1: \newcommand{\ds}{\displaystyle}
2: %\documentstyle[aps,epsf,prl,twocolumn]{revtex}
3: %\documentstyle[preprint,aps,epsf]{revtex}
4: \documentstyle[12pt,epsf]{article}
5: \begin{document}
6:
7: \begin{center}
8: {\bf Transient Dynamics in Magnetic Force Microscopy for a Single-Spin
9: Measurement}\\ \ \\
10:
11: {G.P. Berman,$\!^1$ F. Borgonovi,$\!^{2}$ G. V.
12: L\'opez,$\!^3$ V.I. Tsifrinovich~$\!^{4}$}\\ \ \\
13: \end{center}
14:
15: \begin{center}
16: {$^1$Theoretical Division and CNLS,
17: Los Alamos National Laboratory, Los Alamos, NM 87545}\\
18:
19: {$^2$Dipartimento di Matematica e Fisica, Universit\`a Cattolica,
20: via Musei 41 , 25121 Brescia, Italy, and I.N.F.M., Unit\`a
21: di Brescia and I.N.F.N., sezione di Pavia }\\
22:
23: {$^3$
24: Departamento de F\'{\i}sica de la Universidad de Guadalajara
25: S.R. 500, 44420
26: Guadalajara, Jalisco, M\'exico}\\
27:
28: {$^4$
29: IDS Department, Polytechnic University, Six Metrotech Center, Brooklyn,
30: New York 11201}
31: \end{center}
32:
33:
34: %\maketitle
35: \begin{abstract}
36: \noindent
37: We analyze a single-spin measurement using a transient process in magnetic
38: force microscopy (MFM) which could increase
39: the maximum operating temperature by a factor of $Q$ (the quality factor of
40: the
41: cantilever) in comparison with the static Stern-Gerlach effect. We obtain an
42: exact solution of the master
43: equation, which confirms this result. We also discuss the conditions
44: required to create a macroscopic Schr\"odinger cat state in the cantilever.
45: \end{abstract}
46:
47: \section{\bf Introduction}
48: The routine magnetic force microscopy (MFM) seems to be of a little
49: use for a single spin
50: detection in solids. Even for unrealistically small distances between
51: the ferromagnetic
52: particle on the cantilever tip and the spin, the maximum temperature for a
53: single-spin
54: measurement comes to a millikelvin region [1]. In this paper, we show that
55: using a
56: transient process, one can increase the maximum temperature of a single-spin
57: measurement
58: by a factor of $Q$ (the quality factor of the cantilever). Alternatively, one
59: can increase the distance between the ferromagnetic
60: particle and the cantilever tip if one is willing to work at millikelvin
61: temperatures.\\ \ \\
62: \noindent
63: In section 2 we explain the basic idea of our work. In sections 3-5 we obtain
64: an exact solution
65: of the master equations, which confirms our idea. In section 6 we derive the
66: conditions
67: for creating a macroscopic Schr\"odinger cat state (MSCS) in MFM.
68: In Conclusion
69: we summarize our results.\\ \ \\
70: %
71: \section{\bf Transient Process in MFM}
72: %
73: In this section we describe the basic idea of our work. Suppose that a
74: ferromagnetic particle on the
75: cantilever tip interacts with a single spin on the solid surface.
76: (See Fig. 1.) The equilibrium position
77: of the cantilever tip depends on the spin direction. The distance
78: between two possible
79: equilibrium positions, corresponding to two spin stationary states,
80: is given by $2F/k_c$,
81: where $k_c$ is the cantilever spring constant, $F$ is the magneto-static
82: force between the
83: ferromagnetic particle and the spin. In order to measure the state of the
84: spin, this distance
85: must be greater than twice the uncertainty due to the
86: thermodynamical noise of the
87: cantilever position. This uncertainty can be estimated
88: as $(k_{_B}T/k_c)^{1/2}$, where $k_{B}$
89: is the Boltzmann's constant and $T$ is the temperature.
90: Thus, the condition for a single spin measurement
91: (the static Stern-Gerlach effect) is
92: \cite{g}
93: \begin{equation}
94: T<T_{max}=F^2/k_{_B}k_c.
95: \end{equation}
96: Now, assume that we quickly change the stationary state of the spin and
97: consider the transient
98: cantilever vibrations after this change. The amplitude of the cantilever
99: vibrations at
100: time $t\ll Q/\omega_c$ ($\omega_c$ is the cantilever frequency
101: and $Q/\omega_c$ is
102: the time constant of the cantilever) is $4F/k_c$. In addition, assume
103: that we detect the
104: position and momentum of a point on the cantilever tip with an accuracy
105: that satisfies the quantum limit
106: $(\delta P_{_Z})(\delta Z)\approx\hbar/2$ (the cantilever oscillates
107: along the z-axis.) To
108: find out the state of the spin
109: we are going to compare the observed trajectory of the cantilever
110: tip with the theoretical
111: prediction. The theory predicts the cantilever trajectory
112: within an uncertainty due to the thermal noise. However, the thermal
113: uncertainty of the cantilever
114: position at time $t\ll Q/\omega_c$ increases as
115: $[t(\omega_c/Q)(k_{_B}T/k_c)]^{1/2}$.
116: We can obtain this expression assuming an initial ($t\ll Q/\omega_c$)
117: thermal diffusion
118: with two common properties: 1) The dispersion $\overline{(\delta Z)^2}$
119: is proportional to time, $t$, and 2) The uncertainty of the cantilever
120: position equals its thermodynamical
121: value if we formally put $t=Q/\omega_c$. \\ \\
122: At the time $t=\pi/\omega_c$ (half of the first
123: period), the distance between two possible cantilever positions takes
124: its maximum possible
125: value $4F/k_c$. At the same time, the thermal uncertainty of the position
126: predicted by
127: the theory is still much smaller than its thermodynamical value. Now, the
128: condition for
129: a single spin measurement is
130: \begin{equation}
131: T<T_{max}={ 4QF^2\over\pi k_{_B}k_c}.
132: \end{equation}
133: One can see that the maximum temperature for a single spin measurement
134: increases by a factor
135: of $Q$, compared with the static Stern-Gerlach effect.\\ \ \\
136: In the next three sections, we will confirm this estimation by direct
137: solution of the master
138: equation.
139: %
140: \section{\bf The Hamiltonian and the Master Equation}
141: %
142: We assume that the cantilever tip with an attached ferromagnetic particle
143: can oscillate along the z-axis.
144: (See Fig. 1.) A single paramagnetic atom with spin $1/2$ is placed near the
145: cantilever tip.
146: The dimensionless Hamiltonian of the cantilever tip interacting with a
147: single spin is
148:
149: \begin{equation}
150: \hat H={1\over 2}(\hat p_z^2+ \hat z^2)-2\eta \hat z \hat S_z.
151: \label{eqh}
152: \end{equation}
153: We introduced the following notation (below we omit hats for operators):
154:
155: \begin{equation}
156: z=Z/Z_q,\quad p_z=P_z/P_q,\quad
157: \eta=g\mu_{_B}\left|{\partial B_z\over\partial z}\right|/2F_q\ ,
158: \end{equation}
159: where $Z_q$, $P_q$ and $F_q$ are the ``quants" of the
160: coordinate, momentum and force
161: acting on the cantilever,
162:
163: \begin{equation}
164: Z_q=\left({\hbar\omega_c\over k_c}\right)^{1/2},\quad P_q=\hbar/Z_q,\quad
165: F_q=k_cZ_q .
166: \end{equation}
167: The variables $Z$ and $P_z$ are the ``dimensional" coordinate and
168: momentum of the
169: cantilever tip, $k_c$ is the
170: cantilever ``spring constant", $\omega_c$ is its frequency,
171: $g$ is the ``g-factor" of the
172: spin (below we use $g=2$), $\partial B_z/\partial z$ is the
173: magnetic field gradient
174: produced by the ferromagnetic particle at the spin location
175: when the cantilever is in the equilibrium position with no spin
176: ($z=0$). Note that the
177: cantilever interacts with the z-component of the spin, which is an
178: integral of motion
179: in our system. In the Hamiltonian (3) we omitted the term
180: ($g\mu_{_B}B_0/\hbar\omega_c)\hat S_z$,
181: where $B_0$ is the magnetic field on the spin when the cantilever
182: is at the origin ($z=0$).
183: This term may be eliminated ``physically'' (by application of a
184: uniform external field of
185: magnitude $B_0$ in the negative $z$-direction) or ``mathematically''
186: (by transferring to the
187: system of coordinates rotating with the frequency $g\mu_{_B}B_0/\hbar$).
188: \\ \ \\
189: The master equation describes the evolution of the density matrix of the
190: system
191: interacting with the environment (see for example [2-5]).
192: We are taking into account the interaction of the cantilever with its
193: environment, and ignore the
194: direct interaction between the spin and the environment, assuming that
195: the spin relaxation and decoherence times are large enough. The effect of the
196: environment depends on its ``spectral density", i.e. the density of
197: environmental
198: oscillators at a given frequency. Probably, the simplest model of the
199: environment
200: is the ``ohmic" model, where the spectral density is proportional to the
201: frequency
202: $\omega$ for $\omega<\Omega$, where $\Omega$ is the cutoff frequency
203: for the environment.
204: A master equation for the non-ohmic environment has been derived in [5].
205: For the ohmic model,
206: the simplest master equation has been obtained in [2]. This equation is
207: valid in the
208: ``high temperature limit" $k_{_B}T>>\hbar\Omega$. The master equation
209: derived in [3] is valid
210: for arbitrary temperature. As pointed out in [5], both equations [2]
211: and [3] fail at times shorter than or close to $\hbar/k_{_B}T$.\\ \ \\
212: We are going to consider the ``gedanken experiment" discussed in \cite{g}.
213: Suppose that initially
214: ($t=0$) the spin is in a superposition of the two states
215: with the $z$-projection of the spin $S_z=1/2$ and
216: $S_z=-1/2$. These two states of the spin correspond to two different
217: equilibrium positions
218: of the cantilever tip. Thus, the cantilever (without decoherence) would
219: transform
220: into a MSCS: two simultaneous equilibrium positions. Certainly, decoherence
221: will destroy this state. The master equation describes both the appearance
222: of the MSCS and its destruction due to decoherence.\\ \ \\
223: Following \cite{g}, we consider the ultra-thin cantilever reported on [6].
224: It has the spring
225: constant $k_c=6.5\times 10^{-6}~N/m$, the frequency, $\omega_c/2\pi=1.7~kHz$,
226: and the
227: quality factor, $Q=6700$. The ferromagnetic particle on the cantilever tip is
228: taken
229: to be a sphere of radius $R=15~nm$ at a distance $5~nm$ from the paramagnetic
230: atom. (Below we consider conditions for increasing the distance between the
231: cantilever and spin.)
232: For this case, the static displacement of the cantilever tip due to its
233: interaction
234: with the single spin exceeds the thermal vibrations of the cantilever for
235: temperatures
236:
237: \begin{equation}
238: T<T_{max}={(\mu_{_B}\partial B_z/\partial z)^2\over k_{_B}k_c}\approx 1.7~mK\ .
239: \end{equation}
240: In our gedanken experiment for the temperature $T>>\hbar\omega_c/k_{_B}\approx
241: 8\times 10^{-8}~K$, we can use the simplest high temperature limit in
242: the ``ohmic
243: model". \\ \ \\
244: The master equation in the high temperature limit can be written in the
245: form [2]
246:
247: \begin{equation}
248: \begin{array}{lll}
249: {\ds\partial\rho_{s,s'}\over\ds\partial \tau} &=\Biggl[{\ds i\over 2}
250: \left(\partial_{zz}-\partial_{z'z'}\right)-{\ds i\over\ds 2}(z^2-z'^2)
251: -{\ds 1\over\ds 2} \beta (z-z')(\partial_z-\partial_{z'})
252: -D\beta(z-z')^2\\
253: & -2 i\eta(z's'-zs)\Biggr]\rho_{s,s'}.
254: \label{eqr}
255: \end{array}
256: \end{equation}
257: Here, $s$ and $s'$ take values $\pm 1/2$ (we use $s$ instead of $S_z$),
258: $\tau=\omega_ct$,
259: $\beta=1/Q$ and $D=k_{_B}T/\hbar\omega_c$.
260: Using new coordinates
261: \begin{equation}
262: r =z-z',\quad\quad
263: R = \frac{1}{2} (z+z').
264: \label{nco}
265: \end{equation}
266: Eq. (\ref{eqr}) can be written as:
267: \begin{equation}
268: \begin{array}{lll}
269: &{\ds\partial\rho_{s,s'}(R,r,\tau)\over\ds\partial \tau}=\\ \\
270: &\biggl\{i\partial_{Rr}-i Rr-
271: \beta r\partial_r-D\beta r^2
272: -i\eta\biggl[(2R-r)s'-(2R+r)s \biggr] \biggl\}~\rho_{s,s'}(R,r,\tau).
273: \end{array}
274: \label{eqr1}
275: \end{equation}
276: Performing a Fourier transformation of this equation with respect to
277: the variable ``R'', one obtains, after re-arrangements,
278: \begin{equation}
279: {\partial\hat\rho_{s,s'}\over\partial \tau}+(\beta r-k)
280: {\partial\hat{\rho}_{s,s'}\over
281: \partial r}+
282: \biggl[
283: r+2\eta(s'-s)\biggr]{\partial\hat{\rho}_{s,s'}\over\partial k}=
284: \biggl[-D\beta r^2+i\eta r(s'+s)\biggr]
285: \hat\rho_{s,s'},
286: \label{eqr2}
287: \end{equation}
288: where
289:
290: $$
291: \hat\rho_{s,s'}(k,r,\tau)=\int_{-\infty}^{+\infty}e^{ikR}
292: \rho_{s,s'}(R,r,\tau)~dR
293: .
294: $$
295: We can study separately the spin diagonal case ($s=s'$) and the
296: off--diagonal case
297: ($s\ne s'$). For $s'=s$ (up-up or down-down spins), we have
298: the following equation:
299: \begin{equation}
300: {\partial\hat\rho_{s,s}\over\partial \tau}+(\beta
301: r-k){\partial\hat\rho_{s,s}\over\partial r}+
302: r{\partial\hat\rho_{s,s}\over\partial k}=\biggl(-D\beta r^2+ 2 i\eta r
303: s\biggr)~\hat\rho_{s,s},
304: \label{eqrd}
305: \end{equation}
306: and for $ s' \ne s$ (up-down or down-up spins):
307:
308: \begin{equation}
309: {\partial\hat\rho_{s,-s}\over\partial \tau}+(\beta r-k)
310: {\partial\hat\rho_{s,-s}\over\partial r}+
311: ( r+4\eta s){\partial\hat\rho_{s, -s}\over\partial k}=
312: -D\beta r^2 \ \hat\rho_{s,-s}.
313: \label{eqrnd}
314: \end{equation}
315: We will derive the exact solution of the master equation (\ref{eqr})
316: for the case when
317: the spin is ``prepared" initially in the superposition of two states
318: with $s=1/2$ and
319: $s=-1/2$, while the cantilever tip is in the quasiclassical coherent state
320: \begin{equation}
321: \psi(z,s,0)= \frac{1}{(\pi)^{1/4}}
322: \exp [ ip_0z - (z-z_0)^2/2 ] \otimes\left(\begin{array}{c}a\\ \\
323: b\end{array}\right),
324: \label{psi}
325: \end{equation}
326: where the amplitudes $a$ and $b$ correspond to the values of
327: $s=1/2$ and $s=-1/2$ respectively. The corresponding density
328: matrix can be written as
329:
330: \begin{equation}
331: \rho_{ss'}(z,z',0)=\psi(z,s,0)\otimes\psi^{\dag}(z',s',0).
332: \end{equation}
333: Note that we consider an ensemble of spin-cantilever systems with
334: the same initial
335: state. This implies that the experimenter can detect the position
336: and momentum of
337: a point on the cantilever tip with quantum limit accuracy
338: $\overline{(\delta p_z)^2}~\overline{(\delta z)^2}=1/4$.
339: (In our gedanken experiment, this corresponds to an uncertainty
340: of $300~fm$ for position
341: and $300~nm/s$ for velocity.) Based
342: on the master equation, we can predict the average position of the
343: cantilever tip
344: for its given initial state, depending on the spin state. If
345: the double uncertainty of the position is smaller than the separation
346: between two
347: possible average positions, the cantilever tip will measure the state of
348: the spin.\\ \ \\
349: After Fourier transformation, the ``cantilever part" of the density matrix
350: is represented
351: by
352:
353: \begin{equation}
354: \hat\rho_{s,s'}(k,r,0)\propto\exp{\left[\displaystyle
355: i p_0 r+ikz_0-r^2/4-k^2/4\right]}.
356: \label{rof}
357: \end{equation}
358:
359: \section{\bf Solution for spin diagonal matrix elements}
360:
361: The equations for the characteristics of Eq. (\ref{eqrd}) are
362: \begin{equation}
363: d\tau ={dr\over\beta r-k}={dk\over r}=
364: {d\hat\rho_{s,s'}\over\biggl(-D\beta r^2+2 i \eta
365: s r\biggr) \hat\rho_{ss}},
366: \label{eqcd}
367: \end{equation}
368: or, explicitly
369: \begin{equation}
370: \begin{array} {lll}
371: {\ds dr\over\ds d\tau} &=\beta r -k, \\ \ \\
372: {\ds dk\over\ds d\tau}&= r,\\ \ \\
373: {\ds d\hat\rho_{s,s'}\over\ds d\tau}&=\biggl(-D\beta r^2+ 2 i \eta
374: s r \biggr)~\hat\rho_{s,s'}.
375: \end{array}
376: \label{eqc1}
377: \end{equation}
378: From the first two equations in (\ref{eqc1}), one obtains
379: $$
380: {d^2k\over d\tau^2}-\beta {dk\over d\tau}+ k=0,
381: $$
382: which has the following general solution
383: \begin{equation}
384: k=e^{\beta \tau/2}\biggl(c_1\cos{\theta\tau}+c_2\sin{\theta\tau}\biggr),
385: \label{eqk}
386: \end{equation}
387: where
388: $
389: \theta=\sqrt{1 -{\beta^2\over 4}}
390: $. (Here we are considering the case $\beta < 2$ so $\theta$
391: is a real number. The case $\beta > 2$ can also be solved analytically.)
392: Using the second equation in (\ref{eqc1}) one obtains:
393: \begin{equation}
394: r=e^{\beta \tau/2}\left[\left({\beta\over 2}\cos{\theta\tau}-\theta\sin{\theta
395: \tau} \right)c_1+
396: \left({\beta\over 2}\sin{\theta\tau}+\theta\cos{\theta\tau}\right)c_2\right].
397: \label{eqrr}
398: \end{equation}
399: Inverting Eqs. (\ref{eqk}) and (\ref{eqrr}) as a function of $c_1$ and
400: $c_2$ one obtain the characteristic curves:
401: \begin{equation}
402: c_1=e^{-\beta\tau/2}( q_1k+ q_2r),
403: \label{ch1}
404: \end{equation}
405: and
406: \begin{equation}
407: c_2=e^{-\beta\tau/2}( p_1k+ p_2r),
408: \label{ch2}
409: \end{equation}
410: where the time dependent
411: constants $q_1$, $q_2$, $p_1$ and $p_2$ have been defined as
412: \begin{equation}
413: \begin{array}{lll}
414: q_1&={\ds 1\over\ds\theta}\biggl({\ds\beta\over\ds 2}\sin{\theta\tau}+
415: \theta\cos{\theta
416: \tau}\biggr),\\ \\
417: q_2&=-{\ds 1\over\ds\theta} \sin \theta\tau,\\ \\
418: p_1&={\ds 1\over\ds \theta}\biggl(-{\ds\beta\over\ds 2}\cos{\theta\tau}+
419: \theta\sin{\theta
420: \tau}\biggr),\\ \\
421: p_2&={\ds 1\over\ds\theta} \cos \theta\tau.\\
422: \label{qps}
423: \end{array}
424: \end{equation}
425: Substituting (\ref{eqrr}) into the third equation of (\ref{eqc1})
426: and integrating in time, one obtains:
427:
428: \begin{equation}
429: \hat\rho_{s,s}(k,r,\tau) \propto Q(c_1,c_2) \exp{\left[ i2\eta
430: s(c_1g_1+c_2g_2)-D\beta
431: (c_1^2f_1+2c_1c_2f_3+c_2^2f_2)\right]},
432: \label{ss1}
433: \end{equation}
434: where the functions $f_i's$ and $ g_i's$ are defined as
435: \begin{equation}
436: \begin{array}{lll}
437: \displaystyle f_{_1}(\tau) &={e^{\beta\tau}\over 8}\biggl[\left(\beta+
438: {4\theta^2\over\beta}\right)+
439: \beta\cos{2\theta\tau}-2\theta\sin{2\theta\tau}\biggr],\\ \\
440: \displaystyle f_{_2}(\tau) &={e^{\beta\tau}\over 8}\biggl[\left(\beta+
441: {4\theta^2\over\beta}\right)-
442: \beta\cos{2\theta\tau}+2\theta\sin{2\theta\tau}\biggr],\\ \\
443: \displaystyle f_{_3}(\tau) &={e^{\beta\tau}\over 8}\biggl[2\theta
444: \cos{2\theta\tau}+\beta\sin{2\theta\tau}
445: \biggr],\\ \\
446: g_{_1}(\tau) &=e^{\beta\tau/2}\cos{\theta\tau},\\ \\
447: g_{_2}(\tau) &=e^{\beta\tau/2}\sin{\theta\tau}.
448: \end{array}
449: \label{feg}
450: \end{equation}
451: The arbitrary function $A$ which depends on the characteristics is
452: determined by the initial density matrix
453: $\hat\rho_{s,s}(k(0),r(0),0)$,
454: \begin{equation}
455: \begin{array}{lll}
456: A(c_1,c_2)&=\hat\rho_{s,s}\biggl(c_1,\frac{1}{2}\beta c_1+\theta c_2,0\biggr)
457: ~\exp{\left[\displaystyle -2 i \eta s(c_1g_{10}+c_2g_{20})\right]}\\
458: &\quad\times\exp{\left[\displaystyle
459: D\beta(c_1^2f_{10}+2c_1c_2f_{30}+c_2^2f_{20})\right]},
460: \end{array}
461: \label{qqe}
462: \end{equation}
463: where $f_{i0}=f_i(0)$ and $g_{i0}=g_i(0)$.
464: From the initial density matrix (Eq. (\ref{rof})), we obtain
465:
466: \begin{equation}
467: \begin{array}{lll}
468: \hat\rho_{s,s'}(k,r,0)& \propto
469: \exp{\left\{ i\left[ \left( \frac{1}{2} p_0\beta +z_0 \right)c_1+
470: p_0\theta c_2\right]\right\}}\\ \\
471: &\times\exp{\left\{-\left[
472: \left({\beta^2\over 16}+{1\over 4}\right)c_1^2+
473: {\beta\theta \over 4} c_1 c_2 +
474: {\theta^2\over 4}c_2^2\right]\right\}}.
475: \end{array}
476: \label{rof1}
477: \end{equation}
478: Substituting (\ref{qqe}) and (\ref{rof1}) into (\ref{ss1}) one obtains:
479:
480: \begin{equation}
481: \begin{array}{lll}
482: \hat\rho_{s,s}(k,r,\tau)&\propto\exp{\left\{\displaystyle i\left[
483: \left( \frac{1}{2} p_0\beta +z_0+2\eta sG_1\right)c_1+
484: \left(p_0\theta+2\eta s G_2\right)c_2\right]\right\}}\\ \\
485: &\quad\times\exp{\left\{\displaystyle-\left[
486: \left({\beta^2\over 16}+{1\over 4}\right)c_1^2+
487: {\beta\theta\over 4} c_1c_2+{\theta^2\over 4}c_2^2\right]\right\}}\\ \\
488: &\quad\times\exp{\left\{-D\beta( F_1 c_1^2+2 c_1 c_2 F_3+F_2 c_2^2)\right\}},
489: \end{array}
490: \label{e37}
491: \end{equation}
492: where $F_i$ and $G_i$ are defined as
493: $$F_i(\tau)=f_i(\tau)-f_{i0},\quad\quad\quad\quad G_i(\tau)=g_i(\tau)-g_{i0}.
494: $$
495: Substituting in (\ref{e37}) the values of characteristics as a function of
496: $k$ and $r$
497: (Eqs.(\ref{ch1}) and (\ref{ch2})), one obtains:
498:
499: \begin{equation}
500: \hat\rho_{ss}(k,r,\tau) \propto\exp{\left[\displaystyle
501: -r^2 C_1 +irC_2 +
502: \left( iB_2 -r B_1 \right) k - \sigma_*^2 k^2 \right]},
503: \label{e38}
504: \end{equation}
505: where
506: \begin{equation}
507: \sigma_{*}^2 =e^{-\beta t} \left[ \left(
508: {\beta^2\over 16}+{1\over 4}\right) q_1^2
509: +{\beta\theta \over 4} q_1 p_1
510: +{\theta^2\over 4} p_1^2
511: \quad+D\beta( F_1q_1^2+2q_1p_1F_3+F_2p_1^2) \right],
512: \label{esig}
513: \end{equation}
514:
515: \begin{equation}
516: \begin{array}{lll}
517: B_1&=e^{-\beta t}\biggl\{
518: \left({\beta^2\over 16}+{1\over 4}\right)2q_1q_2+
519: {\beta\theta\over 4}(q_1p_2+q_2p_1) +
520: {\theta^2\over 4}~2p_1p_2\\
521: &\quad+2D\beta [
522: F_1q_1q_2+(q_1p_2+q_2p_1)F_3+F_2p_1p_2]\biggr\},
523: \label{eb1}
524: \end{array}
525: \end{equation}
526:
527: \begin{equation}
528: B_2(s)=e^{-\beta t/2}\biggl[
529: \left(\frac{1}{2}p_0 \beta+z_0+2\eta s G_1\right)q_1+
530: \left(p_0 \theta+2\eta s G_2\right)p_1\biggr],
531: \label{eb2}
532: \end{equation}
533:
534: \begin{equation}
535: \begin{array}{lll}
536: C_1&= e^{-\beta t}\biggl[
537: \left({\beta^2\over 16}+{1\over 4}\right)q_2^2+
538: {\beta\theta \over\ 4 } q_2 p_2 +
539: {\theta^2\over4}p_2^2\\
540: &\quad +D\beta (F_1q_2^2+2q_2p_2F_3+F_2p_2^2)\biggr],\\
541: \label{ec1}
542: \end{array}
543: \end{equation}
544:
545: \begin{equation}
546: C_2(s)= e^{-\beta t/2}\biggl[
547: \left(\frac{1}{2}p_0 \beta+z_0+2\eta s G_1\right)q_2+
548: \left(p_0 \theta+2\eta s G_2\right)p_2\biggr].
549: \label{ec2}
550: \end{equation}
551: In Eqs. (31) and (33) the explicit dependence on $s$ is presented.
552: Performing the inverse Fourier transform
553: one obtains
554: \begin{equation}
555: \begin{array}{lll}
556: \rho_{1/2,1/2}(R,r,\tau)&={\ds |a|^2 \over\ds
557: \sqrt{\pi}~\sigma_{*}}~
558: \exp{\left[\displaystyle - r^2 C_1+ ir C_2(1/2)\right]}\\
559: &\quad\quad\quad\quad\times\exp{\left[ ( -rB_1+iB_2(1/2)-iR)^2
560: /4\sigma^2_{*}\right]},\\ \ \\
561:
562: \rho_{-1/2,-1/2}(R,r,\tau)&={\ds |b|^2 \over\ds
563: \sqrt{\pi}~\sigma_{*}}~
564: \exp{\left[\displaystyle- r^2 C_1+ ir C_2(-1/2)\right]}\\
565: &\quad\quad\quad\quad\times\exp{\left[ ( -rB_1+iB_2(-1/2)-iR)^2/4
566: \sigma^2_{*}\right]}.\\
567: \label{efd}
568: \end{array}
569: \end{equation}
570: Eqs. (\ref{efd}) represent two squeezed Gaussians with modulus
571:
572: \begin{equation}
573: \begin{array}{lll}
574: |\rho_{1/2,1/2}(R,r,\tau)|&={\ds |a|^2 \over\ds
575: \sqrt{\pi}~\sigma_{*}}~
576: \exp{\left[\displaystyle- r^2 (C_1- B_1^2/4\sigma^2_{*})\right]}\\
577: &\quad\quad\quad\quad\times\exp{\left[ -(B_2(1/2)-R)^2
578: /4\sigma^2_{*}\right]},\\ \ \\
579:
580: |\rho_{-1/2,-1/2}(R,r,\tau)|&={\ds |b|^2 \over\ds
581: \sqrt{\pi}~\sigma_{*}}~
582: \exp{\left[\displaystyle- r^2 (C_1- B_1^2/4\sigma^2_{*})\right]}\\
583: &\quad\quad\quad\quad\times\exp{\left[ -(B_2(-1/2)-R)^2
584: /4\sigma^2_{*}\right]}.
585: \label{efds}
586: \end{array}
587: \end{equation}
588: Fig. 2 shows schematically two peaks (seen from the top as ellipses)
589: corresponding to
590: the two matrix elements $|\rho_{1/2,1/2}|$ and $|\rho_{-1/2,-1/2}|$. We denote
591: the centers of the ellipses, which lie on the diagonal $z=z'$, by $M_{++}$
592: and $M_{--}$,
593: the semi-major axis by $\sigma_d$, the semi-minor axis by $\sigma_d'$ and the
594: distance between the centers by $\Delta_d$. The
595: $\rho^{max}_{1/2,1/2}$ is located at $M_{++}=(r=0, R=B_2(1/2))$ or
596: $z=z'=B_2(1/2)$, while the $\rho^{max}_{-1/2,-1/2}$
597: is at $M_{--}=(r=0, R=B_2(-1/2))$ or $(z=z'=B_2(-1/2))$. The distance
598: $ \Delta_d$
599: is given by
600: \begin{equation}
601: \Delta_d = B_2(1/2)-B_2(-1/2).
602: \label{zmax}
603: \end{equation}
604: From Eqs. (\ref{efds}), we obtain $\sigma_d=\sqrt{2}~\sigma_*$, and
605:
606: \begin{equation}
607: 2{{\sigma'}_d}^2 = \displaystyle \frac{4\sigma_*^2}{4\sigma_*^2 C_1 - B_1^2}.
608: \label{scr}
609: \end{equation}
610: For a single spin measurement, the two peaks corresponding to
611: $\rho^{max}_{1/2,1/2}$ and
612: $\rho^{max}_{-1/2,-1/2}$ must be well separated. It follows that the condition
613: $\Delta_d>2\sigma_d$ must be satisfied.\\ \\
614: First, we consider the case $\beta\tau\gg 1$ or $\tau\gg Q/\omega_c$, where
615: $Q/\omega_c$ is
616: the time constant for the cantilever. In this case, we obtain two equilibrium
617: positions for
618: the cantilever, when the transient process is over. We have $\Delta_d=2\eta$
619: and $\sigma_d=\sqrt{D}$.
620: The value $\sigma_d=\sqrt{D}$ is the thermodynamical uncertainty in the
621: cantilever position caused
622: by the thermal noise.
623: The two equilibrium positions can be distinguished if $\eta>\sqrt{D}$ or
624: $T<F^2/k_{_B}k_c$, where
625: $F=\mu_{_B}|\partial B_z/\partial z|$ is the magneto-static force between
626: the ferromagnetic particle
627: and the paramagnetic atom. The last expression exactly coincides with
628: formula (1).\\ \ \\
629: Next, we consider the initial transient process after the instant ($t=0$)
630: at which the paramagnetic
631: spin has been transferred into the superpositional state.
632: For $\beta\tau\ll 1$, we have
633:
634: \begin{equation}
635: \Delta_d=4\eta\sin^2\frac{\tau}{2},\quad \sigma_d=\left[1/2+D\beta\tau-
636: D\beta\cos\tau\sin^3\tau\right]^{1/2}.
637: \end{equation}
638: This expression for $\Delta_d$ describes the oscillating distance between
639: the two peaks. It corresponds to
640: initial vibrations of two classical oscillators near their equilibrium
641: positions $z=\eta$ and
642: $z=-\eta$.
643: The distance between them is
644: given by $\Delta_d$. (For our gedanken experiment the maximum value of
645: $\Delta_d$ is $0.24~nm$.)
646: The formula for $\sigma_d$ contains three terms. The first term, $1/2$,
647: corresponds
648: to the quantum dispersion of the initial wave function. The second term,
649: $D\beta\tau$, describes the
650: initial diffusion of an ensemble of oscillators. Formally, setting
651: $\tau\sim 1/\beta$, we can estimate
652: the final dispersion $\sigma_d=\sqrt{D}$, which corresponds to
653: thermodynamical vibrations of the
654: cantilever tip. The third term describes insignificant oscillations
655: with small amplitude, $D\beta$.\\ \\
656: Note that the condition for distinguishing two cantilever positions at
657: the beginning of the
658: transient process
659: is much less restrictive than the corresponding condition for the
660: equilibrium positions
661: at $\beta\tau>>1$.
662: Indeed, after the first half-period ($\tau=\pi$), we have $\Delta_d=4\eta$
663: and
664: $\sigma_d=(1/2+ \pi D\beta)^{1/2}$. Taking into account that $\beta\ll 1$,
665: the condition for
666: distinguishing two positions,
667: $\eta>(1/2+\pi D\beta)^{1/2}$, is much easier than $\eta>\sqrt{D}$. In our
668: gedanken
669: experiment the condition for distinguishing the two positions for the
670: transient process is
671: \begin{equation}
672: T<T_{max}=\frac{4}{\pi}\frac{QF^2}{k_{_B}k_c}=14~K,
673: \label{eee}
674: \end{equation}
675: compared with $T<T_{max}=1.7~mK$ for the static Stern-Gerlach effect.
676: This estimate seems to be too
677: optimistic. It is connected with the very small distance ($5~nm$) between
678: the ferromagnetic particle
679: and the paramagnetic atom. If we increase this distance to $50~nm$, the
680: temperature $T_{max}$ drops
681: from $14~K$ to $1.1~mK$. Note that expression (\ref{eee}) coincides exactly
682: with our preliminary
683: estimates (2).\\ \ \\
684: \noindent
685: The condition $\Delta_d>2\sigma_d$ is satisfied for the first time at
686:
687: \begin{equation}
688: \tau=\tau_0\approx 2^{1/4}/\sqrt{\eta}.
689: \end{equation}
690: This expression is valid if $\eta>>1$ and $\eta>>(D\beta)^2/\sqrt{8}$.
691: For our gedanken experiment
692: we have $\eta=144$, $D=1.25\times 10^7T$ ($T$ is the temperature in Kelvin),
693: $\beta=1.5\times 10^{-4}$, and $\tau_0=0.1$. Thus, the above conditions
694: are both satisfied. The value of
695: $t_0=\omega_c\tau_0$ is approximately $9.3~\mu s$.
696:
697:
698: \section{\bf Solution for off-diagonal spin matrix elements}
699: The equations for the characteristics are now given by
700: \begin{equation}
701: d\tau={dr\over \beta r-k}={dk\over r-4\eta s}={d\hat\rho_{s,-s}\over
702: -D\beta r^2 \ \hat\rho_{s, -s}},
703: \label{chnd}
704: \end{equation}
705: or
706:
707: \begin{equation}
708: \begin{array}{lll}
709: {\ds dr\over\ds d\tau}&=\beta r-k,\\ \ \\
710: {\ds dk\over\ds d\tau}&= r-4\eta s,\\ \ \\
711: {\ds d\hat\rho_{s, -s}\over\ds d\tau}&=-D\beta r^2~\hat\rho_{s,-s}.\\
712: \end{array}
713: \label{chnd2}
714: \end{equation}
715: The solutions of the first two equations of (\ref{chnd2}) are
716:
717: \begin{equation}
718: \begin{array}{lll}
719: k&=e^{\beta \tau/2}\biggl(c_1\cos{\theta\tau}+c_2\sin{\theta\tau}\biggr)+
720: 4 \beta
721: \eta s \\
722: r&=e^{\beta\tau/2}\biggl[\left({\beta\over 2}\cos{\theta\tau}-\theta
723: \sin{\theta
724: \tau}\right)c_1+ \left({\beta\over 2}\sin{\theta\tau}+
725: \theta\cos{\theta\tau}\right)c_2\biggr]+ 4 \eta s.\\
726: \end{array}
727: \label{solnd2}
728: \end{equation}
729: Following the same steps as above we obtain for the Fourier transform:
730:
731:
732: \begin{equation}
733: \begin{array}{lll}
734: \hat\rho_{1/2, -1/2}(k,r,\tau)& \propto A(c_1,c_2)
735: \exp{\left\{\displaystyle -D\beta\biggl[
736: f_1c_1^2+2c_1c_2f_3+f_2c_2^2\biggr]\right\}}\\
737: &\quad\quad\quad\times\exp{\left\{ -D\beta\biggl[4 \eta
738: (g_1c_1+g_2c_2)+ 4\eta^2\tau\biggr]\right\}},
739: \end{array}
740: \label{rndf}
741: \end{equation}
742: where we fix $s=1/2$ (changing sign of $s$
743: corresponds to a
744: change of sign of $\eta$, see Eq.(\ref{chnd2}), therefore
745: the case $s=-1/2$ can be easily
746: obtained).
747: The functions $f_i$ and $g_i$ have been defined as above,
748: and $c_1$ and $c_2$ are new characteristic curves given by
749: \begin{equation}
750: \begin{array}{lll}
751: c_1 &=e^{-\beta\tau/2}( q_1 k+ q_2 r+ \eta q_3 ),\\
752: c_2 &=e^{-\beta\tau/2}( p_1 k+ p_2 r+ \eta p_3).\\
753: \label{ccnd}
754: \end{array}
755: \end{equation}
756: Here, $q_1, q_2, p_1, p_2$ are defined by Eqs. (\ref{qps}) and $q_3,
757: p_3$ are given by
758:
759: \begin{equation}
760: \begin{array}{lll}
761: q_3 &={\ds {2} \over\ds \theta}\biggl[
762: -\beta \biggl({\beta\over 2}\sin{\theta\tau}+
763: \theta\cos{\theta\tau}\biggr)+\sin{\theta\tau} \biggr],\\
764: p_3 &={\ds {2}\over\ds \theta}\biggl[
765: \beta \biggl({\beta\over 2}\cos{\theta\tau}-
766: \theta\sin{\theta\tau}\biggr)-\cos{\theta\tau} \biggr].\\
767: \label{qp0}
768: \end{array}
769: \end{equation}
770: With the same initial condition Eq. (\ref{rof}), we can determine the
771: function $A (c_1, c_2 )$ and obtain
772:
773: \begin{equation}
774: \begin{array}{lll}
775: &\hat\rho_{1/2,-1/2}(k,r,\tau)\propto\hat\rho_{1/2,-1/2}\biggl(c_1+2
776: \beta \eta,
777: {1\over 2 }\beta c_1
778: + \theta c_2+2\eta ,0 \biggr)\\
779: &\quad \times\exp{\left\{\displaystyle -D\beta\biggl[
780: F_1c_1^2+2c_1c_2F_3+F_2c_2^2+4\eta (G_1c_1+G_2c_2)+4 \eta^2\tau\biggr]
781: \right\}},
782: \label{afu}
783: \end{array}
784: \end{equation}
785: where $F_i(\tau)$ and $G_i(\tau)$ are defined as above.
786: By substituting the initial condition (\ref{rof}) we have
787:
788: \begin{equation}
789: \begin{array}{lll}
790: &\hat\rho_{1/2,-1/2}(k,r,\tau)\propto\exp{\left\{ \displaystyle i\biggl[
791: (\frac{1}{2}p_0 \beta+z_0 ) c_1+p_0\theta c_2 +
792: 2\eta(p_0+z_0\beta) \biggr]\right\}}\\
793: &\quad\quad\quad\times \exp{\left\{\displaystyle -\biggl[
794: \left( {\beta^2\over 16}+
795: {1\over 4} \right) c_1^2+ {\beta\theta \over 4} c_1 c_2+
796: {\theta^2 c_2^2 \over 4} \biggr]\right\}}\\
797: &\quad\quad\quad\times \exp{\left\{\displaystyle -\biggl\{4\eta
798: \left[\left({3\beta\over 8}\right)
799: c_1+{\theta\over 4}c_2\right]+ \eta^2\left(
800: {1}+\beta^2\right)\biggr\}\right\}}\\
801: &\quad\quad\quad\times \exp{\left\{\displaystyle -D\beta \biggl[
802: F_1c_1^2+2c_1c_2F_3+F_2c_2^2+
803: 4\eta s (G_1c_1+G_2c_2)+\eta^2 \tau\biggr]\right\}},\\
804: \label{rof22}
805: \end{array}
806: \end{equation}
807: which can be written as
808:
809: \begin{equation}
810: \begin{array}{lll}
811: \hat\rho_{1/2,-1/2}(k,r,\tau)& \propto\exp{\left[\displaystyle
812: -r^2 C_{12}-r\eta C_{11}- \eta^2 C_{10} +i r C_{21}+ i\eta C_{20}\right]}\\
813: &\quad\times\exp{\left[\left(i B_{20} -r B_{11} - \eta B_{10} \right) k -
814: \sigma_*^2 k^2\right] },
815: \end{array}
816: \label{e38nd}
817: \end{equation}
818: where $\sigma_*$ is given by (\ref{esig}) and
819:
820: \begin{equation}
821: \begin{array}{lll}
822: C_{12} &= e^{-\beta\tau} \biggl[ \left({\ds\beta^2\over\ds 16}+
823: {\ds 1\over\ds 4} +D\beta F_1\right) q_2^2 +\left({\ds\beta\theta\over\ds 4}
824: +2D\beta F_3\right)q_2 p_2 + \left({\ds\theta^2\over\ds 4}+D\beta F_2
825: \right)p_2^2 \biggr],\\ \ \\
826: C_{11} &= e^{-\beta\tau} \biggl[ \left({\ds\beta^2\over\ds 16}+
827: {\ds 1\over\ds 4} +D\beta F_1\right)2 q_2 q_3 +
828: \left({\ds\beta\theta\over\ds 4} +2D\beta F_3\right)(q_2 p_3 + p_2 q_3)
829: \biggr]\\
830: &\quad + e^{-\beta\tau} \biggl[
831: \left({\ds\theta^2\over\ds 4 }+D\beta F_2 \right)2 p_2 p_3 \biggr] \\
832: &\quad+4 e^{-\beta\tau/2} \biggl[ \left({\ds 3\beta\over\ds 8 } +
833: D\beta G_1 \right)q_2 +
834: \left({\ds\theta\over\ds 4}
835: + D\beta G_2 \right)p_2 \biggr],\\ \ \\
836: C_{10} &= e^{-\beta\tau} \biggl[ \left({\ds\beta^2\over 16}+
837: {\ds 1\over\ds 4} +D\beta F_1\right) q_3^2 +\left({\ds\beta\theta\over\ds 4}
838: +2D\beta F_3\right)q_3 p_3 + \left({\ds\theta^2\over\ds 4 }+D\beta F_2
839: \right)p_3^2 \biggr] \\
840: &\quad+4 e^{-\beta\tau/2} \biggl[ \left({\ds 3\beta\over\ds 8} +
841: D\beta G_1 \right)q_3 +
842: \left({\ds\theta\over\ds 4 }
843: + D\beta G_2 \right)p_3 \biggr]\\
844: &\quad+ 4 \left({\ds 1\over\ds 4}+ \frac{\ds\beta^2}{\ds 4}
845: +D\beta\tau \right),\\ \ \\
846: C_{21} &= e^{-\beta\tau/2}\biggl[ \left( \frac{\ds p_0\beta}{\ds 2}+ z_0
847: \right)q_2
848: + p_0\theta p_2 \biggr], \\ \ \\
849: C_{20} &= e^{-\beta\tau/2}\biggl[ \left( \frac{\ds p_0\beta}{\ds 2}+ z_0
850: \right)q_3
851: + p_0\theta p_3 \biggr]+ 2 \left( p_0 + z_0 \beta \right), \\ \ \\
852: B_{11} &= e^{-\beta\tau} \biggl[ \left({\ds\beta^2\over 16}+
853: {\ds 1\over\ds 4} +D\beta F_1\right) 2 q_2 q_1
854: +\left({\ds\beta\theta\over\ds 4} +2D\beta F_3\right)
855: (q_1 p_2 +q_2 p_1)\biggr]\\
856: &\quad+ e^{-\beta\tau} \biggl[ \left({\ds\theta^2\over\ds 4}+ D\beta F_2
857: \right) 2 p_2 p_1 \biggr], \\ \ \\
858: B_{10} &= e^{-\beta\tau} \biggl[ \left({\ds\beta^2\over\ds 16}+
859: {\ds 1\over\ds 4} +D\beta F_1\right) 2 q_3 q_1
860: +\left({\ds\beta\theta\over\ds 4} +2D\beta F_3\right)
861: (q_1 p_3 +q_3 p_1)\biggr]\\
862: &\quad+ e^{-\beta\tau} \biggl[ \left({\ds\theta^2\over\ds 4}+D\beta F_2
863: \right) 2 p_3 p_1 \biggr] \\
864: &\quad+ 4 e^{-\beta\tau/2} \biggl[ \left({\ds 3\beta\over\ds 8 } +
865: D\beta G_1 \right)q_1 +
866: \left({\ds\theta\over\ds 4}
867: + D\beta G_2 \right)p_1 \biggr],\\ \ \\
868: B_{20} &= e^{-\beta\tau/2} \biggl[ \left(\frac{\ds 1}{\ds 2}p_0 \beta + z_0
869: \right)q_1 +p_0 \theta p_1 \biggr].
870: \label{cabnd}
871: \end{array}
872: \end{equation}
873:
874: Performing the inverse Fourier transform we obtain
875:
876: \begin{equation}
877: \begin{array}{lll}
878: \rho_{1/2,-1/2}(R,r,\tau) &=
879: \displaystyle \frac{a b^*}{\sqrt{\pi}\sigma_*}
880: \exp{\left[\displaystyle -r^2 C_{12}-r\eta C_{11}-\eta^2 C_{10} +i r C_{21}+
881: i\eta C_{20}\right] } \\
882: &\quad\times\exp{\left[\displaystyle
883: \left( -r B_{11} - \eta B_{10} +i B_{20} -i R \right)^2/ 4
884: \sigma_*^2\right] },\\
885: \rho_{-1/2,1/2}(R,r,\tau) &=
886: \displaystyle \frac{a^* b}{\sqrt{\pi}\sigma_*}
887: \exp{\left[\displaystyle -r^2 C_{12}+r\eta C_{11}-\eta^2 C_{10} +i r C_{21}-
888: i\eta C_{20}\right] } \\
889: &\quad\times\exp{\left[\displaystyle
890: \left( -r B_{11} + \eta B_{10} +i B_{20} -i R \right)^2/ 4 \sigma_*^2\right] }.
891: \label{aftnd}
892: \end{array}
893: \end{equation}
894: Taking the modulus one obtains
895:
896: \begin{equation}
897: \begin{array}{lll}
898: |\rho_{1/2,-1/2}(R,r,\tau)| &=
899: \displaystyle \frac{|a b^*|}{\sqrt{\pi}\sigma_*} \ e^{\xi \eta^2}
900: \ e^{\displaystyle - (r + r_0\eta)^2/2\tilde\sigma^2}
901: \ e^{\displaystyle -( B_{20} - R )^2/ 4 \sigma_*^2 },\\
902: |\rho_{-1/2,1/2}(R,r,\tau)| &=
903: \displaystyle \frac{|a^* b|}{\sqrt{\pi}\sigma_*} \ e^{\xi \eta^2}
904: \ e^{\displaystyle - (r - r_0\eta)^2/2\tilde\sigma^2}
905: \ e^{\displaystyle -( B_{20} - R )^2/ 4 \sigma_*^2 },
906: \label{aftndm}
907: \end{array}
908: \end{equation}
909: where
910:
911: \begin{equation}
912: \begin{array}{lll}
913: \tilde\sigma^2 &= \frac{\ds 2\sigma_*^2}{\ds 4\sigma_*^2 C_{12} -
914: B_{11}^2},\\ \\
915: r_0&= \frac{\ds 2\sigma_*^2 C_{11}-B_{11}B_{10}}
916: {\ds 4\sigma_*^2 C_{12} - B_{11}^2},\\ \\
917: \xi&= \frac{\ds B_{10}^2}{\ds 4\sigma_*^2}- C_{10}-\frac{\ds r_0^2}
918: {\ds 2\tilde\sigma^2}.
919: \label{wndm}
920: \end{array}
921: \end{equation}
922: The maxima are located at
923: $(R=B_{20}, r=-\eta r_0)$ for $|\rho_{1/2,-1/2}|$)
924: and at $(R=B_{20}, r=\eta r_0)$ for $|\rho_{-1/2,1/2}|$.
925: In $(z,z')$ coordinates this corresponds to
926: $M_{+-}= (z=B_{20} -\eta r_0/2, z'=B_{20}+\eta r_0/2)$
927: for $|\rho_{1/2,-1/2}|$ and
928: $M_{-+}= (z=B_{20} +\eta r_0/2, z'=B_{20}-\eta r_0/2)$
929: for $|\rho_{-1/2,+1/2}|$, so that the distance between them
930: is given by $\Delta_{nd} = \sqrt{2} \eta |r_0|.$
931: Next, we consider the quadratic form
932: $ (r\pm r_0\eta)/2\tilde\sigma^2 + (B_{20}-R)^2/4\sigma_*^2$
933: in the $(z,z')$ plane. Straightforward calculations show that this is an
934: ellipse
935: whose semi-axes are respectively given by $\tilde\sigma$ (across the diagonal)
936: and $2\sqrt{2}\sigma_*$ (along the diagonal).
937: The centers of the peaks $M_{+-}, M_{-+}$ are symmetric with respect to
938: the diagonal line $z=z'$.\\ \\
939: \noindent
940: The most remarkable difference compared with the diagonal case, is
941: the presence of irreversible decoherence. Indeed, the heights of the
942: peaks are exponentially reduced in time by the damping factor
943: $ \sim
944: \exp(-4\eta^2 D \beta\tau)$. This, in turn, defines a characteristic
945: time scale of decoherence: $\tau_d=1/4\eta^2D\beta$. This formula exactly
946: coincides with
947: the expression derived in \cite{g}, based on a semi-qualitative analysis. At
948: time $\tau=\tau_0$,
949: when two diagonal peaks are clearly separated, the damping factor is
950: $4\eta^2D\beta\tau_0$.
951: We expect to observe the coherence between the two peaks (MSCS) if this
952: factor is not
953: much more than one unit. Thus, using the expression for $\tau_0$ from Eq.
954: (40), we can estimate
955: the condition for the quantum coherence as
956: $D\beta\eta^{3/2}<1$, or
957:
958: \begin{equation}
959: T<T_{max}={\ds Q\over\ds k_{_B}}{\ds (\hbar\omega_c)^{7/4}k_c^{3/4}\over\ds
960: F^{3/2}}.
961: \end{equation}
962: For our gedanken experiment the value of $T_{max}$ is approximately
963: $3\times 10^{-7}~K$.\\ \ \\
964: Now, we will check the validity of our estimate for the parameters chosen
965: for our gedanken experiment.
966: Our solution is valid if
967: $\eta>>(D\beta)^2/\sqrt{8}$. Setting $T=T_{max}$ or $D\beta\eta^{3/2}=1$,
968: we obtain
969: $\eta^4>>1/\sqrt{8}$, which is definitely true, assuming $\eta>>1$. Next,
970: the condition of
971: the validity of the high temperature approximation is $D>>1$. For $T=T_{max}$,
972: it follows that
973: $\eta^{3/2}<<Q$. This inequality is roughly satisfied ($\eta^{3/2}=1700$,
974: $Q=6700$). Finally,
975: as we mentioned in the Introduction, the master equation fails at times
976: $t\leq\hbar/k_{_B}T$. Thus,
977: the time considered, $\tau_0=2^{1/4}/\sqrt{\eta}$, must be much greater
978: than $1/D$, which is
979: definitely wrong. Thus, our condition (54) for the creation of MSCS is
980: not justified
981: for the parameters considered. \\ \ \\
982: Next, we discuss at what values of the parameters a MSCS can be generated.
983: First, we emphasize the qualitative difference between two conditions. 1)
984: The condition
985: for distinguishing two positions of the cantilever. 2) The condition for
986: distinguishing two position
987: of the cantilever and the coherence between these two positions.
988: The first condition is relatively simple: $T<T_{max}=F^2/k_{_B}k_c$ for
989: equilibrium positions
990: ($\tau>>Q$) and
991: $T<T_{max}=4QF^2/\pi k_{_B}k_c$ for $\tau=\pi$. The obvious way to
992: increase $T_{max}$ is by decreasing
993: the spring constant $k_c$ or increasing the magneto-static force $F$.
994: For $\tau=\pi$, one additional
995: way is to increase the quality factor $Q$.
996: Condition 2) for generating an MSCS can be satisfied at temperature
997: $T<T_{max}=\hbar^{7/4}Qk_c^{13/8}/k_{_B}F^{3/2}m^{7/8}$,
998: where $m$ is the effective mass of the cantilever, $m=k_c/\omega_c^2$.
999: At $T=T_{max}$, the MSCS will be generated if we satisfy the inequalities
1000: $1\ll\eta^2\ll Q$, or
1001:
1002: \begin{equation}
1003: 1<<{\ds m F^2\over\ds \hbar k_c^{3/2}}\ll Q.
1004: \label{eqq}
1005: \end{equation}
1006: One can see that the regime considered in
1007: our paper does not allow free manipulation of any parameter but $Q$.
1008: Increasing $Q$, we can
1009: satisfy the second inequality and, at the same time, increase $T_{max}$.
1010: In our gedanken experiment,
1011: the tenfold increase of $Q$ ($Q=67000$) provides the validity of the
1012: right-hand inequality in
1013: (\ref{eqq}) and increases $T_{max}$ to $3~\mu K$.
1014:
1015: \section{\bf Conclusion}
1016:
1017: We have shown that the maximum temperature for a single spin measurement
1018: in MFM can be increased
1019: by a factor of $Q$ if one utilizes the initial transient process instead
1020: of the static
1021: displacement of the cantilever tip. We have obtained an exact analytical
1022: solution
1023: of the master equation, which describes the $Q$-times magnification of
1024: the maximum temperature.
1025: In addition, we have found the conditions for generation of macroscopic
1026: Schr\"odinger cat state in MFM.
1027: \\ \ \\
1028: \section{Acknowledgments}
1029:
1030: The work of GPB was supported by the Department of Energy (DOE)
1031: under the contract W-7405-ENG-36. GPB and VIT thank
1032: the National Security Agency (NSA), the Advanced Research and
1033: Development Activity (ARDA), and DARPA (MOSAIC) for partial support.
1034:
1035:
1036: \vfil\eject
1037:
1038: \begin{thebibliography}{99}
1039: %
1040: \bibitem{g}
1041: G.P. Berman, G.D. Doolen, P.C. Hammel, and V.I.Tsifrinovich, Phys. Rev. A,
1042: {\bf 6503} (2002) 032311.\\
1043: \bibitem{b}
1044: A.O. Caldeira and A.T. Leggett, Physica A, {\bf 121} (1983) 587;\\
1045: Ann. Phys. (N.Y.) {\bf 149} (1983) 374.\\
1046: \bibitem{d}
1047: W.G. Unruh and W.H. Zurek, Phys. Rev. D, {\bf 40} (1989) 1071.\\
1048: \bibitem{e}
1049: A. Venugopalan, Phys. Rev. A, {\bf 56} (1997) 4307.\\
1050: \bibitem{f}
1051: B.L. Hu, J.P. Paz, and Y. Zhang, Phys. Rev. D, {\bf 45} (1992) 2843.\\
1052: \bibitem{i}
1053: T.D. Stowe, K. Yasumura, T.W. Kenny, D. Botkin, K. Wago, and D. Rugar, Appl.
1054: Phys. Lett., {\bf 71} (1997) 288.\\
1055: \end{thebibliography}
1056:
1057: \vfil\eject
1058:
1059: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1060:
1061: \begin{figure}
1062: \begin{center}
1063: \epsfxsize 14cm \epsfbox{1.ps}
1064: \caption{
1065: Geometry of the proposed gedanken experiment. ${\bf m}_{_F}$ and ${\bf m}_p$
1066: are
1067: the magnetic moments of the ferromagnetic particle and paramagnetic atom.
1068: }
1069: \label{1}
1070: \end{center}
1071: \end{figure}
1072:
1073:
1074: \begin{figure}
1075: \begin{center}
1076: \epsfxsize 14cm \epsfbox{2.ps}
1077: \caption{
1078: Schematic view of the Gaussians representing the diagonal elements
1079: $|\rho_{-1/2,-1/2}|$
1080: and $|\rho_{1/2,1/2}|$ (seen from the top) in the ($z,z'$) plane. We show
1081: the centers
1082: $M_{--}$ and $M_{++}$, the variances $\sigma_d'$ (transverse) and $\sigma_d$
1083: (parallel), and
1084: the distance between diagonal centers $\Delta_d$.
1085: }
1086: \label{xxx2}
1087: \end{center}
1088: \end{figure}
1089:
1090: \end{document}
1091:
1092:
1093:
1094:
1095:
1096:
1097:
1098:
1099:
1100:
1101: