# How can I use the Annalen der Physik fonts on a more standard article class?

A friend of mine has a very nicely typeset paper on the arXiv and I would like to adapt large parts of the typesetting for my next paper. The paper uses the class file from Annalen der Physik, which is available under Manuscript submission > Article templates in their author guidelines page, or alternatively as andp2012.cls from e.g. here.

I tried to adapt the class to something I like a bit better (onecolumn with better margins, and a bit more room to breathe) and I sort of managed it, but the class is poorly documented and my alterations are a bodge that's going to come apart at the slightest push. (Also, to begin with, the class issues some unavoidable funky warnings right out of the box.) Instead of that, then, I would like to take the elements I like the most and work them into a more standard class like amsart.

This means, in particular, the fonts, which are nicely shaped and away from the serif impertinence of Computer Modern. (No offence, but I'm just deadly tired of it by now.) This should ideally be the lot: the fonts for the text, math, title, author, abstract, and section headings. What fonts or packages are responsible for this, and how can I get them working on amsart?

A sample file compiles into the following look:

The source is below; it needs andp2012.cls and picins.sty to run.

\documentclass{andp2012}% no class options needed by now
\usepackage[english]{babel}
\usepackage{lipsum}

\title{Article title}
\author{J. Doe}
\begin{abstract}
This is an abstract.
\end{abstract}
\shortabstract

\begin{document}
\maketitle

\section{Introduction}
Introduction text.

\section{Content}
\label{section1}
Some initial text, and some equations.
$$V(\vec{x}_A,\vec{x}_B)=d^2\frac{r^2-2\lambda^2}{(r^2+\lambda^2)^{5/2}},$$
being $d$ a letter, $\lambda$ a gathingammy $r$ a letter in $r=|\vec{x}_A-\vec{x}_B|=\sqrt{(x_A-x_B)^2+(y_A-y_B)^2}$, with  $A$, $B$  labels. Moreover $\Lambda=\lambda/a$, and  $\chi = a_{d}/\lambda = m_{eff}d^2/(\hbar^2 \lambda)$, with $m_\mathrm{eff}=\hbar^2/2ta^2$, and $t$, are more maths expressions. So are $k=\sqrt{k_x^2+k_y^2}$ and $V_{latt}(\vec r)= V_0\left(\sin^2(k_x x)+ \sin^2(k_y y)\right))$, and a displayed equation is
$$\left(\hat{T}_A+\hat{T}_B+{\hat V}(\vec{x}_A,\vec{x}_B)\right)\Phi(\vec{x}_A,\vec{x}_B)=E\Phi(\vec{x}_A,\vec{x}_B).$$
Other displayed equations are
$$(\vec{\xi}_{\vec{K}}\cdot\vec{\hat{T}}_D+V(\vec{r}))\psi(\vec{r})=E\psi(\vec{r}), \label{K-r}$$
where $\vec{\xi }_{\vec{K}}=-2t(\cos(K_x a/2),\cos(K_y a/2))$ and $\vec{\hat{I}}\cdot\vec{\hat{T}}_D\psi(\vec{r})=\sum_{i=x,y}\left(\psi(\vec{r}+\vec{\delta}_i)+\psi(\vec{r}-\vec{\delta}_i)\right)$, where $\vec{\delta}_i=a\hat{e}_{i}$, and also
$$\psi(\vec{r})=\frac{1}{N_x N_y}\sum_{\vec{q}}\psi(\vec{q})e^{i\vec{q}\cdot\vec{r}}$$
and
$$E_{\vec{K},\vec{q}}=-4t\left(\cos(K_xa/2)\cos(q_xa)+\cos(K_ya/2)\cos(q_ya)\right)$$
and
$$(E-E_{\vec{K},\vec{q}})\psi(\vec{q})=\sum_{\vec{q'}}V(\vec{q}-\vec{q'})\psi(\vec{q'}).$$
Then you do some blah blah blah and you finish the paper.

\lipsum[1-3]

\end{document}

• Why do they include not only the Word stuff in the same zip, but multiple copies of the LaTeX stuff as well? – cfr Feb 1 '17 at 1:30
• @cfr I know, I was miffed as well. – E.P. Feb 1 '17 at 7:12

The publication version of the code uses a mix of standard and custom fonts. You can't use the journal's custom fonts as you don't have either the fonts or the package to support them. You can, however, use the same base of standard fonts, perhaps tweaking later if required.

\RequirePackage{amssymb,upref}%


\usepackage{amssymb,upref}


This is the 'final' condition

\if@final
\if@fourier
\RequirePackage[expert]{fourier}%
\else%@fourier
\RequirePackage[utopia,expert]{mathdesign}%
\fi%@fourier
\RequirePackage[oldstyle,proportional,bold,regular,scaled=0.92]{profilepro}%


since we can't use this lot, we'll ignore it. Instead, we'll focus on the fonts the class uses to approximate the look of published papers.

\else%@final
\RequirePackage{fourier}%


So,

\usepackage{fourier}


For sans,

  \RequirePackage[scaled=0.86]{helvet}%


so,

\useapckage[scaled=0.96]{helvet}


but consider trying

\usepackage[scaled=0.96]{tgheros}


instead. Bernard's suggestion of erewhon is also a good one.

We also find

\RequirePackage{latexsym,textcomp}%


so I would try

\usepackage[T1]{fontenc}
\usepackage{textcomp}% add latexsym if you need it


The class also includes

\AtEndOfClass{%
\RequirePackage{microtype}%
}%


While this isn't exactly about the fonts, it will make a difference to how the fonts are used and handled and will improve the overall quality of the typesetting, so add

\usepackage{microtype}


Off-topic word to the wise: do not be tempted to emulate the journal's use of tabu. I see that the journal does at least provide its own archived copy, so it is probably safe itself, but you are unlikely to enjoy the typesetting self-sufficiency of a journal in its splendid isolation, so don't be tempted to copy its unfortunate example.

• Thanks, this was the most useful. Here is the finished product if you're interested. – E.P. Feb 2 '17 at 15:18

When you open the pdf file in a suitable pdf viewer, you see the fonts are fourier (based on Adobe Utopia), plus some Mathematica fonts. I suggest, as far as fonts are concerned, to stick to fourier, and add, for text fonts, erewhon, a Utopia clone which brings real small caps, old style digits, superior digits and a few other features.

The file andp2012.cls contains the lines

\if@final
\if@fourier
\RequirePackage[expert]{fourier}%
\else%@fourier
\RequirePackage[utopia,expert]{mathdesign}%
\fi%@fourier
\RequirePackage[oldstyle,proportional,bold,regular,scaled=0.92]{profilepro}%
\else%@final
\RequirePackage{fourier}%
\RequirePackage[scaled=0.86]{helvet}%
\AtEndOfClass{%
\providecommand{\lgseries}{\mdseries}%
\providecommand{\textlg}{\textmd}%
\providecommand{\mbseries}{\bfseries}%
\providecommand{\textmb}{\textbf}%
\providecommand{\sbseries}{\bfseries}%
\providecommand{\textsb}{\textbf}%
\providecommand{\ebseries}{\bfseries}%
\providecommand{\texteb}{\textbf}%
\providecommand{\lnfigures}{\relax}%
\providecommand{\txfigures}{\relax}%
\providecommand{\tbfigures}{\relax}%
\providecommand{\prfigures}{\relax}%
}%
\fi%@final


Try this

\documentclass{minimal}
\usepackage{lipsum}
\RequirePackage{fourier}%
\RequirePackage[scaled=0.86]{helvet}%
\begin{document}
Some initial text, and some equations.
$$V(\vec{x}_A,\vec{x}_B)=d^2\frac{r^2-2\lambda^2}{(r^2+\lambda^2)^{5/2}},$$
being $d$ a letter, $\lambda$ a gathingammy $r$ a letter in $r=|\vec{x}_A-\vec{x}_B|=\sqrt{(x_A-x_B)^2+(y_A-y_B)^2}$, with  $A$, $B$  labels. Moreover $\Lambda=\lambda/a$, and  $\chi = a_{d}/\lambda = m_{eff}d^2/(\hbar^2 \lambda)$, with $m_\mathrm{eff}=\hbar^2/2ta^2$, and $t$, are more maths expressions. So are $k=\sqrt{k_x^2+k_y^2}$ and $V_{latt}(\vec r)= V_0\left(\sin^2(k_x x)+ \sin^2(k_y y)\right))$, and a displayed equation is
$$\left(\hat{T}_A+\hat{T}_B+{\hat V}(\vec{x}_A,\vec{x}_B)\right)\Phi(\vec{x}_A,\vec{x}_B)=E\Phi(\vec{x}_A,\vec{x}_B).$$
Other displayed equations are
$$(\vec{\xi}_{\vec{K}}\cdot\vec{\hat{T}}_D+V(\vec{r}))\psi(\vec{r})=E\psi(\vec{r}), \label{K-r}$$
where $\vec{\xi }_{\vec{K}}=-2t(\cos(K_x a/2),\cos(K_y a/2))$ and $\vec{\hat{I}}\cdot\vec{\hat{T}}_D\psi(\vec{r})=\sum_{i=x,y}\left(\psi(\vec{r}+\vec{\delta}_i)+\psi(\vec{r}-\vec{\delta}_i)\right)$, where $\vec{\delta}_i=a\hat{e}_{i}$, and also
$$\psi(\vec{r})=\frac{1}{N_x N_y}\sum_{\vec{q}}\psi(\vec{q})e^{i\vec{q}\cdot\vec{r}}$$
and
$E_{\vec{K},\vec{q}}=-4t\left(\cos(K_xa/2)\cos(q_xa)+\cos(K_ya/2)\cos(q_ya)\right)$
and
$$(E-E_{\vec{K},\vec{q}})\psi(\vec{q})=\sum_{\vec{q'}}V(\vec{q}-\vec{q'})\psi(\vec{q'}).$$
Then you do some blah blah blah and you finish the paper.

\lipsum[1-3]
\end{document}