tprak

Theoretical Physics Practical Training
git clone git://popovic.xyz/tprak.git
Log | Files | Refs

main.tex (26019B)


      1 \documentclass[a4paper]{article}
      2 
      3 \usepackage[T1]{fontenc}
      4 \usepackage[utf8]{inputenc}
      5 
      6 \usepackage{mathptmx}
      7 
      8 \usepackage[a4paper, total={6in, 8in}]{geometry}
      9 \usepackage{subcaption}
     10 \usepackage[shortlabels]{enumitem}
     11 \usepackage{amssymb}
     12 \usepackage{amsthm}
     13 \usepackage{mathtools}
     14 \usepackage{braket}
     15 \usepackage{bbm}
     16 \usepackage{graphicx}
     17 \usepackage{float}
     18 \usepackage{yhmath}
     19 \usepackage{tikz}
     20 \usetikzlibrary{calc,decorations.markings}
     21 \usepackage[colorlinks=true,naturalnames=true,plainpages=false,pdfpagelabels=true]{hyperref}
     22 \usepackage[parfill]{parskip}
     23 \usepackage[backend=biber, sorting=none]{biblatex}
     24 \newcommand{\hbbar}{{\raisebox{0.05ex}{$\mathchar '26$}\mkern -9mu\raisebox{-0.15ex}{$\mathchar '26$}\mkern -9muh}}
     25 \addbibresource{uni.bib}
     26 \pagestyle{myheadings}
     27 \markright{Popovic, Vogel\hfill Dispertion relations \hfill}
     28 
     29 
     30 \title{Universität Wien\\ Fakultät für Physik\\
     31 \vspace{1.25cm}Laborpraktikum Theoretische Physik 2021S \\ Dispertion relations
     32 }
     33 \author{Milutin Popovic \& Tim Vogel \vspace{1cm}\\ Betreuer: Peter Stoffer}
     34 \date{30. Juni, 2021}
     35 
     36 \begin{document}
     37 \maketitle
     38 \noindent\rule[0.5ex]{\linewidth}{1pt}
     39 \begin{abstract}
     40 \end{abstract}
     41 \noindent\rule[0.5ex]{\linewidth}{1pt}
     42 \newcommand{\PV}{\mathop{\mathrlap{\pushR}}\!\int}
     43 \newcommand{\pushR}{\mathchoice
     44   {\mkern2.5mu P}
     45   {\scriptstyle P}
     46   {\scriptscriptstyle P}
     47   {\scriptscriptstyle P}
     48 }
     49 
     50 \tableofcontents
     51 
     52 \section{Introduction}
     53 Within the tools of complex analysis, there exists the possibility, to form
     54 relations between observable quantities of physicals systems, for example
     55 dispersion in a dielectric medium. This can be taken even further, by using the
     56 same methods within particle-physical problems, where the now more popular
     57 methods of qauntum chromodynamics do not apply, which would be low energy
     58 hadronic processes. We will firstly apply these concepts to the simple example
     59 of the harmonic oscillator and finally work out more complex problems,
     60 regarding the pion vector form factor. The reader is expected to be familiar
     61 with the subject of complex analysis, especially analyticity/holomorphicity of
     62 a function, integration of complex functions, the residue theorem and the
     63 Schwartz reflection principle.
     64 
     65 
     66 
     67 
     68 
     69 
     70 
     71 
     72 
     73 
     74 \section{Damped harmonic oscillator}
     75 Considering a free harmonic oscillator, the equation of motion accounts to:
     76 \begin{equation}
     77     \Ddot{x}(t)+\gamma\Dot{x}(t)+\omega_0^2x(t)=0
     78 \end{equation}
     79 where $\gamma > 0$ is the damping coefficient, and $\omega_0$ the angular frequency
     80 of the oscillator. Using the exponential ansatz we can arrive at an general
     81 solution to this ordinary differential equation
     82 \begin{align}
     83     &x(t) = a e^{-i\omega_1 t} + b e^{-i\omega_2 t} \\
     84     &\nonumber \\
     85     \text{with:} \nonumber\\
     86     &\omega_{1/2} = \pm \sqrt{\omega_0^2 - (\frac{\gamma}{2})^2} -
     87     i\frac{\gamma}{2}
     88 \end{align}
     89 where $a$ and $b$ are calculated based on the Couchy boundary conditions.
     90 
     91 For the case $\omega_0 > \frac{\gamma}{2}$ we can rewrite the solution
     92 \begin{align}
     93     &x(t) = \left(a e^{-i\tilde{\omega}_0 t} + b e^{-i\tilde{\omega}_0
     94     t}\right) e^{-\frac{\gamma}{2}t}\\
     95     &\nonumber \\
     96     \text{with:} \nonumber\\
     97     &\tilde{\omega}_{0} = \sqrt{\omega_0^2 - \left(\frac{\gamma}{2}\right)^2}
     98 \end{align}
     99 \subsection{External Force}
    100 
    101 Now consider a harmonic oscillator with an external force $F(t)$ driving it
    102 \begin{align}\label{eq:force}
    103     \Ddot{x}(t)+\gamma\Dot{x}(t)+\omega_0^2x(t)=\frac{F(t)}{m} =: f(t).
    104 \end{align}
    105 By Fourier transforming the equation we can arrive at an equation for the
    106 greens function in Fourier space. Note that the Fourier transform of
    107 $x(t)$ is
    108 \begin{align}
    109    \hat{x}(t) &= \frac{1}{2\pi}\int_\infty^\infty d\omega
    110     X(\omega) e^{-i\omega t}
    111 \end{align}
    112 so the Fourier transforms of $\dot{x}$ and $\Ddot{x}$ are
    113 \begin{align}
    114     \mathcal{F}(\dot{x}) &= -i\omega X(\omega)\\
    115     \mathcal{F}(\Ddot{x}) &= -\omega^2 X(\omega)\\
    116 \end{align}
    117 and the equation \ref{eq:force} turns into
    118 \begin{align}
    119     (-\omega^2 - i\gamma \omega + \omega_0^2) X(\omega) = F(\omega)
    120 \end{align}
    121 
    122 The Green's function can be represented in Fourier space like the following
    123 \begin{align}
    124     G(\omega)= \frac{1}{-\omega^2 - i\gamma \omega + \omega_0^2}
    125 \end{align}
    126 
    127 The Maximum of the squared modulus $|G(\omega)|^2$ for $\gamma \ll \omega_0$ is
    128 roughly at $\omega_0$, thus the width at half maximum can be calculated by
    129 looking for two $\omega$'s that satisfy
    130 \begin{align}
    131     \frac{1}{2}|G(\omega_0)|^2 &= |G(\omega)|^2\\
    132     \frac{1}{2} \frac{1}{\omega_0^2\gamma^2} &= |G(\omega)|^2
    133 \end{align}
    134 
    135 The exact solutions are
    136 \begin{align}
    137     \tilde{\omega}_1 &= \omega_0\sqrt{-0.5\left(\frac{\gamma}{\omega_0}\right)^2
    138     - 1.0\frac{\gamma}{\omega_0}(0.25\left(\frac{\gamma}{\omega_0}\right)^2
    139     + 1)^{\frac{1}{2}} + 1}\\
    140     \tilde{\omega}_2 &= \omega_0\sqrt{-0.5\left(\frac{\gamma}{\omega_0}\right)^2
    141     + 1.0\frac{\gamma}{\omega_0}(0.25\left(\frac{\gamma}{\omega_0}\right)^2
    142     + 1)^{\frac{1}{2}} + 1}
    143 \end{align}
    144 With help of Taylor expansion in the linear order in $\frac{\gamma}{\omega_0}$
    145 gives us the approximation for the with at half maximum
    146 \begin{align}
    147     \tilde{\omega}_2 - \tilde{\omega}_1 \simeq \gamma
    148 \end{align}
    149 
    150 In the figure below we plotted the squared modulus of $|G(\omega)|^2$
    151 \begin{figure}[H]
    152     \centering
    153     \includegraphics[width=\textwidth]{plots_sec2.png}
    154     \caption{On the left the squared modulus $|G(\omega)|^2$ in $\omega
    155 \in [\omega_0 - 2\gamma, w_0 + w\gamma]$ for $\gamma \ll \omega_0$, precisely
    156 $\gamma = \omega_0/20$ for $\omega_0 = 1$ and on the right $arg(G(\omega))$}
    157 \end{figure}
    158 
    159 
    160 Next we want calculate the Green's function in terms of time
    161 \begin{align}
    162     g(t) = \frac{1}{2\pi} \int^\infty_\infty d\omega G(\omega)e^{-i\omega t}.
    163 \end{align}
    164 Furthermore we can transform this to the complex integral where we have two singularities
    165 at We have two singularities at $z_{1/2} =  - \frac{i\gamma}{2} \pm
    166 \tilde{\omega}_{0}$. We have the following integral
    167 path
    168 
    169 \begin{center}
    170 \begin{tikzpicture}[decoration={markings,
    171 mark=at position 13cm with {\arrow[line width=2pt]{>}}
    172 }
    173 ]
    174 % The axes
    175 \draw[help lines,->] (-4,0) -- (4,0) coordinate (xaxis);
    176 \draw[help lines,->] (0,-3.5) -- (0,1) coordinate (yaxis);
    177 
    178 % The path
    179 \path[draw,line width=0.8pt,postaction=decorate]
    180     (3,0) node[above right] {} arc (0:-180:3) -- (-3, 0)
    181     node[above left] {} -- (-3, 0) -- (3, 0);
    182 
    183 % The labels
    184     \draw[thick, ->] (0,0) -- (2.1, -2.1) node[midway, fill=white] {$R$};
    185     \node[below] at (xaxis) {$\text{Re}$};
    186     \node[left] at (yaxis) {$\text{Im}$};
    187     \node at (0.2,-1.5) {$\nu$};
    188     \node at (1, -3.2) {$C(R)$};
    189 \end{tikzpicture}
    190 \end{center}
    191 
    192 The complex integral representation is
    193 \begin{align}
    194     \oint_\nu dz G(z) e^{-izt} &=
    195      \lim_{R \rightarrow \infty }
    196      \bigg(
    197     \int_{C(R)} + \int_{-R}^{R}
    198     \bigg)
    199     dz\ G(z) e^{-izt}\\
    200     &= 2\pi i\sum_j \text{I}(C_j, z_j) \text{Res}_j
    201 \end{align}
    202 Keep in mind that the integral from $R$ to $-R$ is the integral we are
    203 trying to solve, that is pulling the limit we have one integral over the real
    204 axis. Because of Jordan's lemma, the integral over the complex curve vanishes
    205 \begin{align}
    206     \big| \int_{C(R)} dz G(z) e^{-izt}\big| \leq \frac{\pi}{t}M_R
    207 \end{align}
    208 where $M_R:= \max_{C(R)}\{G(Re^{i\varphi})\}$.It can easily be seen that $M_R$
    209 converges to $0$ as $R$ goes to infinity. Thus the only value the integral can
    210 take is $0$ and we can calculate the real integral with the residues
    211 \begin{align}
    212     \text{Res}_1 &= \frac{e^{i z t}}{(z - z_1)(z - z_2)} (z - z_1)\bigg|_{z=z_1}
    213     \\
    214     &= -\frac{e^{-iz_1t}}{z_1 - z_2} = \frac{e^{-\frac{\gamma}{2}t}
    215     e^{i\tilde{\omega}_0 t}}{2\tilde{\omega}_0}\\
    216     \nonumber \\
    217     \text{Res}_2 &= \frac{e^{i z t}}{(z - z_1)(z - z_2)} (z - z_2)\bigg|_{z=z_2}
    218     \\
    219     &= -\frac{e^{-iz_1t}}{z_2 - z_1} = -\frac{e^{-\frac{\gamma}{2}t}
    220     e^{-i\tilde{\omega}_0 t}}{2\tilde{\omega}_0}\\
    221 \end{align}
    222 with the index $\text{I}(C_R, z_i)$ being $1$, because the curve goes around the
    223 singularities once.
    224 
    225 The integral evolves to
    226 \begin{align}
    227     \frac{1}{2\pi} \int_{-\infty}^{\infty} d\omega G(\omega) e^{-i\omega t}=
    228     \frac{\sin(\tilde{\omega}_0t)}{\tilde{\omega}_0} e^{-\frac{\gamma}{2}t}.
    229 \end{align}
    230 Treating the cases $t<0$ and $t>0$ separately we can join them with the
    231 Heaviside-theta function $\theta(t)$, the Green's function for the damped
    232 harmonic oscillator is
    233 \begin{align}
    234     g(t) = \frac{\sin(\tilde{\omega}_0t}{\tilde{\omega}_0}
    235     e^{-\frac{\gamma}{2}t} \theta(t)
    236 \end{align}
    237 With convolution we can arrive at a solution for the damped harmonic oscillator
    238 for an arbitrary driving force $f(t)$
    239 \begin{align}
    240     x(t) = \int_{-\infty}^{t} dt'
    241     \frac{\sin(\tilde{\omega}_0 (t-t'))}{\tilde{\omega}_0}
    242     e^{-\frac{\gamma}{2}(t-t')} f(t').
    243 \end{align}
    244 
    245 \subsection{Green's Function and dispersion relations}
    246 Next we want to compute the following integral
    247 \begin{align}
    248     0 = \oint_C d\omega' \frac{G(\omega')}{\omega - \omega'}, \;\;\;\;
    249     \text{with} \;\;\; \omega' = \omega_r + i\omega_i
    250 \end{align}
    251 along the following contour
    252 
    253 \begin{center}
    254 \begin{tikzpicture}[decoration={markings,
    255 mark=at position 0.5cm with {\arrow[line width=2pt]{>}},
    256 mark=at position 5cm with {\arrow[line width=2pt]{>}},
    257 mark=at position 13cm with {\arrow[line width=2pt]{>}},
    258 mark=at position 15cm with {\arrow[line width=2pt]{>}}
    259 }
    260 ]
    261 % The axes
    262 \draw[help lines,->] (-3.5,0) -- (3.5,0) coordinate (xaxis);
    263 \draw[help lines,->] (0,-0.5) -- (0,3.5) coordinate (yaxis);
    264 
    265 % The path
    266 \path[draw,line width=0.8pt,postaction=decorate]
    267     (2,0) -- (3, 0) node[below right] {} arc (0:180:3) -- (0.5, 0)
    268     arc (180:0:0.75);
    269 
    270 % The labels
    271     \draw[thick, ->] (0,0) -- (-2.1, 2.1) node[midway, fill=white] {$R$};
    272     \draw[thick, ->] (1.25,0) -- (1.75, 0.5) node[midway, above] {$\varrho$};
    273     \node[below] at (xaxis) {$\text{Re}$};
    274     \node[left] at (yaxis) {$\text{Im}$};
    275     \node[circle,inner sep=1pt,label=below:{$\omega$}, fill=black] at (1.25,0) {};
    276 \end{tikzpicture}
    277 \end{center}
    278 so the integral representation is
    279 \begin{align}
    280     \oint_C d\omega' \frac{G(\omega')}{\omega - \omega'} =
    281     \lim_{\substack{R\rightarrow \infty \\ \varrho \rightarrow 0^+}}
    282     \bigg( \int_{C(R)} + \int_{(C(\rho)} + \int_{-R}^{\omega -
    283     \varrho} + \int_{\omega +\varrho}^R
    284     \bigg)
    285     \;d\omega'\ \frac{G(\omega')}{\omega - \omega'}
    286 \end{align}
    287 We need show that the integral over the big circle goes to $0$. We know that for
    288 $\omega' \neq \omega$ we have
    289 \begin{align}
    290     \big|\frac{G(\omega')}{\omega' - \omega}\big| &=
    291     \big|\frac{1}{\omega'^3}\frac{1}{(1-\frac{\omega_1}{\omega'})(1-\frac{\omega_2}{\omega'})
    292     (1 - \frac{\omega}{\omega'})}\big|
    293     \leq \frac{1}{R^3}
    294 \end{align}
    295 thus
    296 \begin{align}
    297         \bigg|
    298     \int_{C(R)} d\omega' \frac{G(\omega')}{\omega - \omega'}
    299         \bigg| \leq \frac{2\pi R}{R^3} = \frac{2\pi}{R^2}
    300         \xrightarrow[R\rightarrow \infty]{} 0.
    301 \end{align}
    302 The small circle can be calculated with the Residue theorem with the pole at
    303 $\omega$
    304 \begin{align}
    305 \int_{C(\varrho)}d\omega' \frac{G(\omega')}{\omega - \omega'} = 2\pi i
    306     \text{I}(C(\varrho), \omega) \text{Res}(\frac{G(\omega')}{\omega - \omega'},
    307     \omega) = i\pi G(\omega).
    308 \end{align}
    309 Note that we go around $\omega$ only $1/2$ times. Reconstructing the integral
    310 equation we get
    311 \begin{align}
    312     -i\pi G(\omega) = \lim_{\varrho \rightarrow 0^+}
    313     \big(
    314      \int_{-R}^{\omega -\varrho} + \int_{\omega +\varrho}^R
    315     \big) \;d\omega'\ \frac{G(\omega')}{\omega' - \omega}
    316 \end{align}
    317 which is exactly the Cauchy Principal Value. Furthermore we can rewrite
    318 $G(\omega)$ into real and imaginary parts
    319 \begin{align}
    320     \text{Re} (G(\omega)) = \frac{1}{\pi} \PV  d\omega' \frac{\text{Im}
    321     (G(\omega'))}{\omega' - \omega}\\
    322     \text{Im} (G(\omega)) = \frac{1}{\pi} \PV  d\omega' \frac{\text{Re}
    323     (G(\omega'))}{\omega' - \omega}\\
    324 \end{align}
    325 which are Hilbert transforms of each other, the equations are also known
    326 as ``dispersion relations''. It should be noted that these equations also allow
    327 negative frequencies. Let us derrive an representation for only positive
    328 frequencies. We start off by a simple statement
    329 \begin{align}
    330     G(-\omega^*) = G(\omega)^*.
    331 \end{align}
    332 In our case this is obviously true
    333 \begin{align}
    334     &G(-\omega^*) = \frac{1}{-(\omega^*)^2 + i\gamma \omega^* + \omega_0^2}\\
    335     \nonumber \\
    336     &G(\omega)^* = \frac{1}{-(\omega^2)^* + i\gamma \omega^* + \omega_0^2} =
    337     G(-\omega^*)
    338 \end{align}
    339 Now we choose $\omega \in \mathbb{R}^+$, our relation then becomes
    340 $G(-\omega) = G(\omega)^*$. Then we get
    341 \begin{align}
    342     \text{Re} (G(\omega)) &= \frac{1}{\pi} \PV_0^\infty  d\omega' \frac{2\omega'\text{Im}
    343     (G(\omega'))}{\omega'^2 - \omega^2}\\
    344     \text{Im} (G(\omega)) &= -\frac{1}{\pi} \PV_0^\infty  d\omega' \frac{2\omega'\text{Re}
    345     (G(\omega'))}{\omega'^2 - \omega^2}
    346 \end{align}
    347 
    348 To round this chapter up we would like to show one last identity in the sense
    349 of distributions
    350 \begin{align}
    351     \lim_{\varepsilon \rightarrow 0^+} \frac{1}{\omega' -\omega \mp
    352     i\varepsilon} = \text{P}(\frac{1}{\omega' - \omega}) \pm i\pi\delta(\omega'
    353     - \omega).
    354 \end{align}
    355 Let us extend the fraction with $\omega' - \omega \pm i\varepsilon$.
    356 \begin{align}
    357 \frac{\omega' - \omega \pm i\varepsilon}{(\omega' -\omega \mp
    358     i\varepsilon)(\omega' - \omega \pm i\varepsilon)} = \frac{\omega' - \omega
    359     \pm i\varepsilon}{(\omega' - \omega)^2 + \varepsilon^2}.
    360 \end{align}
    361 That means for a test function $f(\omega')$ we have
    362 \begin{align}
    363     \lim_{\varepsilon \rightarrow 0^+} \int_{-\infty}^\infty d\omega'
    364     \frac{f(\omega')}{\omega' - \omega \mp i\varepsilon} &=
    365     \lim_{\varepsilon \rightarrow 0^+}
    366     \int_{-\infty}^\infty d\omega'
    367     \frac{(\omega' - \omega \pm i\varepsilon) f(\omega')}{(\omega' - \omega)^2
    368     + \varepsilon^2}\\
    369     & =
    370     \lim_{\varepsilon \rightarrow 0^+}\bigg(
    371     \int_{-\infty}^\infty d\omega'
    372     \frac{f(\omega')(\omega' - \omega )}{(\omega' - \omega)^2 + \varepsilon^2}
    373     \pm i\varepsilon
    374     \int_{-\infty}^\infty d\omega'\frac{f(\omega')}{(\omega' - \omega)^2 + \varepsilon^2}
    375     \bigg)\label{eq:id}.
    376 \end{align}
    377 Let us look into the first integral in equation \ref{eq:id}, we can rewrite it
    378 \begin{align}
    379     \lim_{\varepsilon \rightarrow 0^+}\int_{-\infty}^\infty
    380     d\omega'\frac{f(\omega')(\omega' - \omega )}{(\omega' - \omega)^2 + \varepsilon^2}
    381     &= \lim_{\varepsilon,\varrho \rightarrow 0^+}
    382     \bigg(
    383         \int_{-\infty}^{\omega - \varrho}d\omega'\frac{f(\omega')(\omega' -
    384         \omega )}{(\omega' - \omega)^2 + \varepsilon^2}
    385         + \int_{\omega
    386         \varrho}^{\infty}d\omega'\frac{f(\omega')(\omega' - \omega )}{(\omega'
    387         - \omega)^2 + \varepsilon^2} + \\
    388         &+\int_{\omega -\varrho}^{\omega+
    389         \varrho}d\omega'\frac{f(\omega')(\omega' - \omega )}{(\omega' -
    390         \omega)^2 + \varepsilon^2}
    391         \bigg)
    392         \\
    393         &= \PV_{-\infty}^{\infty}d\omega' \frac{f(\omega')}{(\omega' - \omega)}
    394 \end{align}
    395 The integral from $\omega - \varrho$ to $\omega + \varrho$ can be calculated by
    396 pulling out $f(\omega)$ out of the integral and directly computing it, which
    397 gives then vanishes. In second integral we approximate $f(\omega')$ to
    398 $f(\omega)$ in the region $\omega' \simeq \omega$
    399 \begin{align}
    400     \varepsilon \int_{-\infty}^{\infty}d\omega' \frac{f(\omega')}{(\omega' -
    401     \omega)^2 + \varepsilon^2} &\simeq  \varepsilon f(\omega) \int_{-\infty}^{\infty}
    402     \frac{1}{(\omega' -\omega)^2 + \varepsilon^2}\\
    403     &= \pi f(\omega).
    404 \end{align}
    405 Which means the identity is
    406 \begin{align}
    407     \lim_{\varepsilon \rightarrow 0^+} \int_{-\infty}^{\infty} d\omega'
    408     \frac{f(\omega')}{\omega' -\omega \mp i\varepsilon} =
    409     \PV_{-\infty}^{\infty} d\omega' \frac{f(\omega')}{\omega' -\omega} \pm
    410     i\pi f(\omega)\label{eq:pv}
    411 \end{align}
    412 \section{Potential scattering in quantum mechanics}
    413 If we consider elastic scattering of a spinless particle off a
    414 time-independent, spherically symmetric potential of finite range, we look for
    415 stationary solutions $\psi$ of the Schrödinger equation
    416 \begin{equation}
    417     -\frac{\hbbar^2}{2m}\Vec{\nabla^2}\psi(\Vec{x})+V(r)\psi(\Vec{x})=E\psi(\Vec{x})
    418 \end{equation}
    419 Since the potential is spherically symmetric, it only depends on $r$ and for
    420 large values of $r$, it can be shown, that the asymptotic form of $\psi$ looks
    421 like:
    422 \begin{equation}
    423     \psi(r,\theta)\approx A[e^{ikz}+f(E,\theta)\frac{e^{ikr}}{r}]
    424 \end{equation}
    425 Where $kr\gg 1$, and k given by
    426 \begin{equation}
    427     k=\frac{\sqrt{2mE}}{\hbbar}
    428 \end{equation}
    429 We also define the scattering angle $\theta$ by $z=r\cos{\theta}$, and since
    430 there is no dependence on the azimuthal angle $\phi$. we can define the
    431 incoming and outgoing parts of the wave function as follows:
    432 \begin{equation}
    433     \psi_{in}=Ae^{ikz}
    434 \end{equation}
    435 and
    436 \begin{equation}
    437     \psi_{out}=Af(E,\theta)\frac{e^{ikr}}{r}
    438 \end{equation}
    439 Where the factor $\frac{1}{r}$ is carried, to conserve probability. The complex
    440 function $f(E,\theta)$ is the so called scattering amplitude. We are now
    441 interested in the differential cross section $\frac{d\sigma}{d\Omega}$, which
    442 is defined as the ratio of number of particles per unit time, that are
    443 scattered into the surface element $dS=r^2d\Omega(\theta,\phi)$ and the number
    444 of incoming particles per unit time, per are orthogonal to the beam direction.
    445 Expressed via probability currents, we thus obtain:
    446 \begin{equation}
    447     \frac{d\sigma}{d\Omega}=\frac{\Vec{j}_{out}\cdot \Vec{e}_rr^2}{|\Vec{j}_{in}|}
    448 \end{equation}
    449 With $\Vec{e}_r$ being a unit vector in direction of the radius, and the currents given as:
    450 \begin{equation}
    451 \Vec{j}=\frac{i\hbbar}{2m}(\psi\Vec{\nabla}\psi^*-\psi^*\Vec{\nabla}\psi)
    452 \end{equation}
    453 By applying these equations, we obtain the differential crosssection as:
    454 \begin{equation}
    455     \frac{d\sigma}{d\Omega}=|f(E,\theta)|^2
    456 \end{equation}
    457 With the scattering amplitude, being given as:
    458 \begin{equation}
    459     f(E,\theta)=\sum_{l=0}^\infty(2l+1)f_l(E)P_l(cos\theta)
    460 \end{equation}
    461 where $l$ denotes the magnitude of orbital angular momentum, and
    462 $P_l(cos\theta)$ are the Legendre polynomials.
    463 We can now work out the total crosssection $\sigma$ via the integral:
    464 \begin{equation}
    465     \sigma=\int{d\Omega\frac{d\sigma}{d\Omega}}
    466 \end{equation}
    467 We do this, by applying the orthogonality relation
    468 \begin{equation}
    469     \int{d\Omega P_l(cos\theta)P_{l'}(cos\theta)}=\frac{4\pi}{(2l+1)}\delta_{ll'}
    470 \end{equation}
    471 Thus, we obtain:
    472 \begin{equation}
    473     \sigma=\sum_{l,l'}(2l+1)(2l'+1)f^*_l(E)f_{l'}(E)\int{d\Omega P_l(cos\theta)P_{l'}(cos\theta)}
    474 \end{equation}
    475 Which, finally leads to:
    476 \begin{equation}
    477     \sigma=4\pi\sum_l(2l+1)|f(E)|^2
    478 \end{equation}
    479 
    480 
    481 
    482 \section{The pion vector from factor and the Omn\`es Problem}
    483 Insert Text
    484 \subsection{Unitarity of the scattering matrix}
    485 Insert Text
    486 \begin{align}
    487     \label{eq:rec}
    488     \text{Im}(F^V_\pi(s) = F^V_\pi(s) e^{-i\delta_{\pi\pi}(s)}\sin\delta_{\pi\pi}(s)
    489 \end{align}
    490 \subsection{The Omn\`es Problem}
    491 The equations \ref{eq:rec} allow us to carefully reconstruct the pion Vector Form
    492 Factor, based on strictly formulated conditions. This is known as the Omn\`es
    493 Problem. First of all the equation tells us that $F_\pi^V(s)$ is a complex
    494 valued function, as $s$ is an analytic variable in the complex plane, apart
    495 from a cut complex s-plane $\Gamma = [s_0, \infty) \subset \mathbb{R}$, where
    496 $s = 4M_\pi^2 > 0$. To summerize the conditions are
    497 \begin{itemize}
    498     \item $F_\pi^V(s)$ is analytic on the cut complex s-plane
    499         $\mathbb{C}\backslash \Gamma$
    500     \item $F_\pi^V(s)\in \mathbb{R} \;\;\; \forall\; s \in
    501         \mathbb{R}\backslash \Gamma$
    502     \item $\lim_{\varepsilon \rightarrow 0}(F_\pi^V(s+i\varepsilon)
    503         e^{-i\delta_{\pi\pi}(s)}) \in \mathbb{R}$ on $\Gamma$ for a real
    504         bounded fucntion $\delta_{\pi\pi}(s)$
    505     \item We assume $F_\pi^V(0) = 1$ and $F_\pi^V(s)$ has no zeros.
    506 \end{itemize}
    507 We start off with the Couchy Integral
    508 \begin{align}
    509     \ln(F(s)) = \frac{1}{2\pi i}\oint_C ds' \frac{\ln(F(s'))}{s'-s}
    510 \end{align}
    511 over the following contour
    512 \begin{center}
    513 \begin{tikzpicture}[decoration={markings,
    514 mark=at position 0.5cm with {\arrow[line width=2pt]{>}},
    515 mark=at position 5cm with {\arrow[line width=2pt]{>}},
    516 mark=at position 13cm with {\arrow[line width=2pt]{>}},
    517 mark=at position 21cm with {\arrow[line width=2pt]{>}},
    518 mark=at position 23cm with {\arrow[line width=2pt]{>}}
    519 }
    520 ]
    521 % The axes
    522 \draw[help lines,->] (-3.5,0) -- (3.5,0) coordinate (xaxis);
    523 \draw[help lines,->] (0,-3.5) -- (0,3.5) coordinate (yaxis);
    524 
    525 % The path
    526 \path[draw,line width=0.8pt,postaction=decorate]
    527     (1.5,0.1) -- (3, 0.1) node[below right] {} arc (4:360:3) -- (1.5, -0.1)
    528     node[below right] {} arc(345:15:0.4);
    529 
    530 % The labels
    531     \draw[thick, ->] (0,0) -- (-2, 2) node[midway, fill=white] {$R$};
    532    \draw[thick, ->] (1.1,0) -- (1.1, 0.4) node[above] {$\varrho$};
    533     \node[below] at (xaxis) {$\text{Re}$};
    534     \node[left] at (yaxis) {$\text{Im}$};
    535     \node[circle,inner sep=1pt,label=below:{$s_0$}, fill=black] at (1.1,0) {};
    536 \end{tikzpicture}
    537 \end{center}
    538 This means the integral can be separated into
    539 \begin{align}
    540     \oint_C ds' \frac{\ln(F(s'))}{s'-s} =
    541     \lim_{\substack{\varepsilon \rightarrow 0}}
    542     \bigg(
    543     \int_{C(R)} + \int_{C(\varrho)} + \int_{s_0+i\varepsilon}^{\infty
    544     +i\varepsilon} +  \int^{s_0-i\varepsilon}_{\infty
    545     -i\varepsilon}
    546     \bigg) ds' \frac{\ln(F(s'))}{s'-s}
    547 \end{align}
    548 The integrals over $C(R)$ and $C(\varrho)$ dissapear. For the last two
    549 integrals we can use the Schwarz reflection principle and then we get
    550 \begin{align}
    551     \oint_C ds' \frac{\ln(F(s'))}{s'-s} = \frac{1}{\pi} \int_{s_0}^{\infty}ds'
    552     \text{Im}\left(
    553     \frac{\ln(F(s'))}{s'-s}\right)
    554 \end{align}
    555 where we used $\text{Im}(z) = \frac{z - z^*}{2i}$ to write the imaginary part
    556 here. We look now at equation \ref{eq:rec} and refactor it
    557 \begin{align}
    558     &\frac{F_\pi^V(s)-F_\pi^V(s)^*}{2i} = F_\pi^V e^{i\delta_{\pi\pi}}(s)
    559     \sin(\delta_{\pi\pi}(s)) \\
    560     &F_\pi^V(s) = F_\pi^V(s)^* e^{2i \delta_{\pi\pi}(s)}\\
    561     &\ln(F_\pi^V(s)) = \ln((F_\pi^V e^{-i\delta_{\pi\pi}})^*)+
    562     i\delta_{\pi\pi}(s).\label{eq:use}
    563 \end{align}
    564 Now we use this equation to compute the integral with variation in $s
    565 \rightarrow s+i\varepsilon$ as $\varepsilon$ goes to infinity.
    566 \begin{align}
    567     \ln(F_\pi^V(s)) &= \lim_{\substack{\varepsilon \rightarrow \infty}}
    568     \ln(F_\pi^V(s+i\varepsilon)) = \\
    569     &= \lim_{\substack{\varepsilon \rightarrow \infty}}\frac{1}{\pi}
    570     \int_{s_0}^{\infty}
    571     \text{Im}\left(\frac{F_\pi^V(s')}{s'-s-i\varepsilon}\right)=\\
    572     &=\lim_{\substack{\varepsilon \rightarrow \infty}}\frac{1}{\pi}
    573     \int_{s_0}^{\infty}
    574     \text{Im}\left(\frac{\ln((F_\pi^V e^{-i\delta_{\pi\pi}})^*)+
    575     i\delta_{\pi\pi}}{s'-s-i\varepsilon}\right)
    576 \end{align}
    577 the part $F_\pi^V e^{-i\delta_{\pi\pi}}$ needs to be real that means
    578 \begin{align}
    579     \ln(F_\pi^V(s)) &= \lim_{\substack{\varepsilon \rightarrow \infty}}\frac{1}{\pi}
    580     \int_{s_0}^{\infty}
    581     \frac{\delta_{\pi\pi}(s')}{s'-s-i\varepsilon} =\\
    582     &= \ln(F_\pi^V(0)) + \lim_{\substack{\varepsilon \rightarrow \infty}}
    583     \frac{s}{\pi}
    584     \int_{s_0}^{\infty}
    585     \frac{\delta_{\pi\pi}(s')}{s'(s'-s-i\varepsilon)}
    586 \end{align}
    587 with the condition $F_\pi^V(0) = 1$ and the relation from \ref{eq:pv} we get
    588 \begin{align}
    589     F_\pi^V(s) =  \exp
    590     \bigg(
    591         \frac{s}{\pi}\PV_{s_0}^\infty ds' \frac{\delta_{\pi\pi}(s')}{s'(s'-s)}
    592         + i\delta_{\pi\pi}(s)
    593     \bigg).
    594 \end{align}
    595 To compute the principal value integral we use the following trick
    596 \begin{align}
    597     \frac{s}{\pi}\PV_{s_0}^\infty ds' \frac{\delta_{\pi\pi}(s')}{s'(s'-s)} =
    598     \frac{2}{\pi} \int_{s_0}^\infty ds' \frac{\delta_{\pi\pi}(s')
    599     -\delta_{\pi\pi}(s)}{s'(s'-s)}+
    600     \delta_{\pi\pi}\frac{s}{\pi}\PV_{s_0}^\infty \frac{1}{s'(s'-s)}.
    601 \end{align}
    602 The first integral can be computed numerically, the second one has an analytic
    603 solution for $s > s_0$. We use the definition of the principal value and circle
    604 around $s$ in a small half circle with the radius $r$.
    605 \begin{align}
    606     \PV_{s_0}^\infty \frac{1}{s'(s'-s)} = \lim_{\substack{r\rightarrow0}}
    607     \bigg(
    608     \int_{s_0}^{s-r}ds'  + \int_{s+r}^{\infty} ds'
    609     \bigg)\frac{1}{s'(s'-s)}
    610 \end{align}
    611 then we simply integrate and plug in
    612 \begin{align}
    613     \PV_{s_0}^\infty \frac{1}{s'(s'-s)} &= \lim_{\substack{r\rightarrow 0}}
    614     \bigg(
    615     \frac{\ln(s'-s) - \ln(s')}{s}\big|_{s'= s_0}^{s'= s-r}
    616     \frac{\ln(s'-s)-\ln(s')}{s}\big|_{s'=s+r}^{s'=\infty}
    617     \bigg) =\\
    618     &=\frac{1}{s}\ln\left(\frac{s_0}{s_0-s}\right)
    619 \end{align}
    620 that means the second integral is
    621 \begin{align}
    622     \delta_{\pi\pi}\frac{s}{\pi}\PV_{s_0}^\infty \frac{1}{s'(s'-s)}
    623     =\delta_{\pi\pi}(s)\frac{1}{\pi} \ln\left(\frac{s_0}{s_0 -s }\right)
    624 \end{align}
    625 
    626 Lastly we would like to plot the modulus of the Omn\`es representation of the
    627 pion vector form
    628 factor. For the phase we would usually use experimentall value, but in our case
    629 we will use the Breit-Wigner representation of the pion VVF to compute the
    630 phase $\delta_{\pi\pi}$. A reminder the Breit-Wigner representation is the
    631 following
    632 \begin{align}
    633     F^V_\pi(s)_{BW} = \frac{M_\varrho^2}{M_\varrho^2 - s - iM_\varrho
    634     \Gamma_\varrho(s)}
    635 \end{align}
    636 where
    637 \begin{align}
    638     \Gamma_\varrho(s) := \Gamma_\varrho\frac{s}{M_\varrho^2} \left(
    639     \frac{\sigma_\pi(s)}{\sigma_\pi(M_\varrho^2)}
    640     \right)^3 \theta(s-4M_\pi^2), \;\;\;\;\;\; \sigma_\pi(s) :=
    641     \sqrt{1-\frac{4M_\pi^2}{s}}.
    642 \end{align}
    643 Thus our phase shift will be
    644 \begin{align}
    645     \delta_{\pi\pi}(s) := \arg\left(F_\pi^V(s)_{BW}\right)
    646 \end{align}
    647 where we will use numerical values for $M_\varrho = 0.77\ \text{GeV}$,
    648 $\Gamma_\varrho = 0.15\ \text{GeV}$, $M_\pi = 0.14 \text{GeV}$.
    649 \begin{figure}[H]
    650     \centering
    651     \includegraphics[width=0.9\textwidth]{./omnes_bw.png}
    652     \caption{Plot of the modulus of the Breit-Wigner(red) and the Omn\`es
    653     representation(black) of the pion Vector From Factor for $s \in [0, 1]$ in $GeV^2$}
    654 \end{figure}
    655 \nocite{mathe}
    656 \nocite{stoffer}
    657 \nocite{omnes}
    658 \printbibliography
    659 \end{document}