 Quasiparticle Description of Spin Soliton

Refer to caption
Figure 8: The evolution of soliton velocity vsv_{\rm s}, background density velocity vbv_{\rm b}, and relative velocity vrv_{\rm r} for spin soliton driven by a weak force FF in a toroidal Bose condensate. The dots and lines correspond to the results of numerical simulation and quasiparticle theory, respectively. The parameters are the same as those in Fig. 3 in main text.

Considering a spin soliton along the ring, the dynamics of quasi-one-dimensional coupled system can be given by

iψdst=\displaystyle i\frac{\partial\psi_{\rm ds}}{\partial t}= [122R2θ2+gds|ψds|2+gcp|ψbs|2]ψds,\displaystyle\left[-\frac{1}{2}\frac{\partial^{2}}{R^{2}\partial\theta^{2}}+g_{\rm ds}|\psi_{\rm ds}|^{2}+g_{\rm cp}|\psi_{\rm bs}|^{2}\right]\psi_{\rm ds}, (11a)
iψbst=\displaystyle i\frac{\partial\psi_{\rm bs}}{\partial t}= [122R2θ2+gcp|ψds|2+gbs|ψbs|2FRθ]ψbs.\displaystyle\left[-\frac{1}{2}\frac{\partial^{2}}{R^{2}\partial\theta^{2}}+g_{\rm cp}|\psi_{\rm ds}|^{2}+g_{\rm bs}|\psi_{\rm bs}|^{2}-FR\theta\right]\psi_{\rm bs}. (11b)

We try to use the Lagrangian variational method to evaluate the dynamics of soliton by introducing the following trial wavefunctions (taking cs=1c_{\rm s}=1) ZhaoPRA2020 ; MengPRA2022

ψds(t)=\displaystyle\psi_{\rm ds}(t)= {i1p(t)2+p(t)tanh[Rθθc(t)w(t)]}ei[γ0(t)+vb(t)Rθ],\displaystyle\left\{i\sqrt{1-p(t)^{2}}+p(t)\tanh\left[R\frac{\theta-\theta_{\rm c}(t)}{w(t)}\right]\right\}e^{i[\gamma_{0}(t)+v_{\rm b}(t)R\theta]}, (12a)
ψbs(t)=\displaystyle\psi_{\rm bs}(t)= p(t)sech[Rθθc(t)w(t)]ei{γ1(t)+γ2(t)R[θθc(t)]},\displaystyle p(t){\rm sech}\left[R\frac{\theta-\theta_{\rm c}(t)}{w(t)}\right]e^{i\left\{\gamma_{1}(t)+\gamma_{2}(t)R[\theta-\theta_{\rm c}(t)]\right\}}, (12b)

where p(t)p(t) accounts for the depth of dark soliton, w(t)w(t) for its width, and θc(t)\theta_{\rm c}(t) for its position angle. Considering the single-valued order parameter whenever a closed loop is followed, we need introduce the continuity condition Δϕ(t)+2πRvb(t)=2πQw(t)\Delta\phi(t)+2\pi Rv_{\rm b}(t)=2\pi Q_{\rm w}(t). Then the Lagrangian of the system can be written as

L(t)\displaystyle L(t) =\displaystyle= ππ{i2[(ψdstψdsψdstψds)(11|ψds|2)+(ψbstψbsψbstψbs)]\displaystyle\int_{-\pi}^{\pi}\bigg{\{}\frac{i}{2}\bigg{[}\bigg{(}\psi_{\rm ds}^{*}\frac{\partial}{\partial t}\psi_{\rm ds}-\psi_{\rm ds}\frac{\partial}{\partial t}\psi_{\rm ds}^{*}\bigg{)}\bigg{(}1-\frac{1}{|\psi_{\rm ds}|^{2}}\bigg{)}+\bigg{(}\psi_{\rm bs}^{*}\frac{\partial}{\partial t}\psi_{\rm bs}-\psi_{\rm bs}\frac{\partial}{\partial t}\psi_{\rm bs}^{*}\bigg{)}\bigg{]} (13)
[12(|ψdsRθ|2+|ψbsRθ|2)+gds2(|ψds|21)2+gbs2|ψbs|4+gcp(|ψds|21)|ψbs|2FRθ|ψbs|2]}Rdθ\displaystyle\ \ \ \ \ \ \ \ -\bigg{[}\frac{1}{2}\bigg{(}\bigg{|}\frac{\partial\psi_{\rm ds}}{R\partial\theta}\bigg{|}^{2}+\bigg{|}\frac{\partial\psi_{\rm bs}}{R\partial\theta}\bigg{|}^{2}\bigg{)}+\frac{g_{\rm ds}}{2}\big{(}|\psi_{\rm ds}|^{2}-1\big{)}^{2}+\frac{g_{\rm bs}}{2}|\psi_{\rm bs}|^{4}+g_{\rm cp}\big{(}|\psi_{\rm ds}|^{2}-1\big{)}|\psi_{\rm bs}|^{2}-FR\theta|\psi_{\rm bs}|^{2}\bigg{]}\bigg{\}}Rd\theta
=\displaystyle= 2{arcsin[p(t)]p(t)1p(t)2}Rθ˙c(t)+2p(t)2w(t)[γ˙0(t)γ˙1(t)+γ2(t)Rθ˙c(t)12γ2(t)2]\displaystyle\ 2\{\arcsin\big{[}p(t)\big{]}-p(t)\sqrt{1-p(t)^{2}}\big{\}}R\dot{\theta}_{\rm c}(t)+2p(t)^{2}w(t)\bigg{[}\dot{\gamma}_{0}(t)-\dot{\gamma}_{1}(t)+\gamma_{2}(t)R\dot{\theta}_{\rm c}(t)-\frac{1}{2}\gamma_{2}(t)^{2}\bigg{]}
+Φ[p(t),w(t),t]p(t)2w(t)+2p(t)2w(t)FRθc(t),\displaystyle+\Phi\big{[}p(t),w(t),t\big{]}-\frac{p(t)^{2}}{w(t)}+2p(t)^{2}w(t)FR\theta_{\rm c}(t),

where Φ[p(t),w(t),t]=p(t)1p(t)2[2πQw(t)Δϕ(t)πR]+[p(t)2w(t)πR][2πQw(t)Δϕ(t)2πR]2\Phi\big{[}p(t),w(t),t\big{]}=p(t)\sqrt{1-p(t)^{2}}\big{[}\frac{2\pi Q_{\rm w}(t)-\Delta\phi(t)}{\pi R}\big{]}+\big{[}p(t)^{2}w(t)-\pi R\big{]}\big{[}\frac{2\pi Q_{\rm w}(t)-\Delta\phi(t)}{2\pi R}\big{]}^{2}. It is now straightforward to apply the Euler-Lagrangian equation ddt[L(t)α˙(t)]=[L(t)α(t)]\frac{d}{dt}\big{[}\frac{\partial L(t)}{\partial\dot{\alpha}(t)}\big{]}=\big{[}\frac{\partial L(t)}{\partial\alpha(t)}\big{]}, where α(t)={p(t),w(t),θc(t),γ0(t),γ1(t),γ2(t)}\alpha(t)=\{p(t),\ w(t),\ \theta_{\rm c}(t),\ \gamma_{0}(t),\ \gamma_{1}(t),\ \gamma_{2}(t)\}. Solving them, the most essential relations for soliton motion can be derived as follows

ddt{arcsin[p(t)]p(t)1p(t)2+p(t)2w(t)Rθ˙c(t)}=p(t)2w(t)F,\displaystyle\frac{d}{dt}\big{\{}\arcsin\big{[}p(t)\big{]}-p(t)\sqrt{1-p(t)^{2}}+p(t)^{2}w(t)R\dot{\theta}_{\rm c}(t)\big{\}}=p(t)^{2}w(t)F, (14)

and

Φ[p(t),w(t),t]p(t)2w(t)p(t)Φ[p(t),w(t),t]w(t)=4p(t)2Rθ˙c(t)1p(t)24p(t)w(t),\displaystyle\frac{\partial\Phi[p(t),w(t),t]}{\partial p(t)}-\frac{2w(t)}{p(t)}\frac{\partial\Phi[p(t),w(t),t]}{\partial w(t)}=\frac{4p(t)^{2}R\dot{\theta}_{\rm c}(t)}{\sqrt{1-p(t)^{2}}}-\frac{4p(t)}{w(t)}, (15)

where 2p(t)2w(t)=Nbs2p(t)^{2}w(t)=N_{\rm bs} denotes the particle number of bright solitons. In order to describe the evolution of solitons in Fig. 3 of main text, we firstly numerically calculate p(t)p(t) from Eqs. (14) and (15); then, we can obtain the soliton velocity vs(t)=Rθ˙c(t)v_{\rm s}(t)=R\dot{\theta}_{\rm c}(t) from Eq. (14). The background density velocity can be calculated through vb(t)=[2πQw(t)Δϕ(t)]/2πRv_{\rm b}(t)=[2\pi Q_{\rm w}(t)-\Delta\phi(t)]/2\pi R. Thus, we can directly get the relative velocity vr=vsvbv_{\rm r}=v_{\rm s}-v_{\rm b}, which is presented as the red lines in Fig. 3 (b) of main text. As shown in Fig. 8, results of quasiparticle theory are in agree with those of numerical simulation.

SM C: Weak Force-Driven Multiple Dark Solitons Protocol for Manipulating Winding Number

We would like to further discuss the abundant manipulations of winding number QwQ_{w} by considering more solitons in a toroidal Bose condensate.

.5 Weak Force-Driven Two Solitons in Solitons-Impurities System

Considering two well-separated dark solitons coupled with impurity atoms, the initial states can be expressed as

ψtds\displaystyle\psi_{\rm tds} =\displaystyle= {ivr1c+1vr12c2tanh[c2vr12R(θ+π2)]}{ivr2c+1vr22c2tanh[c2vr22R(θπ2)]}\displaystyle\bigg{\{}i\frac{v_{\rm r1}}{c}+\sqrt{1-\frac{v_{\rm r1}^{2}}{c^{2}}}\tanh\left[\sqrt{c^{2}-v_{\rm r1}^{2}}R\left(\theta+\frac{\pi}{2}\right)\right]\bigg{\}}\bigg{\{}i\frac{v_{\rm r2}}{c}+\sqrt{1-\frac{v_{\rm r2}^{2}}{c^{2}}}\tanh\left[\sqrt{c^{2}-v_{\rm r2}^{2}}R\left(\theta-\frac{\pi}{2}\right)\right]\bigg{\}} (16)
×ei[(vb1+vb2)θ+(vb1vb2)π/2]Rψgs,\displaystyle\times e^{i[(v_{\rm b1}+v_{\rm b2})\theta+(v_{\rm b1}-v_{\rm b2})\pi/2]R}\psi_{gs},
ψtim\displaystyle\psi_{\rm tim} =\displaystyle= ε{sech[c2vr12R(θ+π2)]eivs1R(θ+π/2)+sech[c2vr22R(θπ2)]eivs2R(θπ/2)}ψgs,\displaystyle\varepsilon\bigg{\{}{\rm sech}\left[\sqrt{c^{2}-v_{\rm r1}^{2}}R\left(\theta+\frac{\pi}{2}\right)\right]e^{iv_{\rm s1}R(\theta+\pi/2)}+{\rm sech}\left[\sqrt{c^{2}-v_{\rm r2}^{2}}R\left(\theta-\frac{\pi}{2}\right)\right]e^{iv_{\rm s2}R(\theta-\pi/2)}\bigg{\}}\psi_{gs}, (17)

Two solitons-impurities are driven by the sinusoidal driving forces F1=(1)t/T1F01sin(2πt/T1)F_{1}=(-1)^{\lfloor t/T_{1}\rfloor}F_{01}\sin(2\pi t/T_{1}) and F2=(1)t/T2F02sin(2πt/T2)F_{2}=(-1)^{\lfloor t/T_{2}\rfloor}F_{02}\sin(2\pi t/T_{2}), respectively. The amplitudes of force are set to F01=F02=0.03F_{01}=F_{02}=-0.03 in Fig. 9 (a) and F01=F02=0.03F_{01}=-F_{02}=-0.03 in Fig. 9 (b), respectively. One can investigate different evolutions of winding numbers by adjusting the periodic rate T1/T2T_{1}/T_{2}. For T1/T2=1T_{1}/T_{2}=1, the winding number QwQ_{\rm w} jumps two times that of the one dark soliton (see in Fig. 1 in in main text) when F01=F02F_{01}=F_{02}; or there are no change for winding number QwQ_{\rm w} when F01=F02F_{01}=-F_{02}. For T1/T21T_{1}/T_{2}\neq 1, the winding number QwQ_{\rm w} will evolve in a near random way due to the different frequencies and directions of the density zero crossings, when both F01=±F02F_{01}=\pm F_{02}.

Refer to caption
Figure 9: The scheme for two well-separated dark solitons with coupling impurity atoms driven by forces F1,2F_{1,2} in a toroidal Bose condensate, and evolutions of winding number QwQ_{\rm w}. The periods of sinusoidal force are both set to T1=300T_{1}=300 and T2=(300, 150, 200)T_{2}=(300,\ 150,\ 200). In (a), the parameters are F01=F02=0.03F_{01}=F_{02}=-0.03, vr1|t=0=vr2|t=00.01v_{\rm r1}|_{t=0}=v_{\rm r2}|_{t=0}\approx 0.01 (with vs1|t=0=vs2|t=0=0.02v_{\rm s1}|_{t=0}=v_{\rm s2}|_{t=0}=0.02 and vb1|t=0=vb2|t=00.01v_{\rm b1}|_{t=0}=v_{\rm b2}|_{t=0}\approx 0.01). In (b), the parameters are F01=F02=0.03F_{01}=-F_{02}=-0.03, vr1|t=0=vr2|t=00.01v_{\rm r1}|_{t=0}=-v_{\rm r2}|_{t=0}\approx 0.01 (with vs1|t=0=vs2|t=0=0.02v_{\rm s1}|_{t=0}=-v_{\rm s2}|_{t=0}=0.02 and vb1|t=0=vb2|t=00.01v_{\rm b1}|_{t=0}=-v_{\rm b2}|_{t=0}\approx 0.01). The other parameters are gds=gcp=1g_{\rm ds}=g_{\rm cp}=1, gim=0g_{\rm im}=0, n=1n=1, ε=1/10\varepsilon=1/\sqrt{10}, ω=1\omega=1, and R=50R=50.

.6 Weak Force-Driven Two Spin Solitons

Considering two well-separated spin solitons, the initial states can be expressed as

ψtds\displaystyle\psi_{\rm tds} =\displaystyle= {ivrcs+1vr2cs2tanh[cs2vr2R(θ+π6)]}{ivrcs+1vr2cs2tanh[cs2vr2R(θπ6)]}e2ivbRθψgs,\displaystyle\bigg{\{}i\frac{v_{\rm r}}{c_{\rm s}}+\sqrt{1-\frac{v_{\rm r}^{2}}{c_{\rm s}^{2}}}\tanh\left[\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\left(\theta+\frac{\pi}{6}\right)\right]\bigg{\}}\bigg{\{}i\frac{v_{\rm r}}{c_{\rm s}}+\sqrt{1-\frac{v_{\rm r}^{2}}{c_{\rm s}^{2}}}\tanh\left[\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\left(\theta-\frac{\pi}{6}\right)\right]\bigg{\}}e^{2iv_{\rm b}R\theta}\psi_{gs}, (18)
ψtbs\displaystyle\psi_{\rm tbs} =\displaystyle= 1vr2cs2{sech[cs2vr2R(θ+π6)]eivsR(θ+π/6)+sech[cs2vr2R(θπ6)]eivsR(θπ/6)}ψgs.\displaystyle\sqrt{1-\frac{v_{\rm r}^{2}}{c_{\rm s}^{2}}}\bigg{\{}{\rm sech}\left[\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\left(\theta+\frac{\pi}{6}\right)\right]e^{iv_{\rm s}R(\theta+\pi/6)}+{\rm sech}\left[\sqrt{c_{\rm s}^{2}-v_{\rm r}^{2}}R\left(\theta-\frac{\pi}{6}\right)\right]e^{iv_{\rm s}R(\theta-\pi/6)}\bigg{\}}\psi_{gs}. (19)

Two spin solitons are both driven by a weak constant force F=0.005F=0.005. One can see that the winding number QwQ_{\rm w} jumps two times that of the one spin soliton in the step-like behavior (see in Fig. 3 in in main text), as shown in Fig. 10. More manipulation of winding number can be realized by involving multiple spin solitons.

Refer to caption
Figure 10: The scheme for two well-separated spin solitons driven by a weak constant force F=0.005F=0.005 in a toroidal Bose condensate, and the evolution of winding number QwQ_{\rm w}. The parameters are gds=3g_{\rm ds}=3, gcp=2g_{\rm cp}=2, gbs=1g_{\rm bs}=1, vr|t=00.01v_{\rm r}|_{t=0}\approx-0.01 (with vs|t=0=0.02v_{\rm s}|_{t=0}=-0.02 and vb|t=00.01v_{\rm b}|_{t=0}\approx-0.01), ω=3\omega=3, and R=50R=50.

References