<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
	<head>
		<meta http-equiv="content-type" content="application/xhtml+xml;charset=utf-8" />
		<title>Spin Ratchets</title>
		<meta name="author" content="Manuel Strehl" />
		<meta name="DC.date" content="2007-01-17" scheme="DCTERMS.W3CDTF" />
		<meta name="DC.language" content="en" scheme="DCTERMS.RFC3066" />
		<meta name="DC.coverage" content="Universität Regensburg" />
		<meta name="slideshow.defaultFontSize" content="32" />
		<meta name="slideshow.showSplashScreen" content="yes" />
		<!--link rel="stylesheet" href="print.css" type="text/css" /-->
		<script type="text/javascript" src="slideshow.js"></script>
	</head>
	<body>
		<div id="content">

			<div class="slide">
				<h1>Spin-Orbit Based<br /> Spin Ratchets</h1>
				<h3 style="margin-top:2em;">Manuel Strehl</h3>
			</div>

			<div class="slide">
				<h2>Outline</h2>
				<ul>
					<li>A Short Introduction</li>
					<li>Spin Currents</li>
					<li>Spin-Orbit Based Spin Ratchets</li>
					<li>Numerical Results</li>
					<li>Ratchet Mechanism</li>
					<li>Role of Dresselhaus SO-Coupling</li>
					<li>Outlook, Conclusions</li>
				</ul>
			</div>
			
			<div class="slide">
				<h2 title="Ratchets">A Short Introduction</h2>
				<h3>Ratchets</h3>
				<p>
					Classical ratchets have two general properties:
				</p>
				<p style="float:right">
					<img alt="" src="images/linke.gif" /><br />
					<cite>H. Linke et.al.,<br /> Science <strong>286</strong>, 2314</cite>
				</p>
				<ul>
					<li>broken spacial symmetry (usually by a ratchet potential)</li>
					<li>away from thermal equilibrium</li>
				</ul>
				<p>
					Aim: <strong>Extract useful work out of unbiased fluctuations</strong>
				</p>
				<p>
					Ratchet condition to get directed current: <code class="tex">I(+U) \ne -I(-U)</code>
				</p>
				<p>
					<cite>P. Reimann, Phys. Rep. <strong>361</strong>, 57 (2002)</cite>
				</p>
			</div>

			<div class="slide">
				<h2 title="Spin Ratchets">Spin Ratchets</h2>
				<ul>
					<li>Precondition: <strong>Spin-selective ratchet potential</strong></li>
					<li>Possible realisation: <strong>Rashba spin-orbit coupling</strong> in a quantum wire</li>
					<li>Aim: <strong>Spin current</strong></li>
				</ul>
				<p>
					<cite>A. Pfund, D. Bercioux, K. Richter <code>cond-mat/0601118</code></cite>
				</p>
			</div>

			<div class="slide">
				<h3>Spin Currents</h3>
				<ul>
					<li><em>Electric circuit</em>: Electrons</li>
					<li><em>Spin circuit</em>: Spins</li>
					<li>Ratchet works as "battery", that is, as deliverer of defined spin states</li>
				</ul>
			</div>
			
			<div class="slide">
				<ul>
					<li>Spin current contributions, e.g., for the left lead:
						<ul>
							<li>Transmission from right: <code class="tex">I_{E,n\sigma}^{S,R}(x\in L) = -\frac{\hbar^2}{2m^\star}\sum_{n'}\left(T_{n'+,n\sigma} - T_{n'-,n\sigma}\right)</code></li>
							<li>Reflection from left: <code class="tex">I_{E,n\sigma}^{S,L}(x\in L) = -\frac{\hbar^2}{2m^\star}\left[\sigma - \sum_{n'}\left(R_{n'+,n\sigma} - R_{n'-,n\sigma}\right)\right]</code></li>
						</ul>
					</li>
				</ul>
				<p>
					<cite>Rashba, J. Supercond. <strong>18</strong>, 137 (2005)</cite>
				</p>
			</div>
			
			<div class="slide">
				<ul>
					<li>Add both: <code class="tex">I_S^L = -\frac{1}{4\pi} \int_0^\infty dE \left[ f(E;\mu_L)R_S(E) + f(E;\mu_R)T_S'(E) \right]</code></li>
					<li>...where <code class="tex">T_S'(E) = T_{++}' + T_{+-}' - T_{--}' - T_{-+}'</code></li>
					<li>From <code class="tex">\mathbb{S}^\dagger\mathbb{S} = \mathbb{I}</code> we get <code class="tex">T_S'(E) = -R_S(E)</code></li>
				</ul>
			</div>
			
			<div class="slide">
				<ul>
					<li>... so the spin current is <code class="tex">I_S^L = \frac{1}{4\pi} \int_0^\infty dE \left[ f(E;\mu_L) - f(E;\mu_R) \right]T_S'(E)</code></li>
				</ul>
				<p>
					We are interested in <code class="tex">\left&lt;I_S\right> = \frac{1}{2} [I_S^{R/L}(+U_0) + I_S^{R/L}(-U_0)]</code>
				</p>
				<p>
					<em>Problem:</em> Leads must be selected, because spin current is not conserved &rarr; 2 possible definitions
				</p>
			</div>
			
			<div class="slide">
				<h4>Two definitions</h4>
				<p>
					Spin current measured in the lead with <strong>lower chemical potential</strong>:
				</p>
				<p class="block">
					<code class="tex">\left&lt;\mathcal{I}_S(U_0)\right> = \frac{1}{2}\left[</code><code style="background-color:yellow" class="tex">I_S^R</code><code class="tex">(+U_0) + </code><code style="background-color:yellow" class="tex">I_S^L</code><code class="tex">(-U_0)\right] </code><br/><code class="tex">= \frac{1}{8\pi} \int_{E_C}^\infty \Delta f(E,U_0)\times[T_S(E,U_0)-T'_S(E,-U_0)]d\epsilon</code>
				</p>
				<p>
					with<br />
					<code class="tex">T_S(E,U_0) = \sum_{\sigma=\pm1\in L}\left[T_{+,\sigma}(E,U_0)-T_{-,\sigma}(E,U_0)\right]</code><br />
					<code class="tex">T'_S(E,U_0) = \sum_{\sigma=\pm1\in R}\left[T'_{+,\sigma}(E,U_0)-T'_{-,\sigma}(E,U_0)\right]</code>
				</p>
			</div>

			<div class="slide">
				<p>
					Spin current measured in <strong>the right lead</strong> for the two rocking situations:
				</p>
				<p class="block">
					<code class="tex">\left&lt;I_S^R(U_0)\right> = \frac{1}{2}\left[</code><code style="background-color:yellow" class="tex">I_S^R</code><code class="tex">(+U_0) + </code><code style="background-color:yellow" class="tex">I_S^R</code><code class="tex">(-U_0)\right] </code><br/><code class="tex">= \frac{1}{8\pi} \int_{E_C}^\infty \Delta f(E,U_0)\times[T_S(E,U_0)-T_S(E,-U_0)]d\epsilon</code>
				</p>
				<p>
					with<br />
					<code class="tex">T_S(E,U_0) = \sum_{\sigma=\pm1\in L}\left[T_{+,\sigma}(E,U_0)-T_{-,\sigma}(E,U_0)\right]</code>
				</p>
				<p>
					<cite>Matthias Scheid, Master's thesis</cite>
				</p>
			</div>
			
			<div class="slide">
				<h4>Comparison in Linear Response</h4>
				<p>
					In linear response (that is, for <code class="tex">\pm U_0 \to 0</code>):
				</p>
				<p class="block">
					<code class="tex">\lim_{U_0\to 0} \left&lt;I_S(U_0)\right> = \frac{1}{8\pi} \int_{E_C}^\infty \Delta f(E,0)\times[T_S(E,0)-T_S(E,0)]d\epsilon</code><br />
					<strong><code class="tex"> = 0</code></strong>
				</p>
				<p class="block">
					<code class="tex">\lim_{U_0\to 0} \left&lt;\mathcal{I}_S(U_0)\right> = \frac{1}{8\pi} \int_{E_C}^\infty \Delta f(E,0)\times[T_S(E,0)-T'_S(E,0)]d\epsilon</code><br />
					<strong>can be different from 0</strong>
				</p>
				<p>
					<em>Meaning</em>: Spin current &#8660; polarization differs in the two directions.
				</p>
			</div>
			
			<div class="slide">
				<p>
					We concentrate mostly on <code class="tex">\left&lt;I_S(U_0)\right></code> (measuring on one side):
				</p>
				<ul>
					<li>easier to measure</li>
				</ul>
			</div>
			
			<div class="slide">
				<h3>Spin Orbit Ratchets</h3>
				<h4>System</h4>
				<p class="block">
					<img alt="" src="images/scatterer.png" />
				</p>
				<p>
					<cite>Mireles and Kirczenow, Phys. Rev B <strong>64</strong>, 024426 (2001)</cite>
				</p>
			</div>
			
			<div class="slide">
				<h3>Ingredients</h3>
				<ul>
					<li>Identical, clean, semi-infinite leads</li>
					<li>
						<p style="float:right">
							<img alt="" src="images/voltage.png" height="150" />
						</p>
						Scattering region, spin-orbit coupling starts adiabatically
					</li>
					<li>energy barrier in the scatterer</li>
					<li>Unbiased "adiabatic" rocking</li>
				</ul>
				<p>
					We look at the system for <code class="tex">\pm U_0</code>:
				</p>
				<p class="block">
					<code class="tex">\left&lt;I_{C/S}\right> = \frac{1}{2} [I(+U_0) + I(-U_0)]</code>
				</p>
			</div>

			<div class="slide">
				<h3>Hamiltonian</h3>
				<p class="block">
					<code class="tex">\mathcal{H}_c = \frac{\hat{p}^2}{2m^\star} + V_c(z) + \frac{\hbar k_{SO}^R}{m^\star}(\hat{\sigma}_x\hat{p}_z-\hat{\sigma}_z\hat{p}_x) + V_b(x)</code>
				</p>
				<ul>
					<li><code class="tex">V_c(z)</code> confinement in z-direction</li>
					<li><code class="tex">\frac{\hbar k_{SO}^R}{m^\star}(\hat{\sigma}_x\hat{p}_z-\hat{\sigma}_z\hat{p}_x)</code> Rashba SO-coupling term</li>
					<li><code class="tex">V_b(x)</code> potential in x-direction</li>
				</ul>
			</div>
			
			<div class="slide">
				<p>
					Rashba interaction leads to a spin splitting in the dispersion relation:
				</p>
				<p class="block">
					<img alt="" src="images/dispersion_relation.png" height="400" />
				</p>
			</div>
			
			<div class="slide">
				<p>
					Additional term for the rocking potential:
				</p>
				<p class="block">
					<code class="tex">\mathcal{H}' = -eU(t)g(x,z;V)</code>
				</p>
			</div>
			
			<div class="slide">
				<h2>Numerical Results</h2>
				<p class="block">
					<img alt="" src="images/spin_current.png" width="600" height="480" /><br />
					<em>... to be explained</em>
				</p>
			</div>
			
			<div class="slide">
				<p style="float:right">
					<img alt="" src="images/scattering_potentials.png" /><br />
					<em>Series of potentials</em>
				</p>
				<ul>
					<li>Series of 5 potentials</li>
					<li>Width: 15a <code class="tex">\propto 3\lambda_F</code></li>
					<li>Energy scales: <code class="tex">[\frac{\hbar^2}{2m^\starL}]</code>, where <code class="tex">L</code> is the length of one scattering potential</li>
					<li>Rocking situation</li>
					<li>&rarr; 3 open modes</li>
				</ul>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/step_tpp.png" width="600" height="480" /><br />
					<em><code class="tex">T_{++}</code> in one rocking situation</em>
				</p>
			</div>
		
			<div class="slide">
				<p class="block">
					<img alt="" src="images/step_tmm.png" width="600" height="480" /><br />
					<em><code class="tex">T_{--}</code> in one rocking situation</em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/step_pol.png" width="600" height="480" /><br />
					<em><code class="tex">T_{++} + T_{+-} - T_{--} - T_{-+} = T_S(+U_0)</code> in one rocking situation</em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/step_i_s.png" width="600" height="480" /><br />
					<em>Resulting <code class="tex">\Delta T_S = T_S(+U_0) - T_S(-U_0)</code></em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/spin_current.png" width="600" height="480" /><br />
					<em>Full picture</em>
				</p>
			</div>
			
			<div class="slide">
				<p>
					<code class="tex">T_S - T'_S</code> for a scatterer with 5 serial adiabatic potentials, 3 open modes in linear response:
				</p>
				<p class="block">
					<img alt="" src="images/t-tp_spin_current.png" width="600" height="480" />
				</p>
			</div>
			
			<!--div class="slide">
				<h3>Increase with Broadening Wire</h3>
				<p>
					3 open modes (width ~ 15a):<br />
					<img alt="" src="images/t-tp_3channels.png" width="600" height="480" />
				</p>
			</div>
			
			<div class="slide">
				<h3>Increase with Broadening Wire</h3>
				<p>
					5 open modes (width ~ 22a):<br />
					<img alt="" src="images/t-tp_5channels.png" width="600" height="480" />
				</p>
			</div>
			
			<div class="slide">
				<h3>Increase with Broadening Wire</h3>
				<p>
					8 open modes (width ~ 35a):<br />
					<img alt="" src="images/t-tp_8channels.png" width="600" height="480" />
				</p>
			</div-->
			
			<div class="slide">
				<h2 title="Ratchet Mechanism">The Ratchet Mechanism</h2>
				<h3>Landau-Zener Approach</h3>
				<p class="block">
					<img alt="" src="images/barrier.png" height="400" /><br />
					<em>Examine a particle moving to the right</em>
				</p>
			</div>
			
			<div class="slide">
				<p style="float:right">
					<img alt="" src="images/landau-zener_prob.png" /><br />
				</p>
				<ul>
					<li>Landau-Zener probability in avoided level crossings <code class="tex">P' = e^{-2\pi \nu}</code> to cross the gap</li>
					<li>We need the probability to leave the original branch: <code class="tex">P = 1 - P' = 1 - e^{-2\pi \nu}</code></li>
					<li>here: <code class="tex">\nu = \frac{|\left&lt;2+|\hat{H}_I|1-\right>|^2}{\hbar|\frac{\partial}{\partial t}[E(2+)-E(1-)]|}</code></li>
					<li>look for large &nu;</li>
				</ul>
			</div>
			
			<div class="slide">
				<p>
					So we get for the spin transmission:
				</p>
				<table style="border-spacing:0">
					<tr>
						<th style="border-bottom:1px solid #000;border-right:1px solid #000;"><code class="tex">T_{m,\sigma',n,\sigma}</code></th>
						<th style="border-bottom:1px solid #000;">(1+)</th>
						<th style="border-bottom:1px solid #000;">(1-)</th>
						<th style="border-bottom:1px solid #000;">(2+)</th>
						<th style="border-bottom:1px solid #000;">(2-)</th>
					</tr>
					<tr>
						<th style="border-right:1px solid #000;">(1+)</th>
						<td>1</td>
						<td>0</td>
						<td>0</td>
						<td>0</td>
					</tr>
					<tr>
						<th style="border-right:1px solid #000;">(1-)</th>
						<td>0</td>
						<td>(1-P)(1-P)</td>
						<td>(1-P)P</td>
						<td>0</td>
					</tr>
					<tr>
						<th style="border-right:1px solid #000;">(2+)</th>
						<td>0</td>
						<td>P(1-P)</td>
						<td>PP</td>
						<td>0</td>
					</tr>
					<tr>
						<th style="border-right:1px solid #000;">(2-)</th>
						<td>0</td>
						<td>0</td>
						<td>0</td>
						<td>0</td>
					</tr>
				</table>
				<p>
					and finally <code class="tex">T_{-+} = T_{+-}</code>, but <code class="tex">T_{++} \ne T_{--}</code>
				</p>
				<p>
					Consequence: <strong>The more adiabatic <code class="tex">U(x)</code> is, the better is the ratchet.</strong>
				</p>
			</div>
			
			<div class="slide">
				<h3>Deficiancies of the Model</h3>
				<!--ul>
					<li>Approximation won't hold for broader wires.</li>
				</ul-->
			</div>
			
			<div class="slide">
				<ul>
					<li>Use of a square scatterer instead of an adiabatic potential also leads to significant spin current:
						<p class="block">
							<img alt="" src="images/spin_current_square_scatterer.png" width="600" height="480" />
						</p>
					</li>
				</ul>
			</div>
			
			<div class="slide">
				<h2 title="Dresselhaus">Role of Dresselhaus Coupling</h2>
				<p class="block">
					<code class="tex">\mathcal{H}_D = \frac{\hbar k_{SO}^D}{m^\star}(\hat{\sigma}_x\hat{p}_x-\hat{\sigma}_z\hat{p}_z)</code>
				</p>
				<p>
					depends on the <strong>direction of the crystallographic axes</strong>.
				</p>
				<p>
					&rArr; additional terms to the tight-binding Hamiltonian <code class="tex">\propto sin(2\phi)</code> and <code class="tex">\propto cos(2\phi)</code>
				</p>
				<p>
					&rArr; <em>dependence of the spin current from the orientation of the scatterer</em>
				</p>
				<p>
					<cite>Ganichev et. al., Phys. Rev. Lett. <strong>92</strong>, 25 (2004)</cite>
				</p>
			</div>

			<div class="slide">
				<p class="block">
					<img alt="" src="images/angle1.png" width="600" height="480" /><br />
					<em><code class="tex">T_{++}</code> for one rocking situation</em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/angle2.png" width="600" height="480" /><br />
					<em>Spin polarization for one rocking situation</em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/angle3.png" width="600" height="480" /><br />
					<em>Resulting <code class="tex">\Delta T_S</code></em>
				</p>
			</div>
			
			<div class="slide">
				<p class="block">
					<img alt="" src="images/angle3.png" width="600" height="480" /><br />
					<em>Note the vanishing <code class="tex">\Delta T_S</code> for <code class="tex">k_\alpha = k_\beta</code> (i.e., for identical Rashba and Dresselhaus coupling)</em>
				</p>
			</div>
			
			<div class="slide">
				<h2>Conclusion</h2>
				<ul>
					<li>Spin ratchet based on Rashba SO coupling</li>
					<li>Spin current even for vanishing charge current</li>
					<li>Spin current can be tuned by
						<ul>
							<li>width of the wire</li>
							<li>Rashba strength and gate voltage</li>
							<li>orientation of the system</li>
						</ul>
					</li>
				</ul>
			</div>
			
			<div class="slide">
				<h2>Outlook</h2>
				<ul>
					<li>Investigate channel openings and closings</li>
					<li>Analyze channel contributions in square barriers</li>
					<li>Analyze shot noise properties of the system</li>
				</ul>
			</div>
			
			<div class="slide">
				<h2>Acknowledgements</h2>
				<ul>
					<li>Dario Bercioux</li>
					<li>Andreas Pfund</li>
					<li>Matthias Scheid</li>
					<li>Klaus Richter</li>
				</ul>
			</div>
			
		</div>
	</body>
</html>

