---
title: "Arrival time, the Fermat potential and the delay of an isothermal lens"
description: "Build the arrival-time surface of a galaxy lens, find its images where the surface is flat and derive the delay of a singular isothermal sphere."
canonical_url: "https://sandboxphysics.com/stories/quasar-time-delays/arrival-time-and-the-fermat-potential/"
language: "en"
page_type: "story"
alternate_languages:
  zh-CN: "https://sandboxphysics.com/zh/stories/quasar-time-delays/arrival-time-and-the-fermat-potential/"
last_modified: "2026-11-14"
---

[Lensed quasar time delays and the Hubble constant](https://sandboxphysics.com/stories/quasar-time-delays/)

# Arrival time, the Fermat potential and the delay of an isothermal lens

Sandbox Physics · 2026-11-14

Every point on the sky near a lens has an arrival time: how long light from the quasar would take if it were forced through that point. The images sit where this arrival-time surface is flat. For the simplest galaxy model, the singular isothermal sphere, the two images are one Einstein radius either side of the source and their delay is the time-delay distance over $c$ times $2\theta_{\rm E}\beta$.

Before this chapter, read [how a delayed quasar flare measures the size of the universe](https://sandboxphysics.com/stories/quasar-time-delays/one-flare-many-paths/). You need vectors in a plane, the small-angle approximation and the idea of a partial derivative. The gradient is introduced below.

## Coordinates, angles and distances

Put the centre of the lens galaxy at the origin of the sky plane. A point on the sky is a two-dimensional vector $\boldsymbol\theta$, measured in arcseconds from the lens centre. The quasar's true direction, where we would see it without the galaxy, is $\boldsymbol\beta$. One arcsecond is $4.848\times10^{-6}$ radians, so every angle here is tiny and $\sin\theta\approx\tan\theta\approx\theta$.

Three distances appear: to the lens, $D_{\rm d}$; to the source, $D_{\rm s}$; and from lens to source, $D_{\rm ds}$. They are angular-diameter distances, defined so that an object of size $\ell$ at distance $D$ spans an angle $\ell/D$. In an expanding universe $D_{\rm s}\neq D_{\rm d}+D_{\rm ds}$, which is why all three are needed. The lens redshift is $z_{\rm d}$.

## The arrival-time surface

Light forced through the point $\boldsymbol\theta$ is late for two reasons. The bent path is longer than the straight one, and the light spends time in the galaxy's gravity, where clocks run slow. Both effects are collected in the Fermat potential,

$$
\phi(\boldsymbol\theta,\boldsymbol\beta)=\frac{(\boldsymbol\theta-\boldsymbol\beta)^2}{2}-\psi(\boldsymbol\theta),
$$

measured in square arcseconds. The first term is the extra geometric length. The second is the gravitational delay, written with the lensing potential $\psi$: a projected, rescaled version of the galaxy's gravitational potential. Two facts about $\psi$ are all we need. Its gradient is the deflection angle, and half its Laplacian is the convergence,

$$
\boldsymbol\alpha=\nabla\psi,\qquad \kappa=\tfrac12\nabla^2\psi .
$$

The convergence $\kappa$ is the galaxy's surface density in units of a critical density; where $\kappa$ is near 1, the galaxy is dense enough to make multiple images.

The time between two points A and B on the surface is

$$
\Delta t_{\rm AB}=\frac{D_{\Delta t}}{c}\bigl[\phi(\boldsymbol\theta_{\rm A},\boldsymbol\beta)-\phi(\boldsymbol\theta_{\rm B},\boldsymbol\beta)\bigr],\qquad D_{\Delta t}=(1+z_{\rm d})\frac{D_{\rm d}D_{\rm s}}{D_{\rm ds}} .
$$

The factor $1+z_{\rm d}$ appears because the delay builds up at the lens, and the expansion stretches it by that factor on its way to us. These are Eqs. (4) to (7) of TDCOSMO IV; Kochanek writes the same surface in his Eqs. (1) and (2). [TDCOSMO IV, Sec. 2.1](https://arxiv.org/abs/2007.02941v3) [Kochanek 2002, Sec. 2](https://arxiv.org/abs/astro-ph/0205319v1)

## Images sit where the surface is flat

The gradient of a function $f(x,y)$ is the vector of its partial derivatives, $\nabla f=(\partial f/\partial x,\ \partial f/\partial y)$. It points uphill, and its length is the steepness. At the bottom of a bowl, at the top of a hill and at the centre of a saddle, the gradient is zero.

Fermat's principle says light travels along paths whose time does not change to first order when the path is nudged. So images form where $\nabla\phi=\mathbf 0$. Differentiating the Fermat potential term by term gives $\boldsymbol\theta-\boldsymbol\beta-\nabla\psi=\mathbf 0$, which is the lens equation

$$
\boldsymbol\beta=\boldsymbol\theta-\boldsymbol\alpha(\boldsymbol\theta).
$$

Each solution is one image. A minimum of $\phi$ arrives first and looks like the source, only brighter. A saddle point arrives later and is mirror-flipped. A maximum sits near the dense centre of the galaxy and is usually too faint to see; the Lab does not track it.

How bright an image is follows from how the lens stretches a small patch of sky. The magnification is $\mu=1/\det(\partial\boldsymbol\beta/\partial\boldsymbol\theta)$. A negative $\mu$ marks a flipped image.

## The singular isothermal sphere

Many lens galaxies are close to a singular isothermal sphere: a round galaxy whose stars move at the same speed at every radius. Its convergence falls as $1/|\boldsymbol\theta|$,

$$
\kappa=\frac{\theta_{\rm E}}{2|\boldsymbol\theta|},\qquad \psi=\theta_{\rm E}|\boldsymbol\theta|,\qquad \boldsymbol\alpha=\theta_{\rm E}\,\frac{\boldsymbol\theta}{|\boldsymbol\theta|} .
$$

You can check the first relation: in polar coordinates $\tfrac12\nabla^2(\theta_{\rm E}r)=\theta_{\rm E}/(2r)$. Every ray is bent by the same angle $\theta_{\rm E}$, the Einstein radius, towards the centre.

Put the source on the positive $x$ axis at distance $\beta<\theta_{\rm E}$. On that axis the lens equation reads $\beta=\theta-\theta_{\rm E}\,{\rm sign}(\theta)$. It has two solutions:

$$
\theta_{\rm A}=\beta+\theta_{\rm E},\qquad \theta_{\rm B}=\beta-\theta_{\rm E} .
$$

Image A is outside the Einstein ring on the source's side. Image B is inside the ring on the far side. Their separation is always $2\theta_{\rm E}$, whatever $\beta$ is.

For this lens $\det(\partial\boldsymbol\beta/\partial\boldsymbol\theta)=1-\theta_{\rm E}/|\boldsymbol\theta|$, so $\mu=|\boldsymbol\theta|/(|\boldsymbol\theta|-\theta_{\rm E})$ and

$$
\mu_{\rm A}=\frac{\theta_{\rm E}+\beta}{\beta},\qquad \mu_{\rm B}=-\frac{\theta_{\rm E}-\beta}{\beta} .
$$

Image B is flipped and fainter, and both images get brighter as the source approaches the centre.

## The delay of the isothermal lens

Evaluate the Fermat potential at each image. At A, $\phi_{\rm A}=\tfrac12\theta_{\rm E}^2-\theta_{\rm E}(\theta_{\rm E}+\beta)$. At B, where $|\theta_{\rm B}|=\theta_{\rm E}-\beta$, $\phi_{\rm B}=\tfrac12\theta_{\rm E}^2-\theta_{\rm E}(\theta_{\rm E}-\beta)$. The geometric terms are equal and cancel, leaving

$$
\phi_{\rm B}-\phi_{\rm A}=2\theta_{\rm E}\beta,\qquad \Delta t=\frac{D_{\Delta t}}{c}\,2\theta_{\rm E}\beta .
$$

Image A, the minimum, arrives first. The delay grows linearly with how far the source sits from the centre.

**Worked example: the distance factor.** Take $z_{\rm d}=0.5$, $z_{\rm s}=2$ in a flat universe with $H_0=70$ km s$^{-1}$ Mpc$^{-1}$ and matter density $\Omega_{\rm m}=0.3$. Integrating the expansion history gives $D_{\rm d}=1259$ Mpc, $D_{\rm s}=1727$ Mpc and $D_{\rm ds}=1097$ Mpc. Then $D_{\Delta t}=1.5\times1259\times1727/1097=2972$ Mpc. Divide by $c$ and convert one square arcsecond, $(4.848\times10^{-6})^2$ radians squared: one square arcsecond of Fermat potential is worth $83.2$ days.

**Worked example: the delay.** For $\theta_{\rm E}=1''$ and $\beta=0.1''$, $\phi_{\rm B}-\phi_{\rm A}=0.2$ square arcseconds and $\Delta t=0.2\times83.2=16.6$ days. The images are at $1.1''$ and $-0.9''$, with $\mu_{\rm A}=11$ and $\mu_{\rm B}=-9$. The leading image is brighter by a factor $11/9$.

**How the distances are computed.** In a flat universe the comoving distance to redshift $z$ is $D_{\rm C}=(c/H_0)\int_0^z{\rm d}z'/E(z')$ with $E(z)=\sqrt{\Omega_{\rm m}(1+z)^3+1-\Omega_{\rm m}}$. Then $D_{\rm d}=D_{\rm C}(z_{\rm d})/(1+z_{\rm d})$, $D_{\rm s}=D_{\rm C}(z_{\rm s})/(1+z_{\rm s})$ and $D_{\rm ds}=[D_{\rm C}(z_{\rm s})-D_{\rm C}(z_{\rm d})]/(1+z_{\rm s})$. Every distance carries the factor $c/H_0$, so $D_{\Delta t}\propto1/H_0$.

## Lab tasks

1. Open [the quasar time-delay Lab](https://sandboxphysics.com/labs/gravitational-lens-telescope/?mode=time-delay), set the mass of the nearby galaxy to 0 and keep the steepness at 2. Drag the quasar's true direction about a tenth of an arcsecond from the centre. Predict the delay and the image separation, then read them.
2. Double the source offset and check that the delay doubles while the separation stays about $2''$.
3. Open the redshift settings and move the lens from $z_{\rm d}=0.5$ to $0.3$. Predict how the delay changes using $D_{\Delta t}=1552$ Mpc at $z_{\rm d}=0.3$. The Lab keeps the Einstein radius fixed when you change a redshift.

## Exercises

1. Show that $\nabla\phi=\mathbf 0$ gives the lens equation, starting from the definition of $\phi$.
2. A lens has $\theta_{\rm E}=1.5''$ and the source is at $\beta=0.2''$, with $D_{\Delta t}=2972$ Mpc. Find the delay.
3. For $\theta_{\rm E}=1''$ and $\beta=0.3''$, find the flux ratio of the two images.
4. An observer measures 16.6 days for the lens of the worked example but assumes $H_0=75$. What delay would that observer have predicted, and what does the mismatch say?
5. A second source behind the same lens lies at $\beta=0.5''$ but at $z_{\rm s}=3$. Which parts of the delay formula change, and which stay the same?

Answers

1. $\nabla\tfrac12(\boldsymbol\theta-\boldsymbol\beta)^2=\boldsymbol\theta-\boldsymbol\beta$ and $\nabla\psi=\boldsymbol\alpha$, so $\nabla\phi=\boldsymbol\theta-\boldsymbol\beta-\boldsymbol\alpha=\mathbf 0$, which rearranges to $\boldsymbol\beta=\boldsymbol\theta-\boldsymbol\alpha$.
2. $\phi_{\rm B}-\phi_{\rm A}=2\times1.5\times0.2=0.6$ square arcseconds, so $\Delta t=0.6\times83.2=49.9$ days.
3. $\mu_{\rm A}=1.3/0.3=4.33$ and $|\mu_{\rm B}|=0.7/0.3=2.33$, a ratio of $1.86$. The ratio $(\theta_{\rm E}+\beta)/(\theta_{\rm E}-\beta)$ depends only on $\beta/\theta_{\rm E}$.
4. $D_{\Delta t}\propto1/H_0$, so the prediction is $16.6\times70/75=15.5$ days. The measured delay is longer, so with this mass model the data prefer $H_0=70$. The conclusion is only as good as the mass model, which the next two chapters question.
5. $D_{\Delta t}$ changes because $D_{\rm s}$ and $D_{\rm ds}$ change. The Einstein radius also changes, because it depends on $D_{\rm ds}/D_{\rm s}$. The formula $\Delta t=(D_{\Delta t}/c)\,2\theta_{\rm E}\beta$ keeps its form.

Next: [mass slope, shear and the time-delay distance](https://sandboxphysics.com/stories/quasar-time-delays/mass-slope-shear-and-the-time-delay-distance/).

## Try it in the Lab

[Quasar time delays in the gravitational-lens Lab](https://sandboxphysics.com/labs/gravitational-lens-telescope/?mode=time-delay): Switch off the nearby galaxy and check that the delay grows in step with the source offset.

## Sources

1. [TDCOSMO IV, Sec. 2.1](https://arxiv.org/abs/2007.02941v3)
2. [Kochanek 2002, Sec. 2](https://arxiv.org/abs/astro-ph/0205319v1)

## Keep reading

Lensed quasar time delays and the Hubble constant · Reading path

### From a delayed flare to the Hubble constant

1. [Article **How a delayed quasar flare measures the size of the universe**](https://sandboxphysics.com/stories/quasar-time-delays/one-flare-many-paths/)
2. [You are here **Arrival time, the Fermat potential and the delay of an isothermal lens**](https://sandboxphysics.com/stories/quasar-time-delays/arrival-time-and-the-fermat-potential/)
3. [Article **Mass slope, shear and the time-delay distance**](https://sandboxphysics.com/stories/quasar-time-delays/mass-slope-shear-and-the-time-delay-distance/)
4. [Article **The mass-sheet degeneracy and what stellar speeds add**](https://sandboxphysics.com/stories/quasar-time-delays/mass-sheet-degeneracy-and-stellar-speeds/)
5. [Article **Recompute the Hubble constant from seven TDCOSMO lenses**](https://sandboxphysics.com/stories/quasar-time-delays/recompute-h0-from-a-frozen-tdcosmo-slice/)
6. [Article **How to read TDCOSMO IV on the mass sheet and the Hubble constant**](https://sandboxphysics.com/stories/quasar-time-delays/reading-tdcosmo-iv/)

1 more group

### Other articles that use the same experiments

Each experiment from this article that another series also uses, followed by the articles there.

1. [Lab **Gravitational Lens Telescope**](https://sandboxphysics.com/labs/gravitational-lens-telescope/)
2. [Article **How a dark planet makes a distant star brighten**](https://sandboxphysics.com/stories/microlensing-planets/a-dark-planet-brightens-a-star/)
3. [Article **From a bent light ray to two stellar images**](https://sandboxphysics.com/stories/microlensing-planets/from-bent-light-to-two-images/)
4. [Article **What happens when a stellar disk crosses a caustic?**](https://sandboxphysics.com/stories/microlensing-planets/a-stellar-disk-crosses-a-caustic/)
5. [Article **What can a microlensing light curve tell us about its lens?**](https://sandboxphysics.com/stories/microlensing-planets/read-a-light-curve-without-weighing-too-soon/)
6. [Article **Read a dark planet’s mass from Earth and space**](https://sandboxphysics.com/stories/microlensing-planets/read-ground-space-and-roman-evidence/)
7. [Article **Fit the published OGLE-2005-BLG-390 photometry**](https://sandboxphysics.com/stories/microlensing-planets/refit-the-published-ogle390-files/)
8. [Article **A background star helped weigh an isolated black hole**](https://sandboxphysics.com/stories/astronomy-gravity/isolated-black-hole-microlensing/)
