1.Center of Theoretical Physics, College of Physical Science and Technology, Sichuan University, Chengdu 610065, China 2.School of Material Science and Engineering, Chongqing Jiaotong University, Chongqing 400074, China
Abstract:Exact solution to the driven quantum system with an explicitly time-dependent Hamiltonian is not only an issue of fundamental importance to quantum mechanics itself, but also a ubiquitous problem in the design for quantum control. In particular, the nonadiabatic transition induced by the time-dependent external field is often involved in order to target the quantum state for the atomic and molecular systems. In this paper we investigate the exact dynamics and the associated nonadiabatic transition in a typical driven model, the Rosen-Zener model and its multi-level extension, by virtue of the algebraic dynamical method. Previously, this kind of driven models, especially of the two-level case, were solved by converting the corresponding Schr?dinger equation to a hypergeometric equation. The property of the dynamical transition of the system was then achieved by the asymptotic behavior of the yielded hypergeometric function. A critical drawback related to such methods is that they are very hard to be developed so as to treat the multi-level extension of the driven model. Differing from the above mentioned method, we demonstrate that the particular kind of the Rosen-Zener model introduced here could be solved analytically via a canonical transformation or a gauge transformation approach. In comparison, we show that the present method at least has two aspects of advantages. Firstly, the method enables one to describe the evolution of the wavefunction of the system analytically over any time interval of the pulse duration. Moreover, we show that the method could be exploited to deal with the multi-level extensions of the model. The explicit expression of the dynamical basis states, including the three-level system and the four-level system, is presented and the transition probabilities induced by the nonadiabatic evolution among different levels are then characterized for the model during the time evolution. In addition, our study reveals further that the dual model of the driven system can be constructed. Since the dynamical invariant of a solvable system can always be obtained within the framework of the algebraic dynamical method, the general connection between the dual model and the original one, including the solvability and their dynamical invariants, are established and characterized distinctly. Keywords:Rosen-Zener model/ canonical transformation/ nonadiabaticity-induced transition/ dynamical invariant
全文HTML
--> --> --> -->
2.1.原有方法回顾
考虑如下的二能级系统哈密顿量:
$ H\left( t \right) = \left( \begin{array}{cc} \varepsilon & V\left( t \right) \\ V\left( t \right) & -\varepsilon \end{array} \right), $
其中$ V\left( t \right) $描述横向磁场, ε为正常数, 描述裸原子能级. 原则上, 形为(1)式的哈密顿量其薛定谔方程可以通过如下过程化为二阶常微分方程. 设系统波函数为$\left| {\psi \left( t \right)} \right\rangle = {c_1}\left( t \right){{\rm e}^{ -{\rm i} \varepsilon t }}\left| \uparrow \right\rangle + {c_2}\left( t \right){{\rm e}^{{\rm i} \varepsilon} t}\left| \downarrow \right\rangle $, 其中$ c_1\left( t \right) $与$ c_2\left( t \right) $为待定系数. 将其代入薛定谔方程(令$ \hbar = 1 $):
$ {\rm i}\frac {\partial}{\partial t}\left| {\psi \left( t \right)} \right\rangle = H\left( t \right)\left| {\psi \left( t \right)} \right\rangle , $
容易得系数$ c_{1,2}\left( t \right) $满足如下方程:
$ {\rm i}\dot c_1\left( t \right) = V\left( t \right){\rm e}^{{\rm i}2\varepsilon t}c_2\left( t \right), $
$ {\rm i}\dot c_2\left( t \right) = V\left( t \right){\rm e}^{-{\rm i}2\varepsilon t}c_1\left( t \right). $
$ \ddot{c_1}\left( t \right)-\left[{\rm i}2\varepsilon +\frac{\dot{V}\left( t \right)}{V\left( t \right)}\right]\dot{c_1}\left( t \right)+V^2\left( t \right)c_1\left( t \right) = 0. $
在$ t\rightarrow +\infty $时, 显然系统仍将回到$ \left| \uparrow \right\rangle $态上, 也就是说经过整体演化跃迁几率为零. 这与(6)式结果(须取$ 1/\tau = \nu $)一致. 值得指出, 上述解析解使得我们可以描述系统在演化过程中的非绝热跃迁. 定义${F_ \pm }\left( t \right) \equiv $${\left| {\left\langle {\psi _{ \pm \frac{1}{2}}^{\rm {ad}}\left( t \right)\left| {{\psi _{\frac{1}{2}}}\left( t \right)} \right.} \right\rangle } \right|^2}$, 其中$ \left| {\psi^{{\rm {ad}}}_{\pm\frac 12}\left( t \right)} \right\rangle $表征系统哈密顿量(7)式的瞬时绝热本征态. 由于$ t\rightarrow -\infty $时$ \left| {\psi_{\pm\frac 12}\left( t \right)} \right\rangle $与$ \left| {\psi^{{\rm {ad}}}_{\pm\frac 12}\left( t \right)} \right\rangle $一致, 故$ F_+\left( t \right) $描述了演化过程中绝热态的保留几率, $ F_-\left( t \right) $则描述非绝热效应所致的绝热态之间的跃迁几率. 图1分别画出了驱动过程中上述布居数$ P\left( t \right) $以及非绝热效应$ F_{\pm}\left( t \right) $ 随时间的演变. 图 1 (a)二能级系统布居数随时间演化; (b)绝热态的保留几率$F_+(t)$以及非绝热跃迁几率$F_-(t)$. 系统初态均取$|\uparrow\rangle$ Figure1. (a) Time evolution of the population of the two-level system; (b) the survival probability $F_+(t)$ of the adiabatic state and the transition probability $F_-(t)$ induced by the nonadiabaticity. In both cases the initial state of the system is in $|\uparrow\rangle$.
-->
3.1.规范变换及系统的动力学演化
下面将上一节中提出的正则变换求解方法推广到相应的多能级系统. 为此, 研究如下哈密顿量:
$ H\left( t \right) = {{{{\varOmega}}}\left( t \right)\cdot {{{J}}}} = 2\nu {\rm{sech}}\left( {\nu t} \right)J_x+\nu J_z, $
$ \left| {\psi \left( t \right)} \right\rangle = G\left( t \right)\left| {\psi^g\left( t \right)} \right\rangle = {\rm e}^{{\rm i}\alpha\left( t \right)J_x}{\rm e}^{{\rm i}\beta\left( t \right)J_y}\left| {\psi^g\left( t \right)} \right\rangle. $
在新表象中容易得到$ \left| {\psi^g\left( t \right)} \right\rangle $满足协变薛定谔方程$ {\rm i}\partial _t\left| {\psi^g\left( t \right)} \right\rangle = H_g\left( t \right)\left| {\psi^g\left( t \right)} \right\rangle $, 相应有效哈密顿量$ H_g\left( t \right) $可表达为
$ H_g\left( t \right) = G^\dagger\left( t \right)H\left( t \right)G\left( t \right)-iG^\dagger\left( t \right)\partial_tG\left( t \right) = {{{X}}\left( t \right)\cdot{{J}}}, $
其中$ {{{X}}\left( t \right) }$三个分量的具体形式为:
$ \begin{aligned} &X_1\left( t \right) =\dot\alpha\cos\beta+2\nu {\rm{sech}} \left( {\nu t} \right)\cos\beta+\nu\cos\alpha\sin\beta,\\ &X_2\left( t \right) = \dot\beta-\nu\sin\alpha,\\ &X_3\left( t \right) = -\dot\alpha\sin\beta-2\nu {\rm{sech}} \left( {\nu t} \right)\sin\beta+\nu\cos\alpha\cos\beta . \end{aligned}$
容易验证, 如果取
$ \alpha\left( t \right) = \frac {\text{π}} 2-\arcsin\left[ {\tanh\left( {\nu t} \right)} \right],\; \; \beta\left( t \right) = 2{\text{π}}-\alpha\left( t \right), $
可以得到$ X_1\left( t \right) = X_2\left( t \right) = 0 $以及$ X_3\left( t \right) = \nu $, 从而
由上述结果可以看出, 类似于二能级情况, 假如系统初始时刻$ t\rightarrow-\infty $处在某一状态$ \left| {\pm 1} \right\rangle $或$ \left| {0} \right\rangle $, 在$ t\rightarrow+\infty $时系统将回到初始状态($ \left| {\pm} \right\rangle $两种状态会多出一相位). 同理, 可以定义跃迁矩阵 $ {F_{mn}}\left( t \right) \equiv {\left| {\left\langle {\psi _m^{\rm {ad}}\left( t \right)\left| {{\psi _n}\left( t \right)} \right.} \right\rangle } \right|^2}$, 其对角元素描述瞬时绝热态的保留几率, 非对角元则代表非绝热跃迁. 在图2中展示了不同初始状态下系统在演化过程中的跃迁几率. 图 2 (a)初态为$|1\rangle$态时系统演化过程中的非绝热跃迁$F_{m1}(t)$$(m=0,\pm 1)$. 其中$|1\rangle \rightarrow |-1\rangle$跃迁几率非常小, 在$t=0$时$F_{-11}\approx 0.0028$; (b)初态为$|0\rangle$态时系统演化过程中的非绝热跃迁$F_{m0}(t)$, 其中$F_{10}(t)=F_{-10}(t)$ Figure2. (a) Nonadiabaticity-induced transition of the initial state $|1\rangle$ during the evolution. The transition probability from $|1\rangle $ to $|-1\rangle$ is very small with $F_{-11}\approx 0.0028$ at $t=0$; (b) nonadiabaticity-induced transition of the initial state $|0\rangle$ during the evolution, where $F_{10}(t)=F_{-10}(t)$.
$ \begin{split} I\left( t \right)& = G\left( t \right)J_zG^{\dagger}\left( t \right) \\ & = -\sin\beta\left( t \right) J_x+\cos\beta\left( t \right)[ \sin\alpha\left( t \right) J_y\\ &\quad +\cos\alpha \left( t \right)J_z ], \end{split} $
其中$ \alpha \left( t \right) $,$ \beta\left( t \right) $由方程(21)式给出. 欲检验$ I\left( t \right) $是否满足
$ {\rm i}\frac {\partial} {\partial t}I\left( t \right) = \left[ {H\left( t \right),I\left( t \right)} \right], $
可考察其分量方程. 记$ I\left( t \right)\equiv {{{R}}\left( t \right)\cdot{{J}}} $, 根据(21)式和(31)式, 可知
$ \begin{split} \dot{R}_x\left( t \right)& = -\varOmega_zR_y\left( t \right),\\ \dot{R}_y\left( t \right)& = \varOmega_zR_x\left( t \right)-\varOmega_x\left( t \right)R_z\left( t \right), \\ \dot{R}_z\left( t \right)& = \varOmega_x\left( t \right)R_y\left( t \right). \end{split} $
根据(31)式, $I{\left( t \right)}$的本征态可表示为$\left| {\phi_m{\left( t \right)}} \right\rangle = $$G{\left( t \right)}\left| {m} \right\rangle ={\rm e}^{{\rm i}\alpha{\left( t \right)}J_x}{\rm e}^{{\rm i}\beta{\left( t \right)}J_y}\left| {m} \right\rangle $. 对比方程(24)式可以看到, $\left| {\phi_m{\left( t \right)}} \right\rangle $与$\left| {\psi_m{\left( t \right)}} \right\rangle $仅相差一相位, 此即所谓的Lewis总相位[1,2]:
$ H^\prime \left( t \right) = {{{\varOmega}}^\prime\left( t \right)\cdot {{J}}} = -\nu{\rm{sech}}\left( {\nu t} \right)J_x+\nu J_z. $
容易检验, 该系统具有如下动力学不变量
$ \begin{split}I^\prime\left( t \right) = \;&\sin\beta\left( t \right)J_x+\cos\beta\left( t \right)\\&\times\left[ {-\sin\alpha\left( t \right)J_y+\cos\alpha\left( t \right)J_z} \right], \end{split}$
其中$ \alpha\left( t \right) $, $ \beta\left( t \right) $仍由方程(21)式给出. 这一结果可以通过对上面动力学不变量分量方程(34)式直接观察得到. 实际上, 这样的对偶变换对两分量形式哈密顿量是普适的. 由于哈密顿量$ H^\prime\left( t \right) $与$ H\left( t \right) $仅x分量相差一负号$ \varOmega_x^\prime\left( t \right) = -\varOmega_x\left( t \right) $, 只要将不变量算子x, y分量做替换$ R_x\left( t \right)\rightarrow -R_x\left( t \right) $, $ R_y\left( t \right)\rightarrow $$-R_y\left( t \right) $, 则分量方程(34)式仍能成立. 上述变换$ I\left( t \right)\rightarrow I^\prime\left( t \right) $ 相当于将角度参数$ \alpha\left( t \right) $, $ \beta\left( t \right) $换成$ -\alpha\left( t \right) $与$ -\beta\left( t \right) $. 也就是说, 对于上面的对偶哈密顿量(36)式, 可以采用规范变换$ G^\prime \left( t \right) = {\rm e}^{-{\rm i}\alpha\left( t \right)J_x}$$\cdot{\rm e}^{-{\rm i}\beta\left( t \right)J_y} $, 前述代数动力学求解方法依然有效.