A solid-state entangled photon pair source with high brightness and indistinguishability

TL;DR

This paper reports a high-brightness, high-indistinguishability entangled photon pair source using GaAs quantum dots embedded in nanostructures, advancing quantum communication and computation capabilities.

Contribution

The authors demonstrate a deterministic quantum dot-based photon source with record brightness, fidelity, and indistinguishability, surpassing previous probabilistic methods.

Findings

Record pair collection probability of 0.65(4)

Entanglement fidelity of 0.88(2)

Photon indistinguishability over 0.9

Abstract

The generation of high-quality entangled photon pairs has been being a long-sought goal in modern quantum communication and computation. To date, the most widely-used entangled photon pairs are generated from spontaneous parametric downconversion, a process that is intrinsically probabilistic and thus relegated to a regime of low pair-generation rates. In contrast, semiconductor quantum dots can generate triggered entangled photon pairs via a cascaded radiative decay process, and do not suffer from any fundamental trade-off between source brightness and multi-pair generation. However, a source featuring simultaneously high photon-extraction efficiency, high-degree of entanglement fidelity and photon indistinguishability has not yet been reported. Here, we present an entangled photon pair source with high brightness and indistinguishability by deterministically embedding GaAs quantum dots in broadband photonic nanostructures that enable Purcell-enhanced emission. Our source produces entangled photon pairs with a record pair collection probability of up to 0.65(4) (single-photon extraction efficiency of 0.85(3)), entanglement fidelity of 0.88(2), and indistinguishabilities of 0.901(3) and 0.903(3), which immediately creates opportunities for advancing quantum photonic technologies.

Tables

Table 1. Table 1: Characterization of more devices from the same chip.

Device $\#$	Single-photon efficiency/Pair rate	Cavity wavelength (nm)	X wavelength (nm)	X Purcell factor	FSS ($\mu$eV)/f	X HOM	XX HOM
1	0.86/0.65	772.43	770.05	3.5	4.8/0.88	0.9	0.9
2	0.80	769.36	770.86	2.6	11.6/N.A.	N.A.	N.A.
3	0.76	767.39	777.51	3.1	18.0/N.A.	N.A.	N.A.
4	0.70/0.44	767.18	778.91	3.1	3.4/0.85	0.81	0.84
5	0.67	766.17	767.08	1.7	5.8/N.A.	N.A.	N.A.
6	0.66	763.86	765.08	2.1	11.0/N.A.	N.A.	N.A.
7	0.63	766.76	766.18	2.7	9.1/N.A.	N.A.	N.A.
8	0.61	767.20	770.20	3.4	5.1/N.A.	N.A.	N.A.
9	0.56	765.53	769.51	2.2	15.8/N.A.	N.A.	N.A.
10	0.55	764.24	766.62	2.0	6.6/N.A.	N.A.	N.A.

Table 2. Table 2: Comparison of the performance of our device to the state-of-the-art entangled sources.

	Pair efficiencya	Entanglement Fidelity	Indistinguishability
InAs QD in micropillar molecule (Ref Dousse2010, )	0.12	0.63	Not shown
InAsP QDs in nanowires (Ref Jons2017, )	0.0025	0.817	Not shown
InAs QDs in planar cavities (Ref Michler2014, )	0.0001	0.81	0.86
GaAs QDs in planar cavities (Ref Huber2017, )	0.0001	0.94	0.93
SPDC USTC (Ref Wang2016, )	$\sim$0.1	0.93	0.91
SPDC Vienna (Ref Giustina2015, )		$>$0.95	$>$0.9
SPDC Geneva (Ref Geneva2005, )		$\sim$0.9	$\sim$0.9
This work	0.65(4)	0.88(2)	$\sim$0.9

Equations

$$C_{\mu }=\frac{g_{XX,X}^{(2)}(0)-g_{XX,\bar{X}}^{(2)}(0)}{g_{XX,X}^{(2)}(0)+g_{XX,\bar{X}}^{(2)}(0)}$$

$$C_{linear}=0.92(2)$$

$$C_{diagonal}=0.81(2)$$

$$C_{circular}=-0.80(2)$$

$$f=\frac{1+C_{linear}+C_{diagonal}-C_{circular}}{4}=0.88(2)$$

Full text

Quantum entanglement is one of the most intriguing properties in quantum physicsEPR , in which the quantum state of a many-particle system cannot be written as a product of the single-particle wave functions, no matter how far they are separated from each other. Entangled photon pairs, which are immune from decoherence and are easy to manipulate and detect, have played an essential role in the epic triumph of quantum physics over local causality through optical tests of Bell’s inequalitiesGiustina2015 ; Shalm2015 . In the modern quantum technology era, entangled photon pairs serve as a key element in many quantum photonic information processing protocolsBouwmeester_book ; Kimble2008 , e.g., the quantum repeaterSimon2007 and device-independent quantum key distributionAcin2007 . To date, spontaneous parametric down conversion (SPDC)Kwiat1995 ; Geneva2005 ; Wang2016 is the most widely used ”working horse” for generating entangled-photon pairs with high degree of entanglement fidelity and photon indistingshabilityPan2012 . However, the Poissonian statistics of such sources intrinsically limits their brightness to an operation rate that is typically $<$ 0.1Pan2012 (the average photon pair generation probability per pulse), thus imposing a great challenge in advancing efficiency-demanding photonic quantum technologies.

Alternatively, epitaxial semiconductor quantum dots (QDs) have been successfully demonstrated as a potentially scalable technology for triggered sources of entangled photon pairs via the biexciton (XX) - exciton (X) cascaded radiative processesBenson2000 ; Stevenson2006 ; Akopian2006 ; Muller2009 ; Michler2014 ; Pelucchi2016 . Their small footprint and compatibility with semiconductor technology make them particular appealing for on-chip integrationOrieux2017 . However, a multitude of challenges have to be overcome to be able to realize optimal semiconductor sources of entangled photon pairs. First, the fine structure splitting (FSS) of the neutral exciton state, a result of the electron-hole exchange interaction in asymmetric QDs, reveals the radiative decay path information and consequently significantly reduces the time-averaged entanglement fidelity. This issue has been very recently alleviated by developing QDs with highly symmetric shapes, either by InGaAs growth in inverted pyramidsPelucchi2016 or by optimized droplet-etching Huo2013 . Furthermore, it has been argued that the comparatively short radiative lifetime of X, the small nuclear spin number of Ga (3/2 compare to 9/2 of In), and the use of two-photon excitation effectively suppress the spin-flip and carrier recapture processes, thus facilitating the achievement of a high-degree of entanglement fidelity and indistinguishabilityKeil2017 ; Huber2017 .

Second, low photon extraction efficiency, a result of the high refractive index of the semiconductor material surrounding the QDs, has long been recognized as a hurdle for quantum light sources based on QDs. Typically, only $<$1% of the photons emitted by QDs in bulk material can be collected by a free-space lens or objective. Photonic nanostructures, e.g, cavitiesDing2016 ; Somaschi2016 ; He2017 , waveguidesClaudon2010 ; Reimer2012 ; Laucht2012 ; Arcari2014 , microlensesGschrey2015 and circular Bragg gratingsDavanço2011 ; Sapienza2015 , exhibit excellent performance in funneling the single-photons emitted by QDs into free-space or optical fibers, but directly implementing these nanostructures for entangled-photon pair generation is not straightforward. The state-of-the-art QD entangled photon pair sources are based on micro-pillar ”molecules”Dousse2010 , photonic nanowiresJons2017 and optical antennasChen2018 , in which each single-photon in the pair efficiently couples into a dedicatedly designed photonic channels, resulting in bright-entangled photon pairs with a high degree of entanglement fidelity. Nevertheless, the overall performance of these entangled pair sources must be significantly improved for most applications, in terms of simultaneously achieving high brightness, entanglement fidelity, and photon indistinguishability.

Here we take a further step towards entangled photon pair sources with high brightness and indistinguishability by combining GaAs QDs with new broadband photonic nanostructures, i.e., circular Bragg resonators on highly-efficient broadband reflectors (CBR-HBR). Using a wide-field QD positioning techniqueSapienza2015 ; He2017 ; Liu2017_RSI , we deterministically fabricate CBR-HBRs in which single GaAs QDs (see S.I. I) are precisely located at the optimal position (the center of the cavity) for high-performance entangled photon pair generation. A single-photon collection efficiency of up to 0.85(3) for both X and XX is achieved, resulting in a record photon pair collection probability of 0.65(4) per excitation pulse. A high degree of single-photon purity of 99.8(1)%, entanglement fidelity of 0.88(2), and indistinguishabilities of 0.901(3) and 0.903(3) are also simultaneously obtained.

Design and fabrication of devices

In order to realize bright entangled photon pairs with Purcell-enhanced emission rates, we have developed a new nanostructure, i.e., CBR-HBR, with a few significant advantages respective to our previous workSapienza2015 . Most notably, the implementation of the HBR strucuture effectively suppresses the downwards photon leakage and therefore significantly improves the collection efficiency over a broadband, see the details in the S.I. II. Our CBR-HBR consists of a circular AlGaAs disk surrounded by a set of concentric AlGaAs rings, sitting on a $\mathrm{SiO_{2}}$ layer with a gold back reflector, as schematically shown in Fig. 1(a). The cavity resonance can be accurately engineered by varying the diameter of the central AlGaAs disk. Meanwhile, the in-plane emission is directed upwards by the concentric rings that meet the second-order Bragg conditions. By carefully designing the thickness of the $\mathrm{SiO_{2}}$ insulator layer, all the photons leaking into the substrate can be effectively reflected from the broadband gold mirror and recaptured by the CBR (See more details in S.I. III). In such a situation, very high collection efficiencies can be obtained in a broadband manner. For QDs located in the center of the CBR, the simulated collection efficiency at the first lens and the Purcell factor as a function of the operation wavelength are plotted in Fig. 1(b). Collection efficiencies above 90 % can be theoretically achieved in a bandwidth of $\approx$33 nm, and Purcell factors above 2 can be obtained for a bandwidth of $\approx$13 nm, which is 6.5 times the X-XX separation ($\approx$2 nm).

We have developed a membrane transfer technique to realize the $\mathrm{AlGaAs/SiO_{2}/Au}$ material platform from which the CBR-HBRs are fabricated, with the details provided in the S.I. IV. We note that the presented photonic design is fully compatible with state-of-the-art piezoelectric-based tuning methodsYan2016 ; Trotta2016 ; Huber2018 , which enable the elimination of the FSS and the tuning of photon energy because of the flexible choice of substrate (here quartz) and flat morphology, which allows efficient strain transfer. By taking advantage of our recently developed QD positioning technique, we are able to identify individual QDs and extract their spatial positions with respect to alignment marks with an uncertainty of $\approx$10 nmSapienza2015 ; Liu2017_RSI , see Fig. 1(c). The CBR-HBR is then deterministically fabricated around the target QD. Figure 1(d) shows the fluorescence image of our device after the CBR-HBR fabrication, in which the targeted single QD in Fig. 1(c) is accurately located in the center of the fabricated CBR-HBR.

I Single-photon emission and brightness assessment

Figure 2(a) presents the photoluminescence (PL) and white light reflectivity (1/R is shown, with R the reflectivity spectrum) of our device at 3.2 K, see the optical setup in S.I. V. Under a pulsed two-photon resonant excitation (TPE) schemeMichler2014 ; Keil2017 ; Huber2017 ; Huber2018 ; Jayakumar2014 , the intensities of the XX and X recombination are comparable, since TPE populates the XX state, which feeds the X state. The cavity mode, with a quality factor of $\approx$150, is clearly identified via the reflectivity measurement (see more details in S.I. VI) and it is resonant with both X and XX. The Purcell enhancement of the radiative decay of each state enabled by the cavity mode is directly quantified from time-resolved measurements in Fig. 2(b), showing comparisons of the lifetimes of X and XX in the CBR-HBR and in bulk (a different reference QD). We note the lifetimes of our reference QD in bulk are very consistent with the values reported in the similar systemsKeil2017 ; Huber2017 . The lifetime of X is shortened from 210 ps to 60 ps by implementing the CBR-HBR, corresponding to a Purcell factor ($F_{p}$) of 3.5. A slightly higher Purcell factor of 4.4 for XX is obtained, due to a better spectral match to the cavity mode, which enables faster triggering rates of entangled photon pair emission compared to those of the QDs in bulk.

Second-order auto-correlation measurements are performed for both X and XX, see Fig. 2(c). The nearly complete absence of coincidence events at zero time delay indicates the ultra-high purity of the emitted single-photons. $g^{(2)}_{X}(0)=0.001(1)$ and $g^{(2)}_{X}(0)=0.007(1)$ are obtained without background correction. The slightly higher $g^{(2)}_{XX}(0)$ value of XX is mostly due to the very weak emission from neighboring QD states.

Different from the single-photon Rabi oscillation in which the population of the excited state is dependent on the pulse area, the pulse area in the two-photon Rabi oscillation is replaced by an adiabatic dynamic phase that turns out to be a non-trivial function of the quantum dot binding energy, pulse area, pulse duration and pulse shapeStufler2006 . Fig. 2(d) shows the detected photon flux from X as a function of the square root of the time-averaged excitation power. The Rabi oscillations of X and XX (not shown) are observed due to the coherent control of the two-level system consisting of the biexciton and crystal-ground-state in the QD. For this device, the photon count rate reaches a maximum for an average laser power of 36 nW, which we denote as ”$\pi$ pulse” condition, similar to recent reportsMichler2014 ; Huber2017 . We note that the laser power needed to reach ”$\pi$ pulse” for the QDs in a CBR-HBR is at least 200 times lower than that for QDs in the bulk and also in simple planar cavities. Such a reduction of ”$\pi$ pulse” power is attributed to the cavity enhanced excitationKaniber2009 and represents an advantage for filtering the excitation laser in the entanglement and indistinguishability measurements we shall present.

For a ”$\pi$ pulse”, we observe a photon count rate up to 3.4(1) MHz under a 79 MHz repetition rate laser excitation. By taking the setup efficiency $\xi$ (7 %, see S.I. VII ), avalanche photodiode (APD) correction factor ($1.25$), and XX preparation fidelity $\eta_{XX}$ ($\approx$0.9) into account, a collection efficiency (with a 0.65 NA objective) $\eta$=85(3) % is extracted for both X and XX. Consequently, the collected photon pair probability per pulse $p=\eta_{XX}\times\eta^{2}\times[1-g^{(2)}_{X}(0)]^{1/2}\times[1-g^{(2)}_{XX}(0)]^{1/2}=0.65(4)$ is obtained. This high photon pair rate per pulse $p\approx 0.65$ outperforms any of the existing entangled photon sources reported in the literature.

Entanglement characterization

The states of the photon pairs emitted by QDs can be written as $\ket{\psi}=1/\sqrt{2}(\ket{H_{X}H_{XX}}+e^{is\tau/\hbar}\ket{V_{XX}V_{X}})$Michler2014 ; Keil2017 ; Huber2017 ; Shields2008 , in which $\tau$ is the, statistically varying, decay time of the XX state relative to the decay time of the X state and s is the value of the FSS. In absence of other dephasing mechanisms, the deviation of the two-photon states from the Bell state $\ket{\psi^{+}}$ originates from the phase factor $s\tau/\hbar$, which has to be minimized in order to obtain high level of entanglement without resorting to inefficient time-filteringWard2014 or spectra-filteringAkopian2006 . Since the X lifetime $\tau_{X}$ is as short as 60 ps in our device thanks to the Purcell effect (corresponding to a lifetime-limited linewidth of $\sim$11 $\mu$eV), we expect that the generation of photon pairs with high time-averaged fidelity is still possible for QDs with a finite FSS. Using polarization-dependent measurements, shown in Fig. 3(a), a FSS of 4.8(2) $\mu$eV for the QD in the CBR-HBR is extracted by subtracting the X transition from the XX transition energy. The theoretical values of fidelityHudson2007 as a function of FSS for different QDs with varied lifetimes are plotted in Fig. 3(b) (see the details in S.I. VIII). For GaAs QDs with a Purcell factor of 3.5, the entanglement fidelity decays slowly with the increase of the FSS. The predicted fidelity for the GaAs QDs with a FSS of 4.8 $\mu$eV is as high as 0.92 and it can still be above 0.75 for a FSS of 10 $\mu$eV. On the contrary, the entanglement fidelity for GaAs QDs in bulk (lifetime of 210 ps) decreases much more quickly with the increase of the FSS and shows a slightly lower value than the Purcell-enhanced source at FSS=0. With the same FSS of 4.8 $u$eV, the entanglement fidelity is only 0.64 for GaAs QDs in bulk. In order to compare the performance against a different material system, we also plot the entanglement fidelities of Purcell-enhanced InAs QDs ($F_{p}=3.5$) and InAs QDs in bulk (typical lifetime of 1000 ps). For the InAs QDs in bulk, the highest fidelity is $<$ 0.75 and the entanglement disappears once the FSS is larger than 1.6 $\mu$eV. Even with the same Purcell factor of 3.5, the entanglement fidelity of InAs QDs is still not reaching to the level of GaAs QDs in bulk. These results are based on the spin-scattering times provided in Ref. Keil2017, ; Huber2017, and need further experimental confirmations.

To evaluate the degree of entanglement of our bright photon-pair source, we perform cross-correlation measurements under $\pi$ pulse excitation for both X and XX photons in linear (HV), diagonal (DA), and circular (LR) basis sets. The cross-correlation histograms in the three basis sets are presented in Fig. 3(c-e). In linear and diagonal basis sets, we clearly observe the antibunching when the photon pairs are co-polarized and bunching for the cross-polarized photon pairs. The correlation in the circular basis is just opposite: co-polarized photon pairs show bunching while cross-polarized ones exhibit antibunching behavior. This set of correlations serves as a strong indication of polarization entanglement in the photon pairs. The degree of correlation in a particular polarization basis is defined byHudson2007 :

[TABLE]

where $g^{(2)}_{XX,X}(0)$ and $g^{(2)}_{XX,\bar{X}}(0)$ are the second order correlation for the co-polarized and cross-polarized photon pairs in that basis at zero delay time. The extracted degree of correlation in the different basis sets from measured coincidence histograms are:

[TABLE]

With these numbers, the entanglement fidelity for the polarization between the emitted photon pairs can be calculated as:

[TABLE]

The theoretically predicted entanglement fidelity of 0.92 can be reduced to 0.88 by using a much shorter spin scattering time (1 ns instead of 15 ns), which strongly indicates the existence of extra dephasing processes. Such extra dephasing processes have been also recently observed in a similar material system and is attributed to the interactions between the confined exciton and charge statesHuber2018 . In contrast to existing entangled sources with vanishing FSS, the pronounced Purcell effect in our work makes the high fidelity of entanglement possible for QDs with a comparatively large FSS. A near-unity entanglement fidelity can be expected in the future by implementing GaAs QDs with very small FSS in CBR-HBRs or by eliminating the residual FSS with a strain-tunable CBR-HBRHuber2018 (see strain-tunable CBR-HBR in the S.I. IX).

Photon indistinguishability

Photon indistinguishablity is a prerequisite for the realization of long-haul quantum information processing, e.g., a quantum repeater via entanglement swappingDuan2001 . We study the indistinguishablity of the emitted photons from our device via Hong-Ou-Mandel (HOM) interference measurementsSantori2002 . The QD is excited by two $\pi$-pulses separated by 1.9 ns with a repetition rate of 79 MHz (13 ns period). The emitted single photons are spectrally filtered within windows of  $\sim$100 $\mu$eV, much larger than the zero-phonon-linewidths of the transitions, and projected to the horizontal polarization before being coupled to an unbalanced Mach-Zehnder interferometer (MZI) equipped with a 1.9 ns delay. A half-wave plate is placed in one arm of the MZI to prepare co-polarized or cross-polarized photons, making them distinguishable or indistinguishable in polarization. The emitted photons are interfered at the beam splitter in co- and cross-polarized configurations. The coincidence histogram of HOM interference for both X and XX are shown in Fig. 4. For both X and XX, the coincidence events at zero delay are greatly suppressed in the co-polarized configuration (Fig. 4(a,b)), which indicates the occurrence of two-photon interference at the beam splitter. In contrast, the area of the central peaks are almost the same as the side peaks in the cross-polarized configuration, showing no indistinguishability (Fig. 4(c,d)). Visibilities of two-photon interference of 0.901(3) and 0.903(3) for X and XX are extracted from the areas of the central peaks in the correlation histogram. We note that the high degree of photon indistinguishability in our pair source is a direct result of the Purcell effect, which has been recently shown as a key element to realizing highly indistinguishable single-photons from InAs QDs in micropillar cavitiesDing2016 ; Somaschi2016 ; He2017 . With a joint force of further improvement of the Purcell factor, charge-stabilizationSomaschi2016 ; Liu2018 and rapid adiabatic passage excitationKaldewey2017 , higher photon indistinguishability can be expected as well as the entanglement fidelity. However, in our devices, the collection effiency tends to decrease with higher Q-factors (therefore higher Purcell factors) of the cavity due to the reduction of the overlap between the far-field pattern and the objective. Ultimately, the simultaneous realization of high collection efficiency and high indistinguishability will be fundamentally limited by the phonon scattering process. An upper bound can be placed on the indistinguishability by considering the microscopic theory developed in Ref. Smith2017, . This theory uses the polaron master equation formalism to capture non-Markovian phonon processes that lead to the emergence of a phonon sideband in the QD emission spectrum, and consequently degrades the indistinguishability of the source. Using this formalism with standard GaAs parameters, we find that our source could have an indistinguishability as high as 0.98 in the absence of any other dephasing processes (e.g. charge noise), see the S.I. X for details of the calculation. However, moderate filtering of the phonon sidebands (at the expense of a few percent count reduction) may be used to further boost the indistinguishability.

While we focus on the performance of a single exemplary device, we have characterized several other devices on the same chip. In table 1 we have listed 10 such devices in the order of the brightness, among which 2 devices with the smallest FSSs are fully characterized (see S.I. XI) and the others are partially characterized. The device 4 has smallest FSS of 3.4 $\mu$eV and the device 1 has the second smallest FSS of 4.8 $\mu$eV with the highest Purcell factor of 3.5. Therefore we do not expect any higher entanglement fidelity in the other devices. The limiting factor of the entanglement fidelity in this batch of devices is the relative large FSSs of the GaAs droplet QDs grown in the thin membrane (140 nm) structure with a thick sacrificial layer (500 nm), as shown in Fig. S1(a). We believe such a limitation can be soon overcome by either optimizing the epitaxial growth process or developing the strain-tunable CBR-HBR that we proposed.

Summary

Given the rapid development of the entangled photon sources both with SPDC and QD technologies, it is very insightful to directly compare the performance of our device to those of the existing sources reported in the literature. Table 2 lists the efficiency, entanglement fidelity, and indistinguishability of the state-of-the-art entangled photon sources together with our device (reference S.I. XII for methodology used in extracting the various parameters.). In general, the SPDC sources exhibit excellent performance in terms of entanglement fidelity and photon indistinguishability; however, their efficiencies are intrinsically limited to $<$0.1 due to the nature of the Poissonian statistics. Increasing the photon pair flux through higher excitation power inevitably adds extra noise and reduces the purity and indistinguishability. For the deterministic approach, the efficiency of QDs in bulk suffers greatly from the total internal reflection and only a few works show a high-degree of indistinguishability. InAsP QDs in nanowires and InAs QDs in micropillar molecules show much improved brightness and decent entanglement fidelity, but still the source efficiency and indistingshability have to be further improved. Our device, for the first time, simultaneously combines a high pair collection probability (0.65(4)), high degree of entanglement fidelity (0.88(2)) and photon indistinguishability (0.901(3) and 0.903(3)), and when taken together outperforms all the existing entangled photon pair sources.

To conclude, we have implemented a broadband photonic nanostrucuture, CBR-HBR, to harvest highly-entangled photon pairs emitted by GaAs QDs, obtained by droplet etching. By employing the QD positioning technique based on fluorescence imaging, the QDs are accurately placed in the center of the CBR-HBR, thus enabling the realization of entangled sources with record performances. Our devices may immediately find applications in both fundamental physics and applied quantum technologies, e.g., quantum random walk with entangled photon pairsPathak2007 , generation of hyper-entanglementPredojevic2017 and quantum repeatersSimon2007 associated with quantum memories. Moving forward, realizing high-performance photon pair sources operating in the telecom bandOlbricha2017 ; Huwer2017 is particularly appealing for long-haul quantum communication. Instead of polarization entanglement, time-bin entanglementsJayakumar2014 can be directly generated from QDs, which makes our devices compatible with the fiber network. The operation wavelength for both droplet QDs and photonic nanostructures can be shifted to the telecom band by changing the filling material of the nanohole and scaling the size of the nanostructures. To scale this technology up to multiple QDs, piezo-tuningYan2016 ; Trotta2016 or on-chip quantum frequency conversion technologiesLi2016 can be directly implemented in our devices to tune the QD emission wavelength, overcome the spectral inhomogeneity between different QDs and eliminate the FSS. Such identical entangled pair sources can serve as individual nodes interconnected via single-photon interference in the future quantum networkGao2013 . With the potential of scalability, our work serves as a landmark in the development of semiconductor quantum information processing chips and may boost new breakthroughs in quantum photonic technologies.