I'd like to typeset equations like: Inline Feynman diagrams 1 Inline Feynman diagrams 2

I have tried this with the tikz-feynman library, but the diagrams generated by it are just way too large, even with the small option (and also just look awkward). Optimally, I want to type simple diagrams even inline with the text, so that I can avoid awkwardly describing the diagram or using a lot of space and breaking the flow of the document to display the diagram.

  • 1
    Welcome to TeX.SE. I like very much your question. Good LaTeX.
    – Sebastiano
    Feb 23, 2019 at 10:44
  • 1
    Using normal inline TikZ by \tikz may be a solution.
    – user156344
    Feb 23, 2019 at 10:55

1 Answer 1


AFAIK you do not get bent arrows with tikz-feynman. And since you seem not to need the graph drawing algorithms (and since they cannot be uploaded to the arXv), you may just work with plain TikZ.

% from https://tex.stackexchange.com/a/430239/121799
\tikzset{% inspired by https://tex.stackexchange.com/a/316050/121799
    arc arrow/.style args={%
    to pos #1 with length #2}{
         mark=at position 0 with {\pgfextra{%
        mark=at position {#1-\tmpArrowTime} with {\coordinate(@1);},
        mark=at position {#1-2*\tmpArrowTime/3} with {\coordinate(@2);},
        mark=at position {#1-\tmpArrowTime/3} with {\coordinate(@3);},
        mark=at position {#1} with {\coordinate(@4);
        (@1) .. controls (@2) and (@3) .. (@4);},
fermion arc arrow/.style={arc arrow=to pos #1 with length 2.5mm},
Vertex/.style={fill,circle,inner sep=1.5pt},
insert vertex/.style={decoration={
         mark=at position #1 with {\node[Vertex]{};},

\[\mathscr{P}(\varphi)=-\sum\limits_{n=1}^\infty\tr\left(\Delta L_{12}\right)^n
 \draw[thick,insert vertex=0,fermion arc arrow={0.55}] (0,0) arc(270:-90:0.6);
 \draw[thick,insert vertex/.list={0,0.5}](0,0) arc(270:-90:0.6);
 \draw[fermion arc arrow/.list={0.3,0.8}] (0,0) arc(270:-90:0.6);
 \draw[thick,insert vertex/.list={0,1/3,2/3}](0,0) arc(270:-90:0.6);
 \draw[fermion arc arrow/.list={0.21,0.55,0.88}] (0,0) arc(270:-90:0.6);

 G(x_1,\dots x_n)=\sum\limits_{m=0}^\infty\frac{1}{m!}
 \node[circle,draw,thick,inner sep=2pt] (X) at (0,0) {$n+m$};
 \foreach \X in {60,90,120}
  {\draw[thick] (\X:0.6) -- (\X:0.9) node[Vertex]{};}
 \foreach \X in {-60,-80,-100,-120}
  {\draw[thick] (\X:0.6) -- (\X:0.9);} 
 \node[rotate=-30,overlay] at (-120:1.1){$x_1$};
 \node[rotate=30,overlay] at (-60:1.1){$x_n$};
 \node at (-90:1.1){$\cdots$};

enter image description here

  • Excuse me very much for this opinion. Can you reduce the size of the three circles? Perhaps the image of the series is more beautiful to see. PS: But what is the matter of this argument in Physics?
    – Sebastiano
    Feb 23, 2019 at 22:11
  • 1
    @Sebastiano You can control the appearance by adjusting Vertex/.style={fill,circle,inner sep=1.5pt},. I do not know what this diagram is. If the propagators were dashed, there would be a resemblance to the loops that one has to compute for the Coleman-Weinberg potential.
    – user121799
    Feb 23, 2019 at 22:15
  • I like very much your work. I not known the Coleman-Weinberg potential. Thank you very much. It will be between my favorities.
    – Sebastiano
    Feb 23, 2019 at 22:17
  • 1
    @Sebastiano To the best of my knowledge the term "dimensional transmutation" has been coined in the paper by Coleman and (Eric) Weinberg. It is a rather important observation. Whether or not the above has anything to do with it I do not know.
    – user121799
    Feb 23, 2019 at 22:22
  • 1
    @Sebastiano The second equation is just expressing the greenfunctions of a QFT in terms of a sum over all diagrams. The first is an intermediate step in a proof about the signs of feynman diagrams with fermionic fields. Both are taken from the excellent book "Functional methods in quantum field theory and statistical physics" by Vasiliev.
    – Leonard
    Feb 24, 2019 at 13:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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