🔬 mesoscale physics

Energy-time entanglement from a monolithically integrated quantum dot on silicon

This paper demonstrates the generation of energy-time entangled photons from a single InGaAs/GaAs quantum dot monolithically grown on a silicon substrate, achieving coherent control via two-photon excitation and observing phase-dependent two-photon interference with visibilities approaching the threshold for Bell inequality violation.

Original authors: Marcel Hohn, Imad Limame, Peter Ludewig, Chirag C. Palekar, Aris Koulas-Simos, Kerstin Volz, Stephan Reitzenstein

Published 2026-07-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Marcel Hohn, Imad Limame, Peter Ludewig, Chirag C. Palekar, Aris Koulas-Simos, Kerstin Volz, Stephan Reitzenstein

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 build a super-advanced computer that uses light instead of electricity. To make this work, you need a special machine that can create pairs of "entangled" light particles (photons). These pairs are like magical twins: no matter how far apart they are, if you check one, you instantly know something about the other.

For a long time, scientists have had two problems with making these machines:

The Material Problem: The best light sources are made of a special material (like III-V semiconductors) that is very hard to glue onto the silicon chips used in our current computers. It's like trying to glue a delicate piece of glass onto a piece of wood without it cracking.
The "Twin" Problem: Usually, these light sources create twins based on their "color" (polarization). But if the material is slightly imperfect, the twins get confused, and the magic connection breaks.

This paper presents a solution to both problems. Here is the breakdown in simple terms:

1. Growing the "Magic Garden" on Silicon

The researchers grew a tiny, artificial "garden" of quantum dots (which act like the light factories) directly on top of a silicon chip.

The Analogy: Imagine trying to grow a tropical plant on a cold, rocky mountain. Usually, the plant dies because the soil is wrong. To solve this, the scientists built a special "buffer layer" (like a thick, warm blanket of GaP and other materials) between the silicon rock and the plant. This blanket smoothed out the roughness and stopped the plant from getting sick.
The Result: They successfully grew a high-quality light factory (an InGaAs/GaAs quantum dot) that is monolithically integrated—meaning it was grown as one single piece right on the silicon, not glued on later.

2. The "Perfect Dance" of Light

To get the entangled twins, they didn't just turn on a light switch. They used a very precise "two-step dance" called coherent two-photon excitation.

The Analogy: Think of the quantum dot as a dancer. To get them to perform a specific routine (the "biexciton-exciton cascade"), the scientists hit them with two laser photons at the exact same time. This is like a dance instructor tapping the dancer on both shoulders simultaneously to get them to spin perfectly.
The Proof: They saw the dancer spinning in perfect rhythm (Rabi oscillations), proving they had complete control over the process.

3. The "Magical Twins" (Energy-Time Entanglement)

Instead of checking if the twins have the same "color" (polarization), which is fragile, they checked if the twins arrived at the same time.

The Analogy: Imagine two runners starting a race. In a normal race, you check who is wearing the same shirt. In this experiment, the scientists checked if the runners started and finished at the exact same moment, even if they took different paths. This is called energy-time entanglement. It is much more robust because it doesn't care if the track is slightly uneven (which solves the "imperfect material" problem).

4. The "Long Hallway" Test

To prove the twins were truly connected, they sent them through a Franson interferometer.

The Analogy: Imagine a hallway with a short path and a very long path. The twins can take the short path together, the long path together, or one short and one long.
- If they are just normal light, the "mixed" paths (one short, one long) would show up clearly.
- If they are entangled, the universe "hides" the mixed paths, and you only see the "short-short" and "long-long" pairs interfering with each other.
The Result: The scientists saw this interference pattern. It was like hearing a perfect echo that proved the twins were linked. They measured a "visibility" (how clear the pattern was) of about 64% for a very short time window. This is high enough to prove the light is "quantum" (non-classical), though not quite high enough yet to break the ultimate "Bell inequality" record (which would be the gold standard).

5. Why the Pattern Wasn't Perfect

The paper admits the pattern wasn't 100% clear.

The Reason: It wasn't because the silicon integration failed. It was because of "background noise" and the timing of the experiment.
The Analogy: Imagine trying to hear a whisper in a quiet room. The whisper was there, but there was a slight hum of a refrigerator (background noise) and the room was a little too big (timing window). The scientists realized that if they could make the room smaller and turn off the fridge, the whisper would be crystal clear.

The Bottom Line

The paper claims that they have successfully built a silicon-compatible factory that can produce entangled light twins.

They proved that you can grow these delicate quantum structures directly on silicon without breaking them.
They proved that these structures can produce "energy-time" entangled pairs, which are tougher and more reliable than the usual "polarization" pairs.
While the signal isn't perfect yet, the "noise" is technical (fixable), not a fundamental flaw in the silicon technology.

This is a major step toward putting quantum computers on the same chips we use today, using a method that fits with existing manufacturing.

Technical Summary: Energy–Time Entanglement from a Monolithically Integrated Quantum Dot on Silicon

Problem Statement
Scalable quantum photonic technologies require deterministic sources of entangled photons that are compatible with established semiconductor manufacturing platforms. While self-assembled III–V semiconductor quantum dots (QDs) are leading candidates for on-demand entanglement generation, their integration with silicon-based architectures remains a significant challenge. Furthermore, while QDs are excellent sources of polarization-entangled photons, the achievable fidelity is fundamentally limited by the fine-structure splitting (FSS) of the exciton state, which arises from structural asymmetries and strain. Energy–time entanglement offers an alternative that is intrinsically insensitive to FSS; however, demonstrating this in a monolithically integrated III–V-on-silicon platform has not yet been achieved.

Methodology
The study utilizes InGaAs/GaAs quantum dots grown monolithically on a silicon (001) substrate. The heterostructure was fabricated via a two-step metal–organic chemical vapor deposition (MOCVD) process. The first step involved growing a GaP nucleation layer and dislocation-filtering layers (AlGaAs/GaAs and GaInP) on a Si:P buffer to create a high-quality virtual GaAs substrate. The second step involved depositing a bottom distributed Bragg reflector (DBR), a GaAs cavity, and a self-assembled In $_{0.33}$ Ga $_{0.67}$ As QD layer.

The experimental characterization involved:

Coherent Excitation: The QD was excited using a continuous-wave (CW) laser tuned to the biexciton (XX) transition energy via two-photon excitation (TPE). This approach drives the biexciton–exciton (XX–X) radiative cascade coherently.
Verification of Cascade: Power- and polarization-resolved micro-photoluminescence ( $\mu$ PL) were used to identify XX and X transitions and verify coherent driving through the observation of Rabi oscillations and dressed-state splitting. Time-resolved photoluminescence (TRPL) and second-order autocorrelation ( $g^{(2)}(\tau)$ ) measurements confirmed the cascaded nature of the emission.
Franson Interferometry: To characterize energy–time entanglement, the emitted photon pairs were analyzed using a four-channel Franson interferometer. The setup consisted of two unbalanced Michelson interferometers (path difference $\Delta L = 43$ cm, $\Delta T = 1.4$ ns) with active phase stabilization using a reference laser. Coincidence counts were recorded using superconducting nanowire single-photon detectors (SNSPDs) with a system temporal resolution of approximately 100 ps.

Key Results

Coherent Control: Under TPE, the system exhibited clear Rabi oscillations and dressed-state formation (splitting into $L_{+,0}$ and $R_{+,0}$ states), confirming coherent light–matter interaction. Cross-correlation measurements showed the characteristic antibunching at negative delays and bunching at positive delays, unambiguously verifying the XX–X cascade.
Coherence Properties: The radiative lifetimes were measured as $T_{1,XX} \approx 286$ ps and $T_{1,X} \approx 574$ ps. The coherence times were extracted as $T_{2,XX} = 128$ ps and $T_{2,X} = 91$ ps. The ratio $T_2 < 2T_1$ indicates the presence of pure dephasing.
Entanglement Verification: Franson interference measurements yielded phase-dependent two-photon interference.
- For an 80 ps integration window, the maximum observed visibility was $(64.0 \pm 7.0)\%$ .
- For a 1600 ps integration window, the visibility decreased to $(49.4 \pm 1.9)\%$ (specifically for the CH2 $\wedge$ CH4 channel combination, which showed the least imbalance).
- The visibility reduction with increasing window size was attributed to residual non-interfering contributions and background dynamics that are more complex in monolithically integrated samples compared to native GaAs substrates.

Significance and Claims
The paper establishes that monolithically integrated III–V quantum dots on silicon can serve as sources of energy–time entangled photons. The primary significance lies in demonstrating non-classical energy–time correlations in a silicon-compatible platform, which circumvents the fine-structure splitting limitations inherent to polarization entanglement.

The authors claim that the observed visibilities approach the threshold for Bell inequality violation at short time scales, though the current results remain below the strict threshold required for violation due to residual technical limitations (e.g., interferometer imbalances and background contributions) rather than fundamental constraints of the material platform. The work highlights that while the platform is not yet perfect, it preserves two-photon coherence throughout the radiative cascade. The authors conclude that this platform represents a promising route toward scalable, CMOS-compatible quantum photonic technologies, offering a pathway to on-chip quantum communication architectures that are resilient to FSS and capable of deterministic coupling to photonic circuits.