AI-generated illustration for Galileo Whispers of a fictional multiply imaged quasar field. The four bright points are a conceptual lensing arrangement, not an observation; no arrival-time delay can be measured from this still image.
Lensed quasars can appear as several images because light from one distant source reaches Earth along different routes around a foreground mass. Those routes have different geometric lengths and different gravitational contributions to travel time. A change in the quasar’s brightness can therefore appear in one image before it appears in another.
The images are views of the same variable source at different emission times. Their relative delay can be measured by monitoring repeated features in their light curves, even though astronomers do not know when the quasar originally emitted the light in an absolute sense.
Multiple images follow different stationary paths
A quasar is a bright active galactic nucleus, compact enough that its central emission often appears point-like in an image. A foreground galaxy can deflect the background light strongly enough to produce two or more separate quasar images on the sky.
Those images correspond to different routes through spacetime, rather than several quasars created by the lens. In the geometric-optics description, an image forms along a path whose travel time is stationary with respect to nearby alternative paths. Stationary means that a small neighboring path change has no first-order effect on the arrival time; it does not mean that every image takes an equally short route.
Narayan and Bartelmann’s lensing lectures, section 3.3, explain this with a time-delay surface. The arrival-time values at its stationary points differ. Those differences are the delays between the corresponding images.
The surface is a mathematical representation of possible arrival times. It is not a physical sheet the light slides across, and the lens does not hold photons in storage before releasing them.
The delay has a geometric and a gravitational part
The geometric contribution accounts for the route’s extra length relative to the appropriate undeflected comparison. A bent route can take longer than the reference path even if light moves locally at its usual vacuum speed throughout.
The gravitational contribution depends on the lens’s gravitational potential along the route. It is related to Shapiro delay: comparing a journey through curved spacetime with the corresponding baseline involves more than a ruler measurement of path length divided by the speed of light.
Both contributions enter the total arrival-time prediction. Comparing two lensed images means taking the difference between their totals. The usual cosmological lens formula also includes the distance geometry and the lens redshift, which relates the timing to the observer’s clock.
This is why multiplying an observed delay by the speed of light does not, by itself, reveal a unique extra geometric distance. That product would express a time interval as an equivalent light-travel length; it would not separate the two terms or determine the lens mass.
A photograph of the image positions helps constrain the geometry, but an image farther from the center is not governed by a universal rule that its delay must be longer. The mass distribution and the complete arrival-time model decide the ordering.
A shared brightness pattern provides the clock
Astronomers measure a light curve by recording an image’s brightness over time. If the quasar changes intrinsically, related changes should appear in its other lensed images after their relative delays, with allowances for magnification and other effects.
Consider an invented monitoring example. Let observers identify three irregular brightness features in two images, called A and B. The day numbers are measured from an arbitrary start of monitoring; they are not dates from a real quasar campaign.
| Matched feature | Day seen in A | Day seen in B | B minus A |
|---|---|---|---|
| First broad peak | 10 | 22 | 12 days |
| Unequal second peak | 25 | 37 | 12 days |
| Later dip | 40 | 52 | 12 days |
All three features support a model in which B lags A by 12 days. Shifting B’s plotted timestamps earlier by twelve days aligns the pattern with A. The two images need not have equal absolute brightness: their magnifications can differ.
The example’s strength is the shared irregular pattern. One isolated bump is easier to match by accident, and a nearly periodic light curve can admit multiple candidate shifts. Real observations have gaps, finite sampling, measurement uncertainties and changes that do not repeat perfectly between images.
A twelve-day delay would not mean a telescope must expose for twelve days. It means that monitoring must cover the relevant changes in both images, with enough sampling and precision to distinguish plausible shifts.
Microlensing adds changes that do not copy the quasar
Stars in the foreground lens galaxy can change the magnification of a quasar image. As the source, stars and observer move relative to one another, this microlensing can introduce brightness changes specific to one image.
The measured curve therefore mixes intrinsic quasar variability with magnification effects and noise. Simply forcing every fluctuation to line up can bias a delay. Researchers need a method that distinguishes a common source pattern from image-specific changes and reports uncertainty in that separation.
The 2020 COSMOGRAIL monitoring paper describes automated curve shifting with explicit treatment of microlensing-model uncertainty. Its real monitoring record illustrates why delay measurement requires a time series and a statistical analysis, rather than a comparison of two photographs.
Different wavelengths can probe different emitting regions and complicate the relation further. A delay should be reported with its image pair, sign convention, observing data and uncertainties; “the quasar delay” is not a complete measurement description.
A measured delay needs a lens model to become a distance
Time-delay cosmography combines the measured interval with a model of how mass deflects the light. That can constrain a cosmological distance scale and, under stated cosmological assumptions, the expansion rate.
The 2018 H0LiCOW analysis of SDSS 1206+4332 combined relative timing with resolved host-galaxy imaging, the lens galaxy’s stellar velocity dispersion, and information about other matter along the line of sight. These observations constrain different parts of the inference.
Different mass models can reproduce some of the same lensing observables. Additional information helps limit those degeneracies; a precise time delay alone does not eliminate them or settle a disagreement about cosmic expansion.
The durable achievement is more specific: one variable quasar supplies a recognizable pattern, gravitational lensing gives that pattern several routes to Earth, and repeated monitoring measures the difference in arrival times. Turning those timings into cosmology requires the rest of the mass and geometry evidence.
