I often find myself typing equations into a node of a tikz picture. I then often find that I want to isolate certain characters in the equation and define them as a node.

For instance, consider the following figure.



  \node {$e^{i\pi}+1=0$};

enter image description here

I'd like it if in this figure I could define each character e, i, pi, +, 1, =, and 0 as its own node. Is this possible?


The \subnode command from tikzmark is designed for this sort of situation. It fakes a node around text that's inside a \node. It does need a few compilation runs to settle (it took 3 for this code for me) and you must remember the remember picture key on the tikzpicture.


\usetikzlibrary{tikzmark, arrows.meta}

\begin{tikzpicture}[>=Latex,remember picture]
  \node (eipi) at (0,0) {\(e^{i\pi}+1=0\)};
\node at (0,-1) {\(\subnode{e}{e}^{\subnode{i}{i}\subnode{pi}{\pi}}+\subnode{one}{1}=\subnode{zero}{0}\)};
\draw[<-] (e) to[out=-90, in=90] ++(-2,-1) node[below] {Base of natural logarithm};
\draw[<-] (i) to[out=135, in=0] ++(-2,1) node[left] {A square root of \(-1\)};
  \draw[<-] (pi) -- ++(0,-2) node[below] {Area of a unit circle};
  \draw[<-] (one) to[out=-45, in=180] ++(2,-1) node[right] {Unity};
\draw[<-] (zero) -- ++(2,0) node[right] {Zilch};
\draw[<-] (eipi) -- ++(0,1) node[above] {An ugly equation};

Annotated equation with subnodes

  • 1
    This has become my new workhorse! It's helping a lot with beamer slides that I'm writing for the course I'm teaching. I'm curious why you've started the math text in the nodes with \(. I was writing a document earlier that didn't work correctly when I used $ but then did work when I switched to \(. Any idea why? – Brian Fitzpatrick Feb 20 at 20:44
  • 1
    @BrianFitzpatrick I don't know of a reason why it wouldn't work with $. If you have an example then feel free to send it my way (open an issue on github.com/loopspace/tikzmark or email me directly - my email is in the tikzmark documentation). In general, though, \( is to be preferred to $, see tex.stackexchange.com/q/510/86 for a bit more on that. – Andrew Stacey Feb 21 at 19:57
  • I'll see if I can build a MWE. The document I was building was very complex so it's definitely possible there was something else going on. – Brian Fitzpatrick Feb 25 at 20:17
  • So I just tried to replicate my issue... and couldn't. However, I'm now almost positive that I was recycling a node name in another tikzfigure, which causes problems with remember picture. When I edited my original document I must have changed the node name in addition to switching the $ to \(. – Brian Fitzpatrick Mar 2 at 19:20
  • @BrianFitzpatrick Well, if it does resurface then do let me know. I'm pleased to hear you got it working. – Andrew Stacey Mar 3 at 8:59

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.