🔬 optics

Quantitative phase gradient microscopy with spatially entangled photons

This paper presents a novel entanglement-based quantitative phase gradient microscopy technique that utilizes spatially entangled photons to non-invasively recover full amplitude and phase profiles with high resolution and sensitivity, eliminating the need for interferometry, scanning, or complex algorithms while offering robustness against background noise.

Original authors: Yingwen Zhang, Paul-Antoine Moreau, Duncan England, Ebrahim Karimi, Benjamin Sussman

Published 2026-03-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yingwen Zhang, Paul-Antoine Moreau, Duncan England, Ebrahim Karimi, Benjamin Sussman

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 ghost. The ghost is invisible to the naked eye because it doesn't block light; it only slightly changes the speed at which light passes through it. In the world of microscopy, these are called "transparent samples" (like living cells). Traditional cameras see nothing but a blank white screen.

For decades, scientists have had to use complex tricks—like lasers, mirrors, and mathematical guessing games—to make these invisible ghosts visible. But a new team of researchers has invented a way to see them using quantum entanglement, and they did it without any of the usual heavy machinery.

Here is how their new technique, called Quantum Correlation Phase Gradient Microscopy (QCPGM), works, explained through simple analogies.

1. The Magic Twin Pair (Entanglement)

Imagine you have a pair of magical dice. You roll them, and no matter how far apart you take them, they always land on matching numbers. In physics, this is called entanglement.

In this experiment, the scientists create pairs of "twin" photons (particles of light).

Photon A (The Signal): This twin stays close to the sample. It acts like a tour guide who walks right next to the object.
Photon B (The Idler): This twin flies far away to a different camera. It acts like a remote observer who never touches the object but knows everything about it because of its connection to Photon A.

2. The "Ghost" Trick (Non-Local Sensing)

Usually, to see a transparent object, you have to shine light through it and measure how the light bends. But this new method is like a magic trick where you don't even need to look at the object directly.

The Setup: The "tour guide" photon (Photon A) passes right through the transparent sample. The sample is so clear that the photon doesn't change its brightness, but it does get slightly "pushed" or delayed (a change in phase).
The Remote Observer: The "remote observer" photon (Photon B) never touches the sample. However, because they are entangled twins, the moment the tour guide gets pushed, the remote observer instantly knows.
The Result: By measuring where the remote observer lands, the scientists can calculate exactly how the tour guide was pushed, even though the remote observer never saw the sample. It's like knowing how hard a door was pushed by feeling the vibration in the floorboards on the other side of the house.

3. Seeing the Invisible Gradient

Transparent objects don't just block light; they create a "slope" or a "gradient" in the light's path. Think of it like a gentle hill. If you roll a marble (a photon) over a hill, it changes direction slightly.

The Old Way: Traditional microscopes often need to take many pictures from different angles or use a grid of tiny lenses (like a honeycomb) to figure out the shape of the hill. This is slow, expensive, and can get blurry.
The New Way: This quantum method looks at the direction the remote observer photon is flying. If the tour guide was pushed to the right by the sample, the remote observer will be detected slightly to the left. By measuring this tiny shift in direction for millions of photon pairs, the computer can reconstruct the exact shape of the "hill" (the sample's phase) with incredible precision.

4. Why This is a Game-Changer

The paper highlights three major superpowers of this new technique:

It's a "Ghost" in the Machine (No Interferometry): Most high-tech microscopes need two beams of light to interfere with each other (like ripples in a pond meeting). This requires perfect stability; if a truck drives by outside, the vibration ruins the picture. This new method needs no interference. It works even if the lab is shaking.
It Ignores the Noise (Background Light): Imagine trying to hear a whisper in a crowded, noisy stadium. Most microscopes get drowned out by background light (like sunlight or room lights). Because this method only counts "twin" photons that arrive at the exact same split-second, it can ignore the noise. It's like having a secret code: "I only listen to people who say 'apple' at the exact same time as you say 'banana'." Everyone else is ignored.
It's Gentle (Low Power): The method is so sensitive it can see with a tiny amount of light (femtowatts). This is like using a single candle to read a book, whereas old methods might need a floodlight. This is perfect for looking at delicate living cells that would be burned or damaged by bright light.

The Bottom Line

The researchers successfully took pictures of cheek cells and tiny patterns, achieving a resolution sharp enough to see details as small as a human hair's width, all while using almost no light and ignoring background noise.

In short: They built a microscope that uses the spooky connection between quantum twins to "feel" the shape of invisible objects without ever touching them, without needing complex lasers, and without being bothered by the noise of the world around it. It's a giant leap toward seeing the invisible world with crystal clarity.

1. Problem Statement

Quantitative Phase Imaging (QPI) is essential for visualizing transparent biological samples and materials without staining. However, existing techniques face significant trade-offs:

Interferometric methods (e.g., holography) offer high accuracy but are highly sensitive to environmental noise and require stable reference beams.
Non-interferometric methods often rely on spatial scanning, microlens arrays (e.g., Shack-Hartmann sensors), or iterative phase-retrieval algorithms. These approaches suffer from limited spatial resolution, slow acquisition speeds, convergence instability, or hardware complexity.
Quantum imaging has shown promise in super-resolution and noise robustness, but previous quantum phase imaging demonstrations have struggled to match the spatial resolution and phase sensitivity of classical QPI microscopes while avoiding the limitations of interferometry.
Background Noise: Many imaging techniques struggle in dynamic or structured lighting environments (e.g., fluorescence in biological samples), where background light obscures phase information.

2. Methodology: Quantum Correlation Phase Gradient Microscopy (QCPGM)

The authors propose a novel, non-interferometric technique called Quantum Correlation Phase Gradient Microscopy (QCPGM). This method leverages the principles of quantum ghost imaging and ghost diffraction using spatially entangled photon pairs.

Photon Source: A Type-II Spontaneous Parametric Down-Conversion (SPDC) process in a periodically-poled potassium titanyl phosphate (ppKTP) crystal generates orthogonally polarized photon pairs (signal and idler) that are strongly correlated in position and anti-correlated in momentum.
Dual-Domain Measurement:
- The Signal photon illuminates the sample in the Near-Field (NF) (position domain) and is detected by a time-tagging camera.
- The Idler photon travels to the Far-Field (FF) (momentum domain) and is detected by a second camera.
- Crucially, the sample is illuminated by both photons of the entangled pair.
Nonlocal Inference: Due to position-momentum entanglement, measuring the position of the signal photon ( $r_s$ ) and the momentum of the idler photon ( $k_i$ ) in coincidence allows for the nonlocal inference of both conjugate observables simultaneously.
Phase Gradient Extraction:
- The phase gradient of the sample at a specific NF pixel ( $r_s$ ) causes a shift in the centroid of the idler photon distribution in the FF.
- The relationship is linear in the paraxial limit: $\langle \nabla \phi(r_s) \rangle \propto (U(r_s) - U_0(r_s))$ , where $U$ is the centroid position.
- By scanning the signal photon position (or using a camera to capture the whole NF), the system maps the full phase gradient field.
Phase Reconstruction: The full phase profile $\phi(r)$ is reconstructed by solving the 2D partial differential equations derived from the measured gradients using the Frankot and Chellappa method (a Fourier-based integration technique), which handles noisy, non-integrable gradient fields better than finite difference methods.
Background Mitigation: The system utilizes temporal coincidence detection (20 ns window) to filter out uncorrelated background photons. Furthermore, it employs a background subtraction algorithm by measuring "accidental coincidences" in a time window offset from the true coincidence peak (e.g., 50 ns earlier) to estimate and subtract dynamic background contributions.

3. Key Contributions

First Non-Interferometric Quantum QPI: Demonstrates a scanning-free, non-interferometric quantum phase imaging technique that recovers full amplitude and phase profiles without microlens arrays or iterative algorithms.
Record-Breaking Performance: Achieves the highest reported performance in quantum phase imaging to date, with a spatial resolution of 2.76 µm and a phase sensitivity of $\lambda/100$ .
Ultra-Low Power Operation: Successfully operates with illuminating powers as low as 100 femtowatts (fW), making it ideal for photosensitive samples.
Robustness to Dynamic Backgrounds: Demonstrates intrinsic resilience to structured and dynamic background light (simulating fluorescence) through coincidence detection and statistical background subtraction, outperforming classical phase gradient microscopes in noisy environments.
Theoretical Comparison: Provides a theoretical comparison showing QCPGM avoids the "centroid pixelization error" inherent to Shack-Hartmann sensors, offering superior precision at equivalent resolutions.

4. Experimental Results

Resolution Validation: The system imaged a 1951 USAF resolution target, resolving line pairs separated by 2.76 µm (Group 8-4).
Phase Sensitivity: Using quantitative phase targets with feature heights from 50 nm to 350 nm, the system achieved a phase sensitivity of $\lambda/100$ (approx. 0.06 rad uncertainty). The method remained accurate for phase jumps up to $\sim\lambda/6$ .
Biological Imaging: Successfully imaged cheek epithelial cells, recovering both amplitude and phase profiles. The phase images revealed cellular structures invisible in standard bright-field microscopy.
Background Suppression:
- In the presence of a dynamic background laser (Signal-to-Background Ratio $\approx$ 1.49), a classical phase gradient microscope yielded a Normalized Root Mean Squared Error (NRMSE) of 1.5.
- Using only coincidence detection (without background subtraction), the NRMSE worsened to 2.0 due to accidental coincidences.
- With the proposed background subtraction algorithm, the NRMSE improved to 0.59, nearly matching the ideal case (0.50) with no background.
Data Acquisition: Images were acquired in 500 seconds. The authors note that current limitations are due to camera technology (7% quantum efficiency, 8 ns time resolution) and that future superconducting nanowire cameras could reduce acquisition times to under one second.

5. Significance and Impact

This work represents a major step forward in quantum-enhanced sensing. By unifying ghost imaging and ghost diffraction, QCPGM overcomes the traditional trade-offs between resolution, sensitivity, and hardware complexity found in classical QPI.

Biomedical Applications: The ability to image at femtowatt power levels enables non-invasive observation of photosensitive biological samples (e.g., live cells, neurons) that would be damaged by conventional high-intensity illumination.
Adaptive Optics: The technique offers a pathway for high-resolution wavefront sensing in low-light or noisy environments, beneficial for astronomy and retinal imaging.
Complex Environments: The demonstrated robustness against dynamic background light opens new possibilities for imaging in uncontrolled lighting conditions or in the presence of autofluorescence.
Future Potential: The authors suggest that combining QCPGM with volumetric imaging techniques could lead to 3D quantum phase microscopy, and the use of higher-resolution time-tagging cameras could push spatial resolution toward the diffraction limit ( $\sim$ 1 µm).