Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let's have the Dirac free lagrangian: $$ L = \bar {\Psi} (i\gamma^{\mu}\partial_{\mu} - m) \Psi . $$ I can rewrite it as $$ L = i\Psi^{\dagger}\partial_{0}\Psi - H_{d}, \quad H_{d} = \Psi^{\dagger}(-i(\hat {\alpha} \cdot \nabla ) + \beta m)\Psi , \quad \hat {\alpha} = \gamma_{0}\gamma . $$ Here $H_{d}$ is the Hamiltonian density. Is it possible to build Hamilton equations and Poisson brackets for spinors (for example, $\Psi$ may be the canonical coordinate and $\Psi^{\dagger}$ be the canonical impulse)?

share|cite|improve this question
Are you referring to quantum field theory or classical field theory? – Hunter Mar 19 '14 at 20:05
@Hunter : first I want to get the classical relations, and after that to quantize the Poisson brackets. – Andrew McAddams Mar 19 '14 at 20:08
@Hunter : and, of course, this formalism must correspond to the anticommutation relations for creation and destruction operators. But I can't think of such canonical variables which would satisfy these conditions. Particularly I don't understand how canonical Poisson brackets may give anticommutation relations for an operators. – Andrew McAddams Mar 19 '14 at 20:14
Related: – Qmechanic Mar 29 '14 at 22:17
up vote 4 down vote accepted

The Hamiltonian density for any classical field is given by: \begin{equation} \mathcal{H} = \pi \dot{\phi} - \mathcal{L} \end{equation} where $\pi$ is the canonical momentum density: \begin{equation} \pi(\mathbf{x},t) = \frac{\partial \mathcal{L}}{\partial \dot{\phi}(\mathbf{x},t)} \end{equation} In classical point particle mechanics the Poisson brackets for two functions $f$ and $g$ are defined as: \begin{equation} \left\{f,g\right\}_{PB} = \frac{ \partial f}{\partial q_j} \frac{\partial g}{\partial p_j} - \frac{\partial f}{\partial p_j} \frac{\partial g}{\partial q_j} \end{equation} where $q_j$ are the generalized coordinates and $p_j$ are the canonical momenta. Clearly: \begin{equation} \left\{q,p\right\}_{PB} = 1 \tag{1} \end{equation} In field theory, the Poisson bracket for two functionals $f$ and $g$ at equal times is defined as: \begin{equation} \left\{f(t),g(t)\right\}_{PB} = \int \mathrm{d}^3 \mathbf{x} \; \left(\frac{\delta f}{\delta \phi(\mathbf{x},t)} \frac{\delta g}{\delta \pi(\mathbf{x},t)} - \frac{\delta f}{\delta \pi(\mathbf{x},t)} \frac{\delta g}{\delta \phi(\mathbf{x},t)} \right) \end{equation} Now, using the rules of functional differentiation, it is easy to see that: \begin{equation} \left\{\phi(\mathbf{x},t),\pi(\mathbf{y},t)\right\}_{PB} = \delta^3(x-y) \end{equation} which is the classical field version of equation $(1)$.

Furthermore, according to Dirac's quantization rule, we can go back and forth between classical point particle mechanics and quantum mechanics via the following recipe: \begin{equation} \begin{array}{ccc} \text{classical mechanics} & \leftrightarrow & \text{quantum mechanics} \\ \displaystyle \left\{A,B\right\}_{PB} & \leftrightarrow & \displaystyle \frac{1}{i\hbar}\left[A,B \right] \end{array} \end{equation} provided the quantities we are considering exist in the classical world (for instance, quantum mechanical spin does not have a classical equivalent and so the rule does not work). To go back and forth between between classical field theory and the operatorial formulation of quantum field theory, we use the rule: \begin{equation} \begin{array}{ccc} \text{classical field theory} & \leftrightarrow & \text{quantum field theory} \\ \displaystyle \left\{A,B\right\}_{PB} & \leftrightarrow & \displaystyle \frac{1}{i\hbar}\left[A,B \right]_\mp \end{array} \end{equation} where the subscript $-$ means the normal commutator and is relevant for bosonic fields, and the $+$ subscript implies the anti-commutator which is relevant for fermionic fields (such as the Dirac field).

share|cite|improve this answer
Don't spinors have the peculiarity that we need to consider the Dirac Bracket? The Lagrangian is linear in velocity, after all...I seem to vaguely recall reading this in Henneaux and Teitelboim's Quantization of Gauge Systems... – Alex Nelson Mar 19 '14 at 20:40
Thank you! But can we get the "+" (anticommutator's) sign in the correspondense of Poisson and the Dirac bracket. May it be done as the result of the grassmanian nature of the spinors? – Andrew McAddams Mar 19 '14 at 20:40
@Hunter, don't get me wrong, your answer is a great overview for most fields. I wish I had your answer handy when I began learning field theory :) But classical spinors are wacky on a good day...I'll try to dig up my notes on this when time allows (it may take weeks), as I myself would be interested in re-learning this. In the meantime, I found this preprint: "Poisson Bracket for Fermion Fields" arXiv:1211.4231 – Alex Nelson Mar 19 '14 at 20:53
@Hunter : I intuitively think that the correspondence principle in a case of spinors already contains postulates which gives us anticommutation relations, but I still don't prove that. So I'll check it in the near future. [:)]. – Andrew McAddams Mar 19 '14 at 21:02
@Hunter : unfortunately, I was wrong: spinors component aren't grassmanian numbers (even in classical limit) as long as we don't use anticommutation relations. – Andrew McAddams Mar 20 '14 at 13:05

The classical Lagrangian for the free electron field is, $$ L=\int d^{3}x(i\psi^{\dagger}\frac{\partial \psi}{\partial t}+i\psi^{\dagger}\alpha_{r}\frac{\partial \psi}{\partial x^{r}}-m\psi^{\dagger}\beta \psi) \ . $$ The q's are $q^{i}(t)\rightarrow q^{(a,x)}\rightarrow \psi^{a}(t,x)$ and so the velocities are $\dot{q}^{i}(t)\rightarrow \frac{\partial \psi^{a}(t,x)}{\partial t}$. Varying the velocities, $$ \delta L= \int d^{3}x i\psi^{\dagger}\delta(\frac{\partial \psi}{\partial t}) $$ shows that the conjugate momenta - the $p_{i}(t)$ - are $i\psi^{\dagger}$. The classical Hamiltonian is, $$ H=\int d^{3}x(-i\psi^{\dagger}\alpha_{r}\frac{\partial \psi}{\partial x^{r}}+m\psi^{\dagger}\beta\psi) $$ The PBs corresponding to $[q^{i},p_{j}]_{PB}=\delta^{i}_{j}$ are, $$ [\psi^{a}(t,x),i\psi^{\dagger}_{b}(t,y)]_{PB}=\delta^{a}_{b}\delta(x-y) $$ Applying Dirac's quantization rule to the above PB gives the commutator, $$ [\psi^{a}(t,x),\psi^{\dagger}_{b}(t,y)]_{-}=\delta^{a}_{b}\delta(x-y) $$ and the $\psi$ are now operators. However, if one quantizes the classical Hamiltonian using this commutator there is no ground state, but one can quantize the classical Hamiltonian by using the anticommutator, $$ [\psi^{a}(t,x),\psi^{\dagger}_{b}(t,y)]_{+}=\delta^{a}_{b}\delta(x-y) $$ It appears that quantum dynamics of the electron field still goes with commutators. In the classical theory, $$ \frac{\partial \psi}{\partial t}=[\psi,H]_{PB} $$ and applying Dirac's quantization rule gives, $$ \frac{\partial \psi}{\partial t}=[\psi,H]_{-} $$ where $\psi$ and $H$ are operators. Everything looks consistent because writing down the quantum EoM directly from the quantum Hamiltonian (quantized using the anticommutator), $$ \frac{\partial \psi}{\partial t}=-i[\psi,H]_{-} $$ shows that the last two equations are identical. Note that the $H$ in the last equation is the quantized Hamiltonian which has been changed from the classical Hamiltonian by the use of the anticommutator, whilst the H in the penultimate equation is the classical Hamiltonian with all the fields simply promoted to operators. This is my understanding of the notion of the classical electron field, which comes from reading section 11 of Dirac's "Lectures on Quantum Field Theory."

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.