This post describes the article “Fremde Welten — Die Suche nach Exoplaneten mit Analogieexperimenten thematisieren” (Strange Worlds: Teaching Exoplanet Detection with Analogy Experiments), published in the teachers’ journal Unterricht Physik (“physics teaching”; issue 194, April 2023) with Alexander Küpper [[1]].

Literature, calculations, project links and mission descriptions checked through 2026-10-24; astro-lab.app showed a maintenance page that day. Classroom quotations and design observations come from the reported unit; the article does not establish population prevalence or learning effects.


Where Students Start

Before the unit, some participating students described a direct-vision model: you point a large telescope at a nearby star, and if there is a planet, you see it. “You could see them [the exoplanets] with a telescope/binoculars” and “You can see them with an extremely powerful telescope” are typical responses from year 8–9 students before they work through an actual detection unit.

This is not an unreasonable starting intuition. Telescopes see things far away. Planets are things far away. The inference seems to follow.

What it misses is the contrast ratio problem. A star is not just brighter than its planets — it is overwhelmingly, almost incomprehensibly brighter. The unit makes that point with an experiment; the numbers here are mine, not the unit’s. In reflected light, the ratio of the planet’s flux to the star’s is

$$C = \frac{F_p}{F_s} = A_g \left(\frac{R_p}{a}\right)^2 \Phi(\alpha) \tag{1}$$

with geometric albedo $A_g$, planet radius $R_p$, orbital distance $a$ and a phase function $\Phi$ that is 1 at full phase [[2]] and $1/\pi$ at quadrature, where the planet appears farthest from its star, for a diffusely reflecting (Lambert) sphere. Jupiter ($A_g = 0.538$, $a = 5.20$ AU [[3]], equatorial radius 71,492 km [[4]]) gives $C \approx 1.4 \times 10^{-9}$ at quadrature, roughly one part in a billion; an Earth analogue gives $2.5 \times 10^{-10}$. The fact sheets do not state the wavelength band of these albedos.

Against that glare and at a tiny angular separation, the planet is difficult to isolate. From 10 parsecs, Jupiter would sit at most about half an arcsecond from the Sun, and moving a planet inward to brighten it moves it into the glare (Fig. 1). Direct imaging of exoplanets is possible in special circumstances — young planets far from their stars, imaged in infrared — but for the vast majority of exoplanets, it is not a viable detection method. Of the 6,451 confirmed planets in the NASA Exoplanet Archive on 24 October 2026, 98 were discovered by imaging and 4,785 by transits [[5]].

Figure 1. In reflected light a planet cannot be both bright relative to its star and far from it: the contrast falls as $a^{-2}$, the separation grows as $a$. Eq. (1) at quadrature (Lambert sphere) around a Sun-like star, with the radii [[4]] and albedos [[3]] of Jupiter and Earth; top axis for a star 10 pc away. Hot giants should reflect far less [[2]]; thermal emission is not included.

The unit described in this article takes that misconception as its entry point and builds from there.


The Direct Imaging Experiment

The first experiment in the unit is a hands-on demonstration of why direct imaging is difficult.

The setup: a student points their smartphone camera at a small light source (a switched-on torch). Directly next to the torch, barely a few centimetres away, is a pin with a coloured head — the “exoplanet”. On the phone’s display, the pinhead is invisible. The torch (star) drowns it out completely.

Students can then investigate what would need to change for the pinhead to become visible. The answer they discover: block the torch with a small disc held in front of the camera at the right distance. With the direct glare suppressed, the illuminated pinhead becomes visible in the image.

This is an analogy for one coronagraphic idea: suppress on-axis starlight while retaining off-axis light. It does not reproduce diffraction, wavefront control, angular separation, detector noise or post-processing in instruments such as SPHERE on ESO’s Very Large Telescope [[6]], and the coronagraph on NASA’s Roman Space Telescope is a technology demonstration [[7]] rather than this tabletop arrangement scaled up.

The experiment also motivates a natural follow-on question: under what conditions does direct imaging work at all? Students can vary the pinhead distance from the torch and its size, exploring qualitatively the conditions under which the “exoplanet” becomes detectable even with partial suppression. The model motivates size and separation, while real surveys also depend on age, thermal emission, wavelength, distance, atmospheric models and instrument selection. Known directly imaged planets have therefore often been young, massive and widely separated; the pinhead demonstration does not derive that selection function.


The Transit Experiment

Once the limits of direct imaging are established, the unit introduces the transit method as the primary indirect technique. The pedagogical structure is deliberate: students have already understood that you cannot usually see exoplanets directly, which motivates the question of how else you might detect them.

The transit experiment uses a lamp as the star, a ball moved by hand (approximately periodically) around the lamp, and an Android smartphone running phyphox as the light sensor. When the ball crosses in front of the lamp from the sensor’s perspective, the measured illuminance dips. Students see a real light curve — not a simulation, not a graph from a database, but something they produced themselves from a physical measurement.

Two phyphox experiment files are provided for download (via QR code in the article and at astro-lab.app):

Basic experiment: records the raw illuminance data and displays the light curve. The focus is qualitative — what shape does the dip have? What determines the depth? What determines the period? Students can formulate the relationship between dip depth and planet-to-star size ratio as a qualitative rule (the larger the planet relative to the star, the deeper the dip) without necessarily working through the mathematics.

Extended experiment: adds real-time calculations of the transit depth, the maximum illuminance and the illuminance during the transit, the transit duration, and the orbital period. For students who are ready for it, this allows a quantitative derivation of the “planet” radius from the light curve. In symbols of my choosing, with $I_s$ the illuminance from the unobstructed lamp and $I_\mathrm{tr}$ the illuminance at the bottom of the dip, the transit depth is the relative drop in light,

$$\delta = \frac{I_s - I_\mathrm{tr}}{I_s} = k^2, \qquad k = \frac{R_p}{R_s} \tag{2}$$

where $R_p$ and $R_s$ are the radii of planet and star: an opaque planet entirely in front of a uniformly bright disc hides the fraction $k^2$ of it [[2]]. With the IAU nominal equatorial radii [[4]], Jupiter in front of the Sun gives $\delta \approx 1.06\%$ and Earth 84 parts per million.

This familiar approximation assumes an opaque planet, an isolated star with a uniform disc and a full transit (between second and third contact in Fig. 2). Limb darkening, grazing geometry, blended light, stellar variability and the apparatus geometry alter the depth. One more condition, my addition, matters in the classroom: the ball is nearer the sensor than the lamp, so the depth compares angular sizes, and $r_p \approx \sqrt{\delta}\, R_\mathrm{lamp}\, d_p / d_\mathrm{lamp}$ (small angles) needs the distances $d_p$ and $d_\mathrm{lamp}$ of ball and lamp from the sensor as well as the lamp radius.

The extended experiment also invites critical engagement with the model: the radius derived from the analogy experiment will differ from the actual ball radius because the lamp is not a uniform stellar disc and the sensor, ambient light, geometry and hand-driven orbit depart from the ideal model. Asking which assumption failed is the scientific point; scale alone is not a sufficient explanation.


Limits of the Transit Method

A recurring theme in the unit is that every detection method has limits, and understanding those limits is part of understanding the method.

For the transit method, the fundamental limit is inclination. A transit is only observable if the planet’s orbital plane is aligned (nearly edge-on) relative to our line of sight. The geometry and numbers that follow are mine, not the unit’s. On a circular orbit of radius $a$ and inclination $i$ (edge-on is $i = 90^\circ$), the planet passes its star at a projected distance $a \cos i$, the impact parameter $b = a \cos i / R_s$ in stellar radii, and blocks light only if

$$a \cos i < R_s + R_p, \qquad \text{that is,} \qquad b < 1 + k. \tag{3}$$

For randomly oriented orbits $\cos i$ is uniformly distributed, so a planet transits with probability $(R_s + R_p)/a \approx R_s/a$ [[2]]. For a Sun-like star $R_s/a$ is 0.47% at 1 AU, 9.3% at 0.05 AU (10.3% for a hot Jupiter with $R_p$ included) and 0.09% at 5.2 AU. Most exoplanetary systems, viewed from Earth, will not be aligned in this way. The transit method is therefore a biased sample: it detects only planets whose orbits we see nearly edge-on, and it misses most planets entirely. Because the probability falls as $1/a$, the bias also favours planets close to their stars.

Students can explore this experimentally: tilt the plane of the ball’s orbit away from edge-on and observe what happens to the light curve. The dip disappears. On a uniformly bright disc it first only shortens; depth and flat bottom last until the planet reaches the limb ($b > 1 - k$), then the dip turns shallow and rounded and vanishes at $b = 1 + k$ (Fig. 2). Gradual shallowing from the first tilt needs a limb-darkened star.

Figure 2. Tilting the orbit shortens a flat-bottomed transit, turns it into a shallow graze and removes it. A path dims the star only if its centre line crosses the highlighted circle of radius $1 + k$, Eq. (3). I–IV mark where the rims of planet and star touch; between II and III the depth is $\delta = k^2$, Eq. (2). Sky view cropped below the centre. Exact disc overlap, illustrative $k = 0.15$, uniform disc; for uniform motion the horizontal axis is proportional to time.

This connects naturally to a broader point about how astronomical surveys work. An occurrence rate counts planets, not detectable planets. Howard and colleagues counted each Kepler candidate (about 5% of them false positives, by their estimate) $a/R_s$ times, the inverse of its transit probability, against only the stars around which it could have been detected: $0.130 \pm 0.008$ planets per star with 2–4 Earth radii and periods under 50 days around GK dwarfs [[8]]. That is a mean per star; a fraction of stars with planets would also need the multiplicity and inclination distributions. In an earlier version of this post I called such a number the fraction of stars with “detectable planets”. That was wrong: the rate is corrected for the planets that do not transit and for those the survey could not detect.

The article includes a differentiation note: the limits investigation works well as an open inquiry task, with students formulating and testing their own hypotheses about what orbital configurations produce detectable transits.


Exoplanets as a Curriculum Bridge

One point the article makes explicitly is that exoplanets are not just an astronomy topic but a context that connects to multiple items in North Rhine-Westphalia’s physics core curriculum for the lower secondary years (Sekundarstufe I) of the academic-track secondary school (Gymnasium) [[9]]. That curriculum already has a content area for astronomy, “stars and the universe” (Sterne und Weltall); the cross-connections reach beyond it:

  • Optics: the seeing process (why does the star outshine the planet?), shadow formation, refraction in telescopes
  • Mechanics: orbital period, Kepler’s laws at a qualitative level, the habitable zone as a consequence of stellar luminosity and distance
  • Thermodynamics: planetary surface temperature, the greenhouse effect, albedo
  • Pressure: atmospheric pressure, habitability (a connection developed more fully in the Mission to Mars experiment)

The motivating question — could this planet host life? — was used to connect these topics. Whether it sustains engagement better than isolated teaching was not tested in the sources cited here.


What Comes After

The transit method is a productive entry point, but the search for extraterrestrial life does not end with planet detection. The article closes by noting that the detected exoplanets need to be analysed for habitability — which depends on orbital radius (habitable zone), stellar temperature, planet radius (mass is not available from transit data alone), atmospheric composition, albedo, and greenhouse effect.

Many of these can be connected back to physics experiment contexts, and the astro-lab project has developed smartphone-based analogy experiments for several of them. Detailed information on these is at astro-lab.app.

The transit analogy experiments have two companion papers [[10]], [[11]]; the second is retold in Hunting Exoplanets with Your Phone. For the full pedagogical sequence — from the original astro-lab student laboratory, through the COVID pivot to home experiments, to the return to school — see The Lab Goes Home. For the exomoon extension, which takes the transit experiment further into the question of moon-hosted life, see Can a Planet Have a Moon?.


References

[1] Küpper, A., & Spicker, S. J. (2023). Fremde Welten — Die Suche nach Exoplaneten mit Analogieexperimenten thematisieren [Strange worlds: teaching exoplanet detection with analogy experiments]. Unterricht Physik, 34(194), 33–36.

[2] Winn, J. N. (2010). Transits and occultations. In S. Seager (Ed.), Exoplanets. University of Arizona Press. arXiv:1001.2010

[3] Williams, D. R. Earth and Jupiter fact sheets. NASA Space Science Data Coordinated Archive. Opened 24 October 2026. https://nssdc.gsfc.nasa.gov/planetary/factsheet/earthfact.html; https://nssdc.gsfc.nasa.gov/planetary/factsheet/jupiterfact.html

[4] Prša, A., Harmanec, P., Torres, G., Mamajek, E., Asplund, M., et al. (2016). Nominal values for selected solar and planetary quantities: IAU 2015 Resolution B3. The Astronomical Journal, 152(2), 41. DOI: 10.3847/0004-6256/152/2/41

[5] NASA Exoplanet Science Institute (2020). Planetary Systems Composite Table. NASA Exoplanet Archive, IPAC. Queried 24 October 2026. DOI: 10.26133/NEA13

[6] ESO. SPHERE: Spectro-Polarimetric High-contrast Exoplanet REsearch instrument. Opened 24 October 2026. https://www.eso.org/public/teles-instr/paranal-observatory/vlt/vlt-instr/sphere/

[7] NASA. Roman Space Telescope: Coronagraph. Opened 24 October 2026. https://science.nasa.gov/mission/roman-space-telescope/coronagraph/

[8] Howard, A. W., Marcy, G. W., Bryson, S. T., Jenkins, J. M., Rowe, J. F., et al. (2012). Planet occurrence within 0.25 AU of solar-type stars from Kepler. The Astrophysical Journal Supplement Series, 201(2), 15. DOI: 10.1088/0067-0049/201/2/15

[9] MSB NRW (2019). Kernlehrplan für die Sekundarstufe I — Gymnasium in Nordrhein-Westfalen: Physik [Core curriculum for lower secondary Gymnasium in North Rhine-Westphalia: physics] (Schule in NRW, Heft 3411). Ministerium für Schule und Bildung NRW.

[10] Küpper, A., Spicker, S. J., & Schadschneider, A. (2022). Analogieexperimente zur Transitmethode für den Physik- und Astronomieunterricht in der Sekundarstufe I [Analogy experiments on the transit method for lower secondary physics and astronomy teaching]. Astronomie+Raumfahrt im Unterricht, 59(188), 46–50.

[11] Spicker, S. J., & Küpper, A. (2024). Exoplanet hunting in the classroom: An easy-to-implement experiment based on video-aided light curve analysis with smartphones. The Physics Teacher, 62(3), 226–229. DOI: 10.1119/5.0125305


Changelog

  • 2026-07-11: Recast the torch and pinhead as a limited coronagraph analogy; added the assumptions behind the transit-depth approximation; bounded student, direct-imaging and engagement claims to the reported unit and evidence.
  • 2025-10-03: Updated the DOI for Spicker & Küpper (2024) to the correct 10.1119/5.0125305.
  • 2026-10-03: Corrected the page range of the Unterricht Physik article to 33–36 (it begins on p. 33 of issue 194, not p. 4).
  • 2026-10-24: Corrected “X% of stars have detectable planets” to an occurrence rate in planets per star, corrected for orbits that do not transit and planets the survey could not detect (Howard et al. 2012). Replaced the depth $\Delta F$, described only as a transit depth, with $\delta$, defined as the relative drop in illuminance and equal to $k^2$ for a uniform disc. Added that a classroom radius estimate also needs the distances of ball and lamp from the sensor. Changed “the German physics curriculum” to North Rhine-Westphalia’s Gymnasium Kernlehrplan, which is the one cited. Removed the symbols $I_*$ and $I_\text{transit}$ from the description of the unit’s phyphox file, whose notation could not be checked. Added, as background beyond the unit, the reflected-light contrast, the transit condition and probabilities, dated archive counts, two figures and the Physics Teacher pages, and converted the references to numbered citations.