🔬 optics

Ghost Imaging with Free Electron-Photon Pairs

This paper demonstrates coincidence-based ghost imaging of complex patterns with 2 $\mu$ m resolution by utilizing correlated electron-photon pairs generated via cathodoluminescence in a transmission electron microscope, thereby extending quantum-enhanced imaging techniques from photonic platforms to electron microscopy.

Original authors: Sergei Bogdanov, Alexander Preimesberger, Harsh Mishra, Dominik Hornof, Thomas Spielauer, Florian Thajer, Max Maurer, Pia Falb, Leo Stöger, Thomas Schachinger, Friedrich Bleicher, Michael S. Seifner

Published 2026-06-23

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Sergei Bogdanov, Alexander Preimesberger, Harsh Mishra, Dominik Hornof, Thomas Spielauer, Florian Thajer, Max Maurer, Pia Falb, Leo Stöger, Thomas Schachinger, Friedrich Bleicher, Michael S. Seifner, Isobel C. Bicket, Philipp Haslinger

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to take a picture of a secret object, but you have a strict rule: you are not allowed to look at the object directly. In fact, the camera you are using is completely blind to the object's shape.

This sounds impossible, right? But that is exactly what scientists in this paper did. They created a "ghost image" of a complex shape (a cat) using a clever trick involving two very different particles: an electron and a photon (a particle of light).

Here is how they pulled it off, using simple analogies:

The Two Messengers

Think of the experiment as a relay race with two runners who are "twins" in a quantum sense, even though they are very different:

The Electron (The Heavy Runner): This is a tiny, charged particle zooming through a powerful microscope. It is heavy, fast, and has a very short "wavelength," meaning it can see incredibly tiny details. However, in this experiment, the electron never actually looks at the shape of the object. It just flies through a silicon sheet.
The Photon (The Light Runner): When the electron hits the silicon sheet, it kicks out a flash of light (a photon). This photon is massless and easy to steer.

The "Ghost" Trick

Usually, to take a picture, you shine light on an object and catch the reflection. Here, they did something backwards:

The Setup: They placed a mask shaped like a cat in the path of the light (the photon), but not in the path of the electron.
The Interaction: The electron flies through the silicon and creates a photon. Because of the laws of physics, the electron and the photon are "entangled" or tightly linked. If the photon goes left, the electron goes right. If the photon goes up, the electron goes down. They are perfect mirrors of each other's movement.
The Filter: The light (photon) hits the cat mask. If the photon tries to go through the cat's ear, it gets blocked. If it tries to go through the eye, it passes through. The electron, however, flies straight through the silicon without ever seeing the cat mask.
The Detective Work: The scientists used a special camera to catch the electrons. But here's the catch: the camera only recorded an electron if its partner photon successfully passed through the cat mask.
- If the photon hit the "black" part of the mask (blocked), the electron was ignored.
- If the photon hit the "white" part of the mask (passed), the electron was saved.

The Result: A Ghost Cat

After collecting millions of these "saved" electrons, the scientists looked at where they landed. Even though no electron ever touched the cat mask, the pattern of where the electrons landed formed a perfect picture of the cat.

It's like if you had a room full of people (electrons) walking through a door, and you only let the people who were holding a specific ticket (photons that passed the mask) into a waiting room. If you looked at the pattern of people in the waiting room, you would see the shape of the ticket booth, even though the people never saw the booth.

Why This Is a Big Deal

Two Different Worlds: Usually, "ghost imaging" is done with two beams of light. This paper is the first to mix a heavy, charged electron with a light photon. It's like mixing a bowling ball with a ping-pong ball and making them dance together perfectly.
Sharpness: They managed to see details as small as 2 micrometers (about the width of a human hair divided by 50). This is sharp enough to see the cat's eyes, ears, and tail clearly.
No "Blur" from the Source: In other types of quantum imaging, the place where the particles are created can be "fuzzy," which blurs the picture. Here, because the light is created exactly where the electron hits the silicon, the "fuzziness" is much less of a problem, allowing for a sharper image.

The Bottom Line

The scientists proved that you can take a picture of a complex object using a particle that never actually touches the object, by relying on its "twin" partner that does. They successfully reconstructed a "ghost" image of a cat, showing that this strange quantum trick works even when mixing electrons and light. This opens the door to new ways of taking pictures with microscopes that are sharper and more sensitive than before.

Technical Summary: Ghost Imaging with Free Electron-Photon Pairs

Problem and Motivation
Coincidence imaging, or ghost imaging, is a technique that reconstructs information about a specimen by exploiting correlations between two particles. In this method, one particle interacts with the object but is not spatially resolved, while the other particle, which never interacts with the object, is spatially resolved. A "ghost image" is formed by matching the detection events of both particles. While this technique has been extensively implemented in photonic quantum optics and recently adapted for massive ion-ion and electron-ion pairs, its application to electron-photon pairs for two-dimensional complex imaging remains largely unexplored.

Current photonic quantum ghost imaging faces intrinsic limitations regarding spatial resolution, often constrained by the trade-offs in entanglement generation within nonlinear crystals (e.g., crystal thickness, phase-matching conditions, and pump beam geometry). Furthermore, previous electron-photon experiments in transmission electron microscopy (TEM) have primarily focused on probing joint uncertainty relations rather than complex image reconstruction. There is a need to adapt quantum-enhanced imaging techniques to the electron microscopy platform, leveraging the unique properties of electrons (atomic-scale resolution, massive, charged) and photons (ease of manipulation, efficient detection) to overcome classical imaging limits.

Methodology
The authors present an experimental setup within a transmission electron microscope (TEM) to perform coincidence imaging using electron–cathodoluminescence (CL) photon pairs.

Pair Generation: A 200 keV electron beam interacts with a 50 nm thick monocrystalline silicon membrane. This interaction generates correlated electron-CL photon pairs, predominantly via transition radiation. In this process, the electron and photon are tightly correlated in time, energy, and position.
Photon Arm (Object Interaction): A custom-built free-space optical system collects the emitted CL photons using a parabolic mirror mounted on a modified TEM sample holder. The photons are guided out of the TEM column through a viewport. An achromatic lens forms a magnified, spatially resolved image of the electron beam interaction point. A transmission mask (shaped like a cat or a calibration grating) is placed in this photon image plane to selectively block or transmit photons based on their position. The transmitted photons are focused onto a single-photon counting module (SPCM), which acts as a "bucket detector" (lacking spatial resolution but providing temporal timestamps).
Electron Arm (Spatial Resolution): Inside the TEM, transmitted electrons are energy-filtered using a Gatan GIF 2001 spectrometer to select only those electrons that have lost 2–3 eV of energy (corresponding to the emission of a visible photon). This step filters out background electrons that did not generate a photon. The filtered electrons are detected by a Timepix3 (TP3) direct electron detector, which provides both spatial resolution and temporal timestamps.
Coincidence Matching: A time tagger synchronizes the detection events of the SPCM and the TP3 camera. Electron-photon pairs are identified by their temporal correlation, specifically looking for time-of-arrival differences ( $\tau = t_e - t_\gamma$ ) within a $\pm 25$ ns window. The ghost image is reconstructed by correlating the spatial position of the detected electrons with the binary transmission state of the mask in the photon arm.

Key Contributions

First 2D Ghost Imaging with Electron-Photon Pairs: The work demonstrates the first two-dimensional ghost image of complex patterns formed from the spatial correlations between a free electron and its corresponding emitted photon.
Hybrid Platform: It establishes a hybrid architecture uniting massive, charged electrons with massless, neutral photons, leveraging the strengths of both platforms.
Resolution Benchmarking: The authors benchmark the system's performance using a standard calibration sample, achieving a spatial resolution down to 2 µm.
Alternative Pair Production Mechanism: Unlike photonic ghost imaging which relies on spontaneous parametric down-conversion (SPDC) in bulk crystals, this method utilizes interface-localized transition radiation. This avoids bulk phase-matching constraints and volume-induced limitations on spatial resolution.

Results

Cat Mask Imaging: The system successfully reconstructed a ghost image of a cat-shaped transmission mask. The raw electron counts showed a featureless beam profile, but the coincidence-filtered data clearly resolved the cat's features, including eyes, ears, and tail. The mask features (e.g., eyes $\sim 70 \times 50$ $\mu m^2$ ) were demagnified by a factor of $\sim 19$ onto the electron sample plane, resulting in features on the order of a few micrometers in the coincidence image.
Resolution Quantification: Using a calibration target with 60 $\mu m$ periodicity grating lines, the authors quantified the spatial resolution. By fitting the experimental data to an analytical model that accounted for optical distortions, they determined a full-width at half-maximum (FWHM) resolution of $2.03 \pm 0.06$ $\mu m$ at the TEM sample plane. This corresponds to a standard deviation of $\Delta x = 0.87 \pm 0.03$ $\mu m$ .
Performance Improvement: Compared to the authors' previous work on electron-photon entanglement (which had a resolution of $\Delta x \approx 1.45$ $\mu m$ ), this study shows an improvement of nearly 50% in spatial resolution, attributed to upgrades in the photon optical system that reduced aberrations and surface imperfections of the parabolic mirror.

Significance and Claims
The paper claims that this work establishes an alternative experimental framework to conventional photon-photon ghost imaging implementations. By utilizing electron-photon pairs, the technique overcomes the fundamental resolution limits imposed by SPDC crystal properties in photonic systems. The electron-photon platform does not suffer from analogous fundamental limitations on spatial resolution because the pair production (transition radiation) is localized to the interface, avoiding bulk phase-matching constraints.

The authors state that this approach opens the door for:

Programmable and Dynamic Imaging: The ability to shape electron wave functions via post-selection based on photon detection, potentially using adaptive masks (e.g., Digital Mirror Devices).
Quantum-Enhanced Sensing: The potential to exploit these states for discrete-variable protocols and to probe sample features with unprecedented precision.
Robustness: The inherent noise robustness of correlation-based imaging, which is valuable for low-signal or interaction-constrained environments.

The work is presented as a step toward future demonstrations of quantum entanglement (such as EPR and Bell states) in electron microscopy and the application of these states for advanced imaging techniques like wavefront shaping and low-dose imaging. The authors remain modest, noting that current resolution is limited by the performance of the photon arm and that future improvements in mirror design and detector technology (e.g., TP4 detectors) are expected to further enhance performance.

The Two Messengers

The "Ghost" Trick

The Result: A Ghost Cat

Why This Is a Big Deal

The Bottom Line

More like this