1: %
2: %\documentstyle[preprint,aps]{revtex} % !!!
3: \documentstyle[epsfig,aps]{revtex}
4: %\documentstyle[twocolumn,aps]{revtex}
5: \setlength{\topmargin}{-0.05in}
6: \begin{document}
7: \draft
8: \title{Motional effects of single trapped atomic/ionic qubit}
9: \author{L. You}
10: \address{School of Physics, Georgia Institute of Technology,
11: Atlanta, Georgia 30332-0430, USA}
12:
13: \date{\today}
14: \maketitle
15: \begin{abstract}
16:
17: We investigate theoretical decoherence effects of the
18: motional degrees of freedom of a single trapped atomic/ionic
19: electronically coded qubit.
20: For single bit rotations from a resonant running wave
21: laser field excitation, we found the achievable
22: fidelity to be determined by a single parameter
23: characterized by the motional states.
24: Our quantitative results provide a useful realistic view
25: for current experimental efforts in quantum information and computing.
26:
27: \end{abstract}
28: \pacs{03.67.Lx, 89.70.+c, 32.80.-t}
29:
30: \narrowtext
31:
32: Since the pioneering work by Shor \cite{Shor} on efficient prime factorization
33: with a quantum computer in 1994, we have witnessed an explosive growth of
34: interests in quantum information and computing. Although still largely a
35: theoretical field, solid progress in experimental efforts have been made
36: within the last few years. Most notably, pure state based
37: quantum gate implementation \cite{nist}, demonstrations of
38: quantum teleportation \cite{tele}, and GHZ state synthesis \cite{anton}
39: have stimulated more vigorous experimental efforts.
40:
41: Two of the most interesting proposals for potentially large scale
42: quantum computing were suggested by the same
43: Innsbruck group, based on trapped ions \cite{Cirac}
44: and cavity QED with atoms \cite{Cirac2}. Various
45: implementations of these ideas are actively pursued in many
46: experiments around the world.
47: In their original analysis as presented in \cite{Cirac,Cirac2},
48: individual qubits are coded in electronic degrees states of atoms/ions,
49: and coherent evolution of the system state
50: requires the qubits to be in selected motional pure states.
51: Experimentally, one needs to attain the
52: strong binding limit \cite{side} and cooling to
53: the motional ground state is also required \cite{monroe}.
54: As is well known, these limits are difficult to maintain
55: due to various decoherence processes \cite{bf}
56: which heat up the motional degrees of freedom.
57: Further more, maintaining motional ground state becomes
58: problematic when strong confinement is not satisfied.
59: Several ideas were proposed recently for computing with
60: `hot' qubits \cite{Cirac3}.
61:
62: This paper attempts to provide quantitative answers to
63: motional effects (ME) on electronically encoded quantum states \cite{Milburn}.
64: It is the first step towards a thorough investigation of
65: the ME. The paper is organized as follows.
66: First our model is presented.
67: We then discuss analytically the decoherence of the ME
68: for an unknown electronic encoded qubit.
69: Finally we present numerical results to support our understanding.
70: In forthcoming papers, we will study the decoherence
71: due to ME on multi-qubit entanglement creation,
72: e.g, the effect on the conditional logic
73: operation CNOT.
74:
75: We consider a single harmonically bound two state atom
76: described by the Hamiltonian \cite{jas,zeng,vogel}
77: \begin{eqnarray}
78: H&&=\sum_{\vec n}\hbar(n_x\omega^g_x+
79: n_y\omega^g_y+n_z\omega^g_z)|g,\vec n\rangle\!\langle g,\vec n|\nonumber\\
80: &&+\sum_{\vec m}\hbar(\omega_{eg}+m_x\omega^e_x+
81: m_y\omega^e_y+m_z\omega^e_z)|e,\vec m\rangle\!\langle e,\vec m|\nonumber\\
82: &&+{1\over 2}\hbar\Omega_L e^{i\omega_L t}\sum_{\vec n,\vec m}
83: \eta_{\vec n\vec m}(\vec k_L)|g,\vec n\rangle\!\langle e, \vec m|
84: +h.c.\ .
85: \label{H_1}
86: \end{eqnarray}
87: $|g,\vec n\rangle=|g\rangle|\vec n\rangle_g$
88: ($|e,\vec m\rangle=|e\rangle|\vec m\rangle_e$) denotes
89: number state in the ground (excited) trap with
90: frequencies are $\omega^g_{i=x,y,z}$ ($\omega^e_i$).
91: $\omega_{eg}$ is the electronic transition frequency.
92: $\Omega_L$ is the Rabi
93: frequency of the plane wave laser field.
94: The motional dipole moments are the
95: familiar Franck-Condon factor \cite{you}
96: \begin{equation}
97: \eta_{\vec n\vec m}({\vec k}_L)= \langle{g},{\vec n}|e^{-i{\vec k}_L\cdot{\vec R}
98: } |{e},{\vec m}\rangle.
99: \label{fc}
100: \end{equation}
101:
102: Radiative coupling to the vacuum reservoir of the atom
103: will not be included here as the effect of the
104: resulting spontaneous emission on the qubit decoherence
105: has been studied and is well understood \cite{Cirac,Cirac2,Cirac3,walls}.
106: Our model can also be viewed as between two ground states
107: in a three level $\Lambda$-type off resonant Raman system.
108: In such a case,
109: $\Omega_L\sim\Omega_P\Omega_S^*/\delta_L$ are the two photon
110: effective Rabi frequency and $\vec k_L=\vec k_P-\vec k_S$.
111: The indices $P$ and $S$
112: denote the dipole connected pump and stokes transitions,
113: and $\delta_L$ is the (large) detuning
114: from the eliminated far off-resonant excited state.
115:
116: Physical models similar to Eq. (\ref{H_1}) have been
117: studied under different context before \cite{Cirac,side,zeng,vogel,gorigi},
118: usually, within the Lamb-Dicke limit (LDL) when
119: ME become considerably simplified. In terms of the trap width
120: $a_i^{g,e}=\sqrt{\hbar/2M\omega_i^{g,e}}$,
121: the LDL corresponds to $k_La_i^{g,e}\ll 1$.
122: This requires the atom/ion to be confined less than the
123: wavelength $a_i^{g,e}\ll\lambda_L$ and is equivalent to
124: require $\hbar\omega_i^g$ to be much larger
125: than the recoil energy $E_R=\hbar^2k_L^2/2M\ll \hbar\omega_i^{e,g}$.
126: For an effective two state system reduced from a
127: near resonant three level $\Lambda$-type configuration,
128: LDL is easily satisfied with co-propagating
129: pump and Stokes fields when $\vec k_P\sim\vec k_S$.
130:
131: In this study we investigate ME for general cases not in the LDL.
132: Such studies will provide
133: much needed theoretical clarification as trapped atoms/ions are
134: among the ``hottest'' qubit candidates in many experimental efforts.
135: We note Franck-Condon factors (\ref{fc}) satisfy
136: \begin{eqnarray}
137: &&\sum_{\vec n}\big[\eta_{\vec n\vec m}({\vec k}_L)\big]^{*} \eta_{\vec n%
138: \vec m^{\prime}}({\vec k}_L)=\delta_{\vec m\vec m^{\prime}}, \nonumber\\
139: &&\sum_{\vec m}\big[\eta_{\vec n\vec m}({\vec k}_L)\big]^{*} \eta_{\vec n%
140: ^{\prime}\vec m}({\vec k}_L)=\delta_{\vec n\vec n^{\prime}},
141: \end{eqnarray}
142: which allows the introduction of a complete and orthonormal basis \cite{you}
143: \begin{eqnarray}
144: |\vec n\rangle_p = \sum_{\vec n'}\eta_{\vec n\vec n'}^*({\vec k}_L)|\vec n'\rangle_g.
145: \label{wp}
146: \end{eqnarray}
147: Physically it corresponds to the motional wave packet
148: of a photon absorption from ground state $|\vec n\rangle_g$.
149: Mathematically, it is the number
150: coherent state basis $|\alpha,n\rangle=D(\alpha)|n\rangle$, with
151: the displacement operator $D(\alpha)=e^{\alpha b^+-\alpha^* b}$.
152: We note $R_i=a_i^g(b_i^{g+}+b_i^g)$
153: with $b_i^g$ ($b_i^{g+}$) the annihilation (creation)
154: operator for$|n_i\rangle_g$, therefore Eq. (\ref{wp}) can be
155: rewritten as $|n_x\rangle_p =|ik_La_x^g,n_x\rangle_g$, exactly
156: representing wave-packets
157: corresponding to excitation from different ground trapping
158: states. As will become
159: clear later, the coherent Rabi coupling between
160: the ground and excited state manifolds can
161: also be decomposed into paired sets
162: $\{|g\rangle|\vec n\rangle_g,|e\rangle|\vec n\rangle_p\}$.
163:
164: With the inverse relation
165: \begin{eqnarray}
166: |\vec m\rangle_e=\sum_{\vec n}\eta_{\vec n\vec m}({\vec k}_L)|\vec n\rangle_p,
167: \end{eqnarray}
168: we can transform Eq. (\ref{H_1}) into the $|\vec n\rangle_p$ basis.
169: Denote $|e_p,\vec n\rangle=|e\rangle|\vec n\rangle_p$, we obtain
170: \begin{eqnarray}
171: \sum_{\vec m}\hbar\omega_{eg}|e,\vec m\rangle\!\langle e,\vec m|
172: &&=\sum_{\vec n}\hbar\omega_{eg}|e_p,\vec n\rangle\!\langle e_p,\vec n|,\nonumber\\
173: \sum_{\vec m,i} m_i\hbar\omega^e_i
174: |\vec m\rangle_e\!\ _e\langle\vec m|
175: =&&\sum_{\vec n\vec n'}
176: (E_{\vec n\vec n'}^D+E_{\vec n\vec n'}^O)
177: |\vec n\rangle_p\!\ _p\langle\vec n'|.
178: \end{eqnarray}
179: We found $E_{\vec n\vec n'}^D$ terms couple
180: nearest neighbors, i.e.
181: states with $n_i=n_i'\pm 1$ ($n_{j\ne i}=n'_j$),
182: while $E_{\vec n\vec n'}^O$ terms couple states with
183: $\vec n=\vec n'$
184: and $n_i=n_i'\pm 2$ ($n_{j\ne i}=n'_j$).
185: $E_{\vec n\vec n'}^O$ terms become diagonal
186: along $i-$axis whenever $\delta\omega_i^2\neq 0$.
187:
188: We will focus on the case
189: $\omega_i^g=\omega_i^e$ ($i=x,y,z$) in the
190: present paper. This is typical for the ion trap system
191: and can also be arranged for optical dipole traps \cite{jeff}.
192: For the plane wave excitation along the x-axis,
193: the ME along the $y$ and $z$ directions are unperturbed.
194: Our Hamiltonian Eq. (\ref{H_1}) can simplifies to
195: a one dimensional model
196: \begin{eqnarray}
197: H&&=\sum_{n_x}n_x\hbar\omega^g_x|g,n_x\rangle\!\langle g,n_x|\nonumber\\
198: &&+\sum_{n_x}\hbar(n_x\omega^g_x-\Delta_L)
199: |e_p,n_x\rangle\!\langle e_p,n_x|\nonumber\\
200: &&+i(k_La_x)\sum_{n_x}\hbar\omega^{g}_x \sqrt{n_x+1}
201: |e_p,n_x+1\rangle\!\langle e_p,n_x|+h.c.\nonumber\\
202: &&+{1\over 2}\hbar\Omega_L
203: \sum_{n_x}|g,n_x\rangle\!\langle e_p,n_x|+h.c. ,
204: \label{hm}
205: \end{eqnarray}
206: where transformation to the interaction picture by
207: \begin{eqnarray}
208: U(t)=\exp\left(-i\hbar\omega_Lt\sum_{n_x}|e_p,n_x\rangle\!\langle e_p,n_x|\right)
209: \end{eqnarray}
210: has also been made. The detuning is
211: $\hbar\Delta_L
212: =\hbar\omega_L-\hbar\omega_{eg}-{\hbar^2k_L^2\over 2M}$,
213: including the recoil shift.
214:
215: This Hamiltonian can be graphically illustrated as
216: in Figure \ref{fig1}. The paired
217: ladder structure resembles the familiar
218: motional state ladders in an ion trap \cite{side}.
219: However, in Fig. \ref{fig1},
220: the excited states are the wave-packet basis states Eq. (\ref{wp}).
221: The nearest neighbor coupling is not
222: due to the LDL approximation.
223: Denote $H=H_0+H_1$ with $H_1$ the nearest neighbor coupling term
224: in the excited state [the third line of Eq. (\ref{hm})], $H_0$
225: becomes
226: \begin{eqnarray}
227: H_0=\hbar\left (\begin{array}{cccccccccc}
228: \cdot & \cdot & & & & \\
229: \cdot & \cdot & & & & \\
230: & & n_x\omega_x^g & {\Omega_L\over 2} & & \\
231: & & {\Omega_L\over 2} & n_x\omega_x^g-\Delta_L & & \\
232: & & & & \cdot & \cdot \\
233: & & & & \cdot & \cdot \\
234: \end{array} \right),
235: \end{eqnarray}
236: where each $2\times2$ block describes
237: the Rabi oscillation
238: between paired states $\{|g\rangle|n_x\rangle_g,|e\rangle|n_x\rangle_p\}$
239: with {\it exactly the same} Rabi
240: frequency $\Omega=\sqrt{\Omega_L^2+\Delta_L^2}$.
241: There are no differential detunings between different pairs either.
242: With wave function coefficients $\{C_{n_x}^g,C_{n_x}^e\}$,
243: the $2\times2$ oscillation is described by
244: \begin{eqnarray}
245: C_{n_x}^e(\tau) &&=e^{-in_x\omega_x^g\tau}
246: \left[C_{n_x}^e(0)\cos\theta-iC_{n_x}^g(0)\sin\theta\right],\nonumber\\
247: C_{n_x}^g(\tau) &&=e^{-in_x\omega_x^g\tau}
248: \left[C_{n_x}^g(0)\cos\theta-iC_{n_x}^e(0)\sin\theta\right],
249: \label{twou}
250: \end{eqnarray}
251: when $\Delta_L=0$. The pulse area is
252: $\theta(\tau)={1\over 2}\int_{0}^\tau\Omega(t) dt$.
253: $H_0$ describes the coherent evolution
254: between paired states with a time scale given $\Omega$.
255: $H_1$, on the other hand couples nearest neighbors
256: of excited motional wave pack states and can causes decoherence
257: of an electronically coded qubit. Its time scale is
258: determined by several factors including the
259: trap frequency $\omega_x^g$, Lamb Dicke parameter $k_La_x^g$,
260: and the highest motional state number $n_x^{\rm max}$.
261: Assuming an electronically coded unknown qubit
262: \begin{eqnarray}
263: |\psi(0)\rangle&&=\alpha |g\rangle +\beta |e\rangle,
264: \label{qb}
265: \end{eqnarray}
266: (normalization $|\alpha|^2+|\beta|^2=1$),
267: an arbitrary single bit
268: rotation is achieved through a multiplication of \cite{barenco}
269: \begin{eqnarray}
270: e^{i\delta}\left( \begin{array}{cc}
271: e^{i\gamma_1} & 0\\
272: 0 & e^{-i\gamma_1} \end{array}\right)
273: \left( \begin{array}{cc}
274: \cos\theta & \sin\theta\\
275: -\sin\theta & \cos\theta \end{array}\right)
276: \left( \begin{array}{cc}
277: e^{i\gamma_2} & 0\\
278: 0 & e^{-i\gamma_2} \end{array}\right).\nonumber
279: \end{eqnarray}
280: $\delta$, $\gamma_1$, $\gamma_2$, and $\phi$ are parameters.
281: With resonant Rabi coupling, our solution Eq. (\ref{twou})
282: achieves the important
283: $\theta$ rotation corresponds to
284: $e^{2i\gamma_1}=-i$, and $e^{2i\gamma_2}=i$. Ideally
285: one hopes to arrive at the target state
286: \begin{eqnarray}
287: |\psi(\tau)\rangle_T&&=
288: (\alpha\cos\theta-i\beta\sin\theta) |g\rangle
289: +(\beta\cos\theta-i\alpha\sin\theta) |e\rangle,\nonumber
290: \end{eqnarray}
291: with the density matrix,
292: \begin{eqnarray}
293: \rho_T(\tau)&&=I_{g}^T|g\rangle\!\langle g|
294: +(1-I_{g}^T)|e\rangle\!\langle e|
295: +(I_{ge}^T|g\rangle\!\langle e|+h.c.),
296: \label{rhot}
297: \end{eqnarray}
298: where
299: \begin{eqnarray}
300: I_{g}^T&&=|\alpha|^2\cos^2\theta+|\beta|^2\sin^2\theta
301: +i(\alpha\beta^*-c.c.)\sin\theta\cos\theta,\nonumber\\
302: I_{ge}^T&&=\alpha\beta^*\cos^2\theta+\beta\alpha^*\sin^2\theta
303: +i(|\alpha|^2-|\beta|^2)\sin\theta\cos\theta.\nonumber
304: \end{eqnarray}
305:
306: Due to ME the electronic qubit Eq. (\ref{qb})
307: does not remain decoupled from the motional degrees of freedom.
308: In general, it evolves within the
309: much larger Hilbert space containing motional states.
310: We now study two concrete examples of
311: decoherence assuming the initial qubit
312: in the enlarged Hilbert space reproduces the
313: density matrix Eq. (\ref{rhot}), i.e. resembles
314: a perfect qubit to the innocent bystanders unaware
315: of the motional degrees of freedom.
316:
317: First, we consider
318: \begin{eqnarray}
319: |\psi(0)\rangle_{\rm tot}
320: &&=(\alpha |g\rangle + \beta|e\rangle)\otimes |\psi(0)\rangle_{\rm cm},
321: \label{psi1}
322: \end{eqnarray}
323: with an initial pure motional state
324: $|\psi(0)\rangle_{\rm cm}=\sum_{n_x}c_{n_x}|n_x\rangle_g$
325: ($\sum_{n_x}|c_{n_x}|^2=1$).
326: Upon tracing the motional degrees of freedom,
327: $\rho(0)=|\psi(0)\rangle\!\langle\psi(0)|$ is correctly
328: reproduced. By rewriting $|\psi(0)\rangle_{\rm tot}$
329: as
330: $$
331: \sum_{n_x}(c_{n_x}\alpha |g\rangle|n_x\rangle_g
332: +\sum_{n_x'}\eta_{n_x n_x'}c_{n_x'}\beta |e\rangle |n_x\rangle_p),
333: $$
334: we can analytically
335: evolve this state with $H_0$ to obtain
336: \begin{eqnarray}
337: \rho(\tau)&&=I_{g}|g\rangle\!\langle g|
338: +(1-I_g)|e\rangle\!\langle e|
339: +(I_{ge}|g\rangle\!\langle e|+h.c.),
340: \label{rho1}
341: \end{eqnarray}
342: with
343: \begin{eqnarray}
344: I_{g} &&=|\alpha|^2\cos^2\theta+|\beta|^2\sin^2\theta
345: +i(\alpha\beta^*\eta^*-c.c.)\cos\theta\sin\theta,\nonumber\\
346: I_{ge} &&=i(|\alpha|^2-|\beta|^2){1\over 2}\sin2\theta
347: \sum_{n_x,q_x}e^{-i(n_x-q_x)\omega_x^g\tau}
348: c_{n_x}c_{q_x}^*\eta_{q_xn_x}\nonumber\\
349: &&+\alpha\beta^*\cos^2\theta
350: \sum_{n_x,q_x}e^{-i(n_x-q_x)\omega_x^g\tau}
351: c_{n_x} \sum_{q_x'}\eta^*_{q_x q_x'}c_{q_x'}^*\eta_{q_xn_x} \nonumber\\
352: &&+\alpha^*\beta\sin^2\theta
353: \sum_{n_x,q_x}e^{-i(n_x-q_x)\omega_x^g\tau}
354: \sum_{n_x'}\eta_{n_x n_x'}c_{n_x'}c_{q_x}^*\eta_{q_xn_x},\nonumber
355: \end{eqnarray}
356: where we have defined the parameter
357: \begin{eqnarray}
358: \eta(k_L)=\sum_{n_x,n_x'}c_{n_x}^*\eta_{n_x n_x'}c_{n_x'}.
359: \end{eqnarray}
360: We see the evolution by Eq. (\ref{twou})
361: will in general not reproduce the intended
362: density matrix Eq. (\ref{rhot}) because of $\eta$.
363: $H_1$ term is the other reason for incomplete control
364: although its effects (when $\Omega_L\gg\omega_x^g$)
365: can be minimized by employing a fast pulse
366: with $\omega_x^g\tau\ll 1$.
367: Within such a limit, or when $\omega_x^g\tau=2\pi$,
368: we obtain
369: \begin{eqnarray}
370: I_{ge} &&=i(|\alpha|^2-|\beta|^2)\sin\theta\cos\theta\,\eta(k_L)\nonumber\\
371: &&+\alpha\beta^*\cos^2\theta+\alpha^*\beta\sin^2\theta\,\eta(2k_L).
372: \end{eqnarray}
373: The necessarily condition for attending
374: a perfect fidelity of the single bit
375: rotation is then $\eta(k_L)\equiv 1$, which can be
376: approximately satisfied in the LDL when $k_La_x^g\ll$
377: or in the $\Lambda$-type Raman systems with
378: co-propagating pump and Stokes fields.
379: As a second example, we consider the case of a
380: thermal motional state
381: \begin{eqnarray}
382: \rho_{\rm tot}(0) &&=|\psi(0)\rangle\!\langle\psi(0)|\otimes\rho_{\rm cm}(0),\nonumber\\
383: \rho_{\rm cm}(0) &&= \sum_{n_x}\rho^{\rm cm}_{n_x} |n_x\rangle\!\langle n_x|,\nonumber\\
384: \rho^{\rm cm}_{n_x} &&
385: =(1-e^{-\hbar\omega_x^g/k_BT})e^{-n_x\hbar\omega_x^g/k_BT}.
386: \end{eqnarray}
387: This is the limiting case of an ensemble average
388: of Eq. (\ref{psi1})
389: with
390: $c_{n_x}=\sqrt{\rho_{n_x}^{\rm cm}}\,e^{-i\phi_{n_x}}$
391: and
392: $\phi_{n_x}$ a uniform random number $\in[0,2\pi)$.
393: The dynamics due to $H_0$ can be evolved analytically
394: and the same density matrix Eq. (\ref{rho1}) is
395: obtained. After averaging over ${\{\phi_n\}}$, we obtain
396: \begin{eqnarray}
397: \langle \eta(k_L)\rangle_{\{\phi_n\}}
398: &&=\sum_{n_x} \rho_{n_x}^{\rm cm} \eta_{n_xn_x}(k_L)\nonumber\\
399: &&=\exp\left[-{1\over 2}(k_La_x)^2\coth\left({1\over 2}{\hbar\omega_x^g\over k_BT}\right)\right].
400: \end{eqnarray}
401: In the low temperature limit when
402: $k_BT< \hbar\omega_x^g$, $\eta$ becomes $1$ as
403: long as LDL $k_La_x^g\ll 1$ is satisfied.
404: At high temperatures when $k_BT\gg\hbar\omega_x^g$,
405: $(k_La_x^g)^2\ll {\hbar\omega_x^g/k_BT}$ needs to satisfied
406: for $\eta$ close to $1$.
407:
408: We now discuss the numerical solutions.
409: Expand the total wave-function as
410: \begin{eqnarray}
411: |\psi(t)\rangle_{\rm tot}=\sum_{n_x}[c_{n_x}^g(t)|g\rangle|n_x\rangle_g
412: +c_{n_x}^e(t)|e\rangle|n_x\rangle_p],
413: \end{eqnarray}
414: we have solved the Schr\"odinger equation
415: including both $H_0$ and $H_1$.
416: The transformation Eq. $(\ref{wp})$ greatly
417: reduces the motional Hilbert space dimension.
418: The perceived fidelity
419: for the electronic coded qubit Eq. (\ref{qb})
420: under transformation Eq. (\ref{rhot}) is
421: \begin{eqnarray}
422: {\cal F} &&=\rm Tr[\rho_T(\tau)\rho(\tau)].
423: \end{eqnarray}
424:
425: We take $\alpha=\beta=1/\sqrt{2}$ as an example to illustrate
426: our numerical results since similar/better fidelities
427: are obtained with other choices. In Figure \ref{fig2}
428: we compare fidelities under arbitrary $\theta(\tau)$
429: rotations for two different pure states.
430: Acceptable fidelities are obtained only for $k_La_x^g\le 0.3$ .
431: In general larger $\Omega_L/\omega_x^g$ ratios
432: also improved fidelity although it saturates
433: around $\Omega_L/\omega_x^g \sim 100$.
434: Noticeable improvements are also recorded for
435: narrower distributions in $|c_n|^2$, e.g. in Fig. \ref{fig2}
436: initial state with $c_{n_x}=\delta_{n_x 0}$
437: produced better fidelity. This is a direct reflection
438: of dephasing among different motional pair states
439: because of their different time scales from $H_0$
440: and $H_1$. The oscillatory behavior is due to dephasing
441: caused by the Rabi oscillation between motional paired
442: states. For comparison, we note that
443: meaningful single bit rotations need to achieve
444: a fidelity of $1/2$, the lower limit from a random (uncontrolled)
445: sampling of final states.
446:
447: Finally we compare with thermal states
448: for several different values of $k_BT/\hbar\omega_x^g$.
449: Surprisingly, we found the fidelity for an initial
450: motional thermal state is always higher than
451: its corresponding pure state. In the temperature
452: regime considered we found
453: acceptable results as long as LDL is maintained.
454: In Fig. \ref{fig3}, we have used
455: $k_La_x^g=0.1$ and $\Omega_L=100$ ($\omega_x^g$).
456:
457: In conclusion, we have performed detailed
458: theoretical studies of the decoherence of an electronically
459: coded atom/ion qubit due to ME. By introducing a
460: wave packet basis in the excited state, we were able
461: to perform considerably cleaner analysis to simplify
462: the ME. We found that a single parameter $\eta$ measures
463: the achievable fidelity of arbitrary single bit rotations.
464: We performed numerical calculations which demonstrates our
465: understanding and provided quantitative limits
466: for experiments: the LDL is always required to
467: maintain a high fidelity for arbitrary single bit rotations.
468: We also found that a pure motional state is not necessarily
469: preferred although a qubit with an initial ground
470: motional state does give rise to the
471: highest recorded fidelity. In actual experimental
472: implementations, a large $\Omega_L$ is also needed
473: to assure negligible motional dephasing during $\tau$.
474: One can always wait for a period $2\pi/\omega_x^g$ for subsequent
475: single bit operations since the motional wave function then
476: always rephases to its initial state.
477: This study will also shed light on devising schemes for overcoming
478: ME decoherence and error corrections in
479: trapped atomic/ionic qubits.
480:
481: We thank Dr. T. Uzer and Dr. M. S. Chapman for helpful discussions.
482: We also thank Dr. M\"{u}stecapl{\i}o\={g}lu
483: for supplying a Fortran
484: subroutine for evaluating $\eta_{n_xm_x}$. This work is supported by
485: the ONR grant No. 14-97-1-0633 and the ARO/NSA grant G-41-Z05.
486:
487: \begin{references}
488: %
489: \bibitem{Shor}P. W. Shor, in {\it Proc. 35th Annual Symposium on the
490: Foundations
491: of Computer Science}, edited by S. Goldwasser (IEEE Computer Society Press,
492: Los Alamitos, California, 1994);
493: A. Ekert and R. Jozsa, Rev. Mod. Phys. {\bf 68}, 733 (1996).
494:
495: \bibitem{nist}Q.A. Turchette {\it et al.},
496: Phys. Rev. Lett. {\bf 81}, 3631 (1998); {\it ibid}, {\bf 75}, 4714 (1995).
497:
498: \bibitem{tele}D. Boschi {\it et al.}, Phys. Rev. Lett. {\bf 80}, 1121 (1998);
499: Dik Bouwmeester {\it et al.}, {\it Nature} {\bf 390},
500: 575 (1997); A. Furusawa {\it et al.},
501: {\it Science} {\bf 282}, 706 (1998).
502:
503: \bibitem{anton}Dik Bouwmeester {\it et al.},
504: Phys. Rev. Lett. {\bf 82}, 1345 (1999).
505:
506: \bibitem{Cirac}J. I. Cirac and P. Zoller,
507: Phys. Rev. Lett. {\bf 74}, 4091 (1995).
508:
509: \bibitem{Cirac2}T. Pellizzari {\it et al.},
510: Phys. Rev. Lett. {\bf 75}, 3788(1995).
511:
512: \bibitem{side}J. Javanainen {\it et al.},
513: J. Opt. Soc. Am. B {\bf 1}, 111 (1984).
514:
515: \bibitem{monroe}C. Monroe {\it et al.},
516: Phys. Rev. Lett. {\bf 75}, 4011 (1995);
517: M. Morinaga {\it et al.}, {\it ibid}, {\bf 83}, 4037 (1999);
518: Ch. Roos {\it et al.}, {\it ibid}, {\bf 83}, 4713 (1999);
519: E. Peik {\it et al.}, Phys. Rev. A {\bf 60}, 439 (1999).
520:
521: \bibitem{bf}R. Bonifacio {\it et al.},
522: (quant-ph/9911100 and /9906115).
523:
524: \bibitem{Cirac3}J. F. Poyatos {\it et al.},
525: Phys. Rev. Lett. {\bf 81}, 1322 (1998);
526: A. Sorensen and K. Molmer, {\it ibid}, {\bf 82}, 1971 (1999);
527: K. Molmer and A. Sorensen, {\it ibid}, {\bf 82}, 1835 (1999).
528:
529: \bibitem{Milburn}S. Schneider and G. J. Milburn,
530: Phys. Rev. A {\bf 59}, 3766 (1999).
531:
532: \bibitem{jas}M. Cirone {\it et al.},
533: Phys. Rev. A {\bf 57}, 1202 (1998).
534:
535: \bibitem{zeng}H. P. Zeng {\it et al.},
536: Phys. Rev. A {\bf 59}, 2174 (1999).
537:
538: \bibitem{vogel}R. L. de Matos Filho and W. Vogel,
539: Phys. Rev. Lett. {\bf 76}, 608 (1996).
540:
541: \bibitem{walls}A. C. Doherty, A. S. Parkins, S. M. Tan,
542: and D. F. Walls, J. Opt. B {\bf 1}, 475 (1999).
543:
544: \bibitem{gorigi}G. M. Gorigi {\it et al.},
545: Phys. Rev. A {\bf 57}, 2909 (1998).
546:
547: \bibitem{you}L. You {\it et al.}, Phys. Rev. A {\bf %
548: 51}, 4712 (1995).
549:
550: \bibitem{jeff}J. Ye {\it et al.},
551: Phys. Rev. Lett. {\bf 83}, 4987 (1999).
552:
553: \bibitem{barenco}A. Barenco {\it et al.}, Phys. Rev. A {\bf 52}, 3457 (1995).
554:
555: \end{references}
556: %
557: %
558: %
559: %
560: %% figure 1
561: \newpage
562: \begin{figure}[t]
563: \centerline{\epsfig{file=fig1.eps,width=2.5in}\\[12pt]}
564: \caption{The paired ladders of a trapped two state atom/ion.
565: Solid curve arrow heads denote Rabi oscillations between
566: paired states, while the dotted curve arrow heads denote
567: nearest neighbor motional coupling.}
568: \label{fig1}
569: \end{figure}
570: %
571:
572: \begin{figure}[t]
573: \centerline{\epsfig{file=fig2.eps,width=2.5in}\\[12pt]}
574: \caption{The fidelity for single bit rotations with
575: $k_La_x^g=0.1$ (dashed lines), $0.3$ (solid lines),
576: and $1.0$ (dot dashed lines). The higher and lower fidelity
577: sets are for initial motional state $c_{n_x}=\delta_{n_x0}$ and
578: $c_{n_x}=(2\,\delta_{n_x0}
579: +\sqrt{2}\,\delta_{n_x1}+\delta_{n_x2})/\sqrt{7}$ respectively.
580: $\Omega_L=100$ $(\omega_x^g)$.}
581: \label{fig2}
582: \end{figure}
583:
584: \begin{figure}[t]
585: \centerline{\epsfig{file=fig3.eps,width=2.5in}\\[12pt]}
586: \caption{The fidelity for single bit rotations
587: with $k_BT/\hbar\omega_x^g=1$ (dashed lines), $3$ (solid lines),
588: and $10$ (dot dashed lines). The higher and lower fidelity
589: sets are for initial motioal thermal state $\rho_{n_x}^{\rm cm}$
590: and pure state $c_{n_x}=\sqrt{\rho_{n_x}^{\rm cm}}$ respectively.}
591: \label{fig3}
592: \end{figure}
593:
594: \end{document}
595: \end
596: