QUANTUM OPTICS

Coherent single-photon emissionfrom colloidal lead halide perovskitequantum dotsHendrik Utzat1, Weiwei Sun1, Alexander E. K. Kaplan1, Franziska Krieg2,3,Matthias Ginterseder1, Boris Spokoyny1, Nathan D. Klein1, Katherine E. Shulenberger1,Collin F. Perkinson1, Maksym V. Kovalenko2,3, Moungi G. Bawendi1*

Chemically made colloidal semiconductor quantum dots have long been proposedas scalable and color-tunable single emitters in quantum optics, but they have typicallysuffered from prohibitively incoherent emission. We now demonstrate that individualcolloidal lead halide perovskite quantum dots (PQDs) display highly efficientsingle-photon emission with optical coherence times as long as 80 picoseconds, anappreciable fraction of their 210-picosecond radiative lifetimes. These measurementssuggest that PQDs should be explored as building blocks in sources of indistinguishablesingle photons and entangled photon pairs. Our results present a starting point forthe rational design of lead halide perovskite–based quantum emitters that havefast emission, wide spectral tunability, and scalable production and that benefit fromthe hybrid integration with nanophotonic components that has been demonstratedfor colloidal materials.

Many proposed schemes of quantum in-formation processing require scalablequantum emitters (QEs) capable of pro-ducing indistinguishable single pho-tons or entangled photon pairs (1, 2).

To realize such QEs, the optical coherencetime of the emitter (T2) needs to approach twicethe spontaneous emission lifetime (2T1). De-sign of solid-state QEs with transform-limitedphoton coherence (T2 = 2T1) is fundamentallyhampered by the exciton-bath interactionleading to optical decoherence due to phonon-scattering (3) and spin-noise (4) often result-ing in T2 ≪ 2T1: In addition, charge densityfluctuations in the environment typically causejumping of the spectral line of the emitter,leading to further decoherence on time scalesof micro- to milliseconds (5). The consequenceis that only a few practical QEs have beendemonstrated. Atom-like defects in diamond(6, 7) and self-assembled III-V quantum dots(QDs) are most commonly studied, often inte-grated with optical microcavities to increasethe degree of coherence in the Purcell regime(8). There has also been some proof-of-conceptwork with single molecules (9). The growthand integration of III-V QDs with high pho-ton coherence requires the highest possiblematerial quality and exceptional control overgrowth conditions. Self-assembled QDs alsosuffer from low scalability and typically ran-

dom growth that complicates deterministicintegration with microphotonic components,shortcomings that are further aggravated bydot-to-dot inhomogeneities (2). For defectsin diamond, the bulky host matrix presents achallenge for coupling to cavities and hampersefficient photon out-coupling, limiting theirbrightness. Conversely, chemically synthesizedcolloidal semiconductor quantum dots (CQDs)exhibit unmatched ease of processability fromsolution and straightforward hybrid integra-tion of single CQD emitters with various pre-fabricated microphotonic components has beendemonstrated. However, CQDs typically sufferfrom incoherent and unstable emission, whichhas prevented their application in quantumoptics (1).We demonstrate that a specific combination

of fast radiative lifetimes and long optical co-herence times gives rise to highly coherentsingle-photon emission from individual cesiumlead halide CsPbX3 (X = Cl, Br, I) perovskitequantum dots (PQDs) (10, 11). Our results sug-gest that—unlike any other colloidal quantumdot material—perovskite-based quantum dotscan be explored as low-cost, scalable QEs withfacile cavity integration to generate wavelength-tunable sources of indistinguishable single pho-tons and entangled photon pairs in the visiblespectral region.PQDs combine the advantages of chemical

synthesis in large batches and precise controlover the size and shape inherent to colloidalmaterials with the extensive compositionaltunability of lead halide perovskites. At roomtemperature, they display narrowband emissionacross the entire visible spectrum and near-unity emission quantum yields owing to thenotable defect tolerance of lead halide pe-

rovskites. The CsPbBr3 PQDs used in this studywere synthesized according to (11) and are sta-bilized with zwitterionic ligands with func-tional amino and sulfonate groups, which havedemonstrated increased stability compared toconventional PQDs that have oleylamine andoleic acid surface ligands (10, 11). Detailed in-formation on the synthetic procedure can befound in (12). PQDs exhibit an orthorhombic(Pnma) crystal structure consisting of corner-sharing [PbBr6]

− octahedra with Cs+ ions fill-ing the interoctahedral voids (Fig. 1A). Weconfirm this structure by x-ray diffraction (fig.S2). High-resolution scanning transmissionelectron micrographs (Fig. 1, B and C) confirmthe high degree of size uniformity and a cubicquantum dot shape as reported previously (11).The PQDs in our study have an average edgelength of 13.5 ± 2 nm (fig. S1). Their ensembleabsorption and emission spectra are presentedin Fig. 1D. The room temperature absorptionedge at 2.42 eV exhibits an excitonic featureconfirming quantum confinement. The emissionpeak centered around 2.38 eV has a full width athalf-maximum (FWHM) of ~90 meV, close tothe room-temperature single-particle emissionlinewidth, confirming the high synthetic qual-ity of our samples (13).The emission of PQDs occurs from weakly

confined excitons. Owing to strong spin-orbitcoupling in lead halide perovskites, spin andorbital angular momenta are strongly mixed,and only the total angular momentum J is con-served. The exciton in PQDs is formed from ahole with s-like symmetry in the valence bandand a twofold degenerate electron in the con-duction band with total angular momentumJh/e = 1/2 (14), where h is hole and e is electron.The electron-hole exchange interaction liftsthe degeneracy between the singlet (Jexciton = 0)and triplet (Jexciton = 1) exciton states (Fig. 1E).Owing to a strong Rashba effect, the degen-eracy of different angular momentum projec-tion states jjj ¼ T1; 0 is additionally lifted as aresult of inversion symmetry breaking in theorthorhombic crystal structure, leading to en-ergetic splitting of the triplet excitonic finestructure with values we define as W1 and W2

(15). It has recently been shown that the lowest-lying triplet state in PQDs is optically bright,whereas the singlet is dark, which is a distinctfeature of lead halide perovskite semiconduc-tors (16).We spin-coated dilute solutions of PQDs on

quartz to perform single-emitter characteriza-tion at low temperatures. Characterization spec-tra of individual PQDs with either one or twosharp emission peaks are shown in Fig. 2, A toC, and Fig. 2D, respectively. The insets show thedegree of emission peak polarization by plot-ting relative transmission intensities througha linear polarizer as a function of polarizerangle. The observed linear polarization forall PQDs confirms emission from the excitonstate, because the trion emission is nonpolar-ized (16). The resolution limit of our spectrom-eter (~500 meV), and potentially fast spectral

RESEARCH

Utzat et al., Science 363, 1068–1072 (2019) 8 March 2019 1 of 5

1Department of Chemistry, Massachusetts Institute ofTechnology, 77 Massachusetts Avenue, Cambridge, MA02139, USA. 2Institute of Inorganic Chemistry, Department ofChemistry and Applied Bioscience, ETH Zurich, 8093 Zurich,Switzerland. 3Laboratory for Thin Films and Photovoltaics,Empa−Swiss Federal Laboratories for Materials Science andTechnology, CH-8600 Dübendorf, Switzerland.*Corresponding author. Email: mgb@mit.edu

on Novem

ber 5, 2020

http://science.sciencemag.org/

Dow

nloaded from

diffusion, limits access to the true homogeneouslinewidth and to the fine structure for PQDswith small energetic splitting between excitonsublevels W1(W2). Figure 2E shows a typicalspectral and intensity time series of a singlePQD at 3.6 K under nonresonant pulsed exci-tation (480 nm and 50 W/cm2), confirming theabsence of major fluorescence intermittency orlarge spectral jumps, in stark contrast to estab-lished II-VI CQDs, which suffer from blinkingand large charge-induced spectral jumps at lowtemperatures (17, 18). We note that some PQDsshow near-Poissonian emission statistics overthe course of minutes (fig. S5).A single PQD intensity correlation [g(2)(t)]

under pulsed excitation is shown in the upperpanel of Fig. 2F. Our PQDs show high biexcitonemission quantum yields at low temperatures, asexpected from their very high radiative rates thatcan outcompete any nonradiative Auger proc-ess. Their superior spectral stability comparedwith other CQDs allows spectral selection of theexciton zero-phonon line (ZPL) and rejection ofbiexciton emission, analogous to single-photonsources based on self-assembled QDs (2). Whenthe emission is isolated with tunable dichroicedgepass filters, PQDs exhibit strong antibunch-ing [g(2)(0) < 0.04], indicating high single-photonpurity of the emission (Fig. 2F, lower panel). Allstudied individual PQDs show fast photolu-minescence decay that can be fit with the sumof two exponentials (Fig. 2G). The fast com-ponent (~210 to 270 ps) is about two orders ofmagnitude higher in intensity than the long tail,confirming their fast emission rates comparedwith any other single-photon emitters owingto the bright nature of the lowest-lying excitonground state and a giant oscillator strengtheffect (18, 19). As the emission quantum yield

of the PQDs used in this study is ~95% (see fig.S9), the fast observed photoluminescence (PL)lifetime is close to the spontaneous radiativelifetime T1 (16). The transform-limited linewidth,calculated as Grad = ħ/T1, is 3.1, 2.4, 2.5, and 2.4meV for PQDs 1 to 4, respectively, where ħ isPlanck’s constant h divided by 2p.We measure the optical coherence time and

resolve the fine-structure splitting with photon-correlation Fourier spectroscopy (PCFS) (20, 21),which can overcome both the temporal and spec-tral resolution limitations of other techniques byencoding the spectral coherence of a single emit-ter in intensity anticorrelations recorded at theoutput of a Michelson interferometer (Fig. 3A).We note that PCFS and Hong-Ou-Mandel spec-troscopy, a commonly used two-photon inter-ference method, both provide optical coherencetimes of single emitters on fast time scales, butPCFS allows easier extraction of spectral diffu-sion dynamics, ZPL fraction, and fine-structuresplitting (21). Unlike conventional Fourier spec-troscopy, in which the interferogram is resolvedby collecting a sufficient number of photons ateach interferometer position (typically hundredsof milliseconds integration time), PCFS corre-lates photon pairs as a function of their temporalseparation t at each interferometer path lengthdifference d. The temporal resolution of PCFS isonly determined by the photon shot noise at agiven t, which enables measurement of the op-tical coherence on time scales inaccessible tomost other techniques. The observable is the PCFSinterferogram G(2)(d, t) which intuitively corre-sponds to the envelope of the squared interfer-ogram compiled from photon pairs separated byt [Fig. 3A and notes in (12)]. The Fourier trans-form of G(2)(d, t) yields the spectral correlationp(z, t) defined as pðz; tÞ ¼ h∫∞�∞sðw; tÞsðwþ

z; t þ tÞ dwi, where h…i denotes the time av-erage, s(w, t) represents the spectrum of thesingle emitter over frequencies w and at timet, and z represents the energy difference. Thespectral correlation is the sum of the autocor-relations of the spectra compiled from photonpairs separated by t. Because the probabilityof spectral wandering vanishes as t approacheszero, the spectral correlation at small t re-duces to the autocorrelation of the homogeneousspectrum (20).The PCFS interferogram G(2)(d, t) and the

corresponding spectral correlation p(z, t) (in-sets) are shown in Fig. 3, B to E, for PQDs 1 to 4at 3.6 K and for small photon lag times of t <100 ms where we observe the homogeneousspectral information, unaffected by spectraldiffusion. For all PQDs,G(2)(d, t < 100 ms) showsan initial fast decay owing to a fast partialdecoherence of the emission. The exact originof the decay is unknown but is likely due to abroad acoustic phonon side band or fast re-laxation between emissive fine-structure states.The long component in G(2)(d, t) extending overpath-length differences of d ≫ 1 ps implies longoptical coherence times of the ZPL emission.Importantly, the fraction of photons emitted intothe coherent ZPL can be calculated as the squareroot of the coherent decay amplitude of G(2)(d)and ranges between ~0.5 and ~0.8, implyingthat the majority of photons is emitted coher-ently. This ZPL fraction is still smaller than forepitaxial III-V QDs but is already comparableto silicon vacancy centers in diamond, whichare often used in quantum photonics (ZPL frac-tion 0.7) (22).With different orientations of an individual

PQD, one, two, or three emissive fine-structurestates can be observed. The beatings in the

Utzat et al., Science 363, 1068–1072 (2019) 8 March 2019 2 of 5

Fig. 1. Properties of cesium lead bro-mide PQDs. (A) The perovskite crystalstructure is formed from [PbBr6]

− unitsas corner-connected octahedra; theCs+ ions occupy the voids in between.(B and C) High-resolution high-angleannular dark-field scanning transmissionelectron microscopy images confirmthe cubic shape with an average edgelength (L) of ~13 nm and high degree ofsize uniformity at scale. (D) Room temper-ature ensemble absorption (Abs) andPL spectra of the PQDs studied. Theemission energy was 2.38 eV with a FWHMof ~90 meV, indicating minimal inhomo-geneous broadening. a.u., arbitrary units.(E) The emission of PQDs exhibits anexcitonic fine structure. Owing to strongspin-orbit coupling, the total angularmomentum of the electron and the holeare good quantum numbers (Je = Jh = 1/2).Exchange interaction splits the excitoninto a singlet (J = 0) and a triplet (J = 1).The triplet state is split further accordingto its angular momentum projections owing to the Rashba effect. For PQDs with orthorhombic crystal structure, the j jj ¼ T1;0 degeneracy is lifted.The triplet states are emissive, whereas the singlet is optically dark. |YX,Y,Zi, different fine-structure states.

RESEARCH | REPORTon N

ovember 5, 2020

http://science.sciencem

ag.org/D

ownloaded from

interferograms arise from the energy differ-ence between the fine-structure states modulat-ing the envelope of the interferogram, as clarifiedin Fig. 3A for a PQD with two observable fine-structure states. The corresponding spectralcorrelations p(z, t) indicate narrow lines witheither three, five, or seven peaks, depending onthe number of observable emissive states andtheir splittings.Assuming exponential dephasing (Lorentzian

spectral lineshapes) for each emissive fine-structure state, we fit the data with the autocor-relation of two or three Lorentzian peaks ofwidthG defined as G ¼ 2ℏ

T2and energetic separations

of Wi. That is

pðzÞ ¼D∫∞

�∞sðwÞsðwþ zÞdw

Eþ c

where c is a constant accounting for compar-atively broad background emission and s(w) isthe spectral lineshape:

sðwÞ ¼12G

w2 � 12G� �2 þ

Xn¼1∨2

i¼1

ai12G

ðw�WiÞ2 � 12G� �2

The decay of the optical coherence with e�2=T2d

and the beating patterns in the interferogramsowing to the fine structure are well capturedby our fit, allowing extraction of the optical co-herence time T2. For PQDs 1 to 4, we find longcoherence times of T2 ~ 78 ps (68, 88), 52 ps(42, 72), 50 ps (46, 54), and 66 ps (52, 92), withthe confidence intervals given in parentheses.These dephasing times correspond to linewidthsof G ~ 17, 25, 27, and 20 meV, respectively. No-tably, comparison of the optical coherence timesT2 with the spontaneous radiative lifetimes T1of 210, 272, 267, and 269 ps shows that thePQD emission linewidth consistently approachesthe transform limit ( T2

2T1~ 0.19, 0.10, 0.09, and

0.12 for PQDs 1 to 4, respectively). This valueof T2

2T1is two orders of magnitude higher than

for the best specially engineered II-VI hetero-structure CQDs studied to date, which exhibitslow photon release from dark exciton statesand small fractions of coherent photon emis-sion (23, 24). Indeed, although no syntheticoptimization of the photon coherence of PQDshas been conducted, our highest value T2

2T1~ 0.2

is already comparable to T22T1

~ 0.16 to 0.8 fortypical and long-studied epitaxial III-V QDs(4, 8, 25, 26). We further show in (12) that spec-tral diffusion of PQDs is much reduced com-

pared with that of established colloidal materials,likely owing to the absence of surface trapstates in lead halide perovskites (see figs. S7and S8).On the basis of these findings, we suggest

that PQDs can serve as scalable building blocksin sources of quantum light with spectral tun-ability over the entire visible range—a prospecthard to envision with any other quantum emit-ter. When integrated with optical cavities, evena very moderate Purcell enhancement of theemission rate by a factor of ~5 to 10 shouldconsistently yield truly transform-limited emis-sion and thus the emission of indistinguish-able single photons. Generation of polarizationentangled photon pairs via the biexciton-exciton cascade emission could potentiallyexhibit high efficiencies owing to near-unitybiexciton and exciton emission quantum yields.Pursuits in this direction will benefit from theunmatched ease of hybrid integration of CQDswith cavities, which has been demonstrated ina multitude of pilot studies. Indeed, straight-forward integration with plasmonic gap cav-ities (27), dielectric slot waveguides (28), orhigh–quality (Q) micropillar cavities (29) hasbeen shown. Efficient light-cavity coupling hasalso been achieved in plasmonic-QD hybrid

Utzat et al., Science 363, 1068–1072 (2019) 8 March 2019 3 of 5

Fig. 2. Single PQD characterization at 3.6 K. (A to C) Single PQDspectra indicate single lines with two phonon sidebands on the lower-energy side. The resolution limit of our spectrometer does not permitresolution of the fine structure underlying each line. The insets show thepolarization-dependent relative emission intensity. (D) The emissionspectrum of a typical PQD displaying two emission lines with orthogonalpolarization and large enough fine-structure splitting to be resolvedwith the spectrometer. Blue and red indicate polarization of the two dif-ferent lines. (E) Single PQDs show stable emission with minimal spectral

and intensity fluctuations over the course of minutes. (F) Single PQDsexhibit high biexciton emission quantum yields, as seen in the second-orderintensity correlation g(2)(t) (upper panel). When the ZPL is spectrallyselected, sub-Poissonian emission with high single-photon purity isobserved (lower panel). Corr. counts, correlation counts. (G) The emissionlifetime of a single PQD is fast (~210 to 280 ps) and follows a mono-exponential decay over two orders of magnitude. A residual long-livedcomponent can be observed at longer time scales, likely owing to partialdark exciton emission or trapping and recapturing.

RESEARCH | REPORTon N

ovember 5, 2020

http://science.sciencem

ag.org/D

ownloaded from

nanostructures, produced by all-chemical means,which offers a radically different approach (30).PQDs can likely also be deterministically inte-grated with photonic-crystal cavities.In view of these prospects, optimization of

the intrinsic PQD coherence time should be ex-plored. Tuning of the dephasing time has beendemonstrated for colloidal II-VI QDs by lever-aging synthetic control over the fine-structuresplitting to reduce phonon-mediated dephasing

(23). Similar strategies may apply to PQDs, al-though future elucidation of the pure dephasingmechanism in single PQDs may suggest differentstructural handles to the dephasing time. Wesuspect that further passivation of the PQD sur-face, through suitable ligands or growth of inor-ganic shells, may reduce the phonon spectraldensity and increase the coherence time andcoherent fraction of the emission. These effortsmay also further reduce spectral diffusion.

Our results suggest that lead halide perovskites—with their high defect tolerance and opticallybright lowest-lying exciton state—are promisingsemiconductors for the scalable production ofquantum emitters with highly coherent emissionthat can be processed onto virtually any substrateand integrated with nanophotonic components.Rational optimization of these emitters will buildon the tools of colloidal chemistry and the struc-tural versatility of lead halide perovskites.

Utzat et al., Science 363, 1068–1072 (2019) 8 March 2019 4 of 5

Fig. 3. Measurements of the optical coherence times of single PQDs.(A) The PCFS experiment for a PQD with two observable fine-structurestates of energetic splitting W1. The interferogram shows modulation ofthe envelope with a frequency corresponding to the energy differencebetween the two fine-structure states jyY > and jyZ > and loss of photon

coherence decaying with e�1=T2t. PCFS measures the envelope of theinterferogram squared, compiled from photons with small temporalseparation. (B to E) Data for PQDs 1 to 4 and the corresponding spectralcorrelation p(z, t) (insets) for short interphoton arrival times (t < 100 ms)where the effect of spectral diffusion is minimal. The blue line shows the

best fit with our lineshape model, and the black dashed lines indicate theexponential dephasing component of the PCFS interferogram decayingwith e�2=T2d. A fast partial loss of coherence to ~0.3 to 0.6 of the initialamplitude—possibly owing to a broad acoustic feature or fast relaxationwithin the fine structure—can be observed in the interferograms. We

extract long optical coherence times of T2 ~ 50 to 78 ps. The widths of theunderlying Lorentzian lines Ghom ¼ 2ћ

T2and the fine-structure splittings

W1(W2) are indicated in the plot of the spectral correlations. For PQD 4, anadditional unknown side peak is observed that is not captured by ourmodel, likely owing to aliasing common in Fourier spectroscopy.

RESEARCH | REPORTon N

ovember 5, 2020

http://science.sciencem

ag.org/D

ownloaded from

REFERENCES AND NOTES

1. I. Aharonovich, D. Englund, M. Toth, Nat. Photonics 10,631–641 (2016).

2. P. Senellart, G. Solomon, A. White, Nat. Nanotechnol. 12,1026–1039 (2017).

3. P. Borri et al., Phys. Rev. Lett. 87, 157401 (2001).4. A. V. Kuhlmann et al., Nat. Phys. 9, 570–575 (2013).5. S. A. Empedocles, M. G. Bawendi, J. Phys. Chem. B 103,

1826–1830 (1999).6. A. Sipahigil et al., Phys. Rev. Lett. 113, 113602 (2014).7. A. Sipahigil et al., Phys. Rev. Lett. 108, 143601 (2012).8. C. Santori, D. Fattal, J. Vucković, G. S. Solomon, Y. Yamamoto,

Nature 419, 594–597 (2002).9. G. Wrigge, I. Gerhardt, J. Hwang, G. Zumofen, V. Sandoghdar,

Nat. Phys. 4, 60–66 (2008).10. L. Protesescu et al., Nano Lett. 15, 3692–3696 (2015).11. F. Krieg et al., ACS Energy Lett. 3, 641–646 (2018).12. See the supplementary materials.13. H. Utzat et al., Nano Lett. 17, 6838–6846 (2017).14. G. L. Bir, G. E. Pikus, Symmetry and Strain-Induced Effects in

Semiconductors (Wiley, 1974).15. M. Isarov et al., Nano Lett. 17, 5020–5026 (2017).16. M. A. Becker et al., Nature 553, 189–193 (2018).17. R. G. Neuhauser, K. T. Shimizu, W. K. Woo, S. A. Empedocles,

M. G. Bawendi, Phys. Rev. Lett. 85, 3301–3304 (2000).18. G. Rainò et al., ACS Nano 10, 2485–2490 (2016).19. T. T. Tran et al., Phys. Rev. Appl. 5, 2–6 (2016).

20. X. Brokmann, M. Bawendi, L. Coolen, J.-P. Hermier,Opt. Express 14, 6333–6341 (2006).

21. L. Coolen, X. Brokmann, J.-P. Hermier, Phys. Rev. A 76, 033824(2007).

22. E. Neu et al., New J. Phys. 13, 025012 (2011).23. N. Accanto et al., ACS Nano 6, 5227–5233 (2012).24. M. Nirmal et al., Phys. Rev. Lett. 75, 3728–3731 (1995).25. A. Högele et al., Phys. Rev. Lett. 93, 217401 (2004).26. J. Houel et al., Phys. Rev. Lett. 108, 107401 (2012).27. T. B. Hoang, G. M. Akselrod, M. H. Mikkelsen, Nano Lett. 16,

270–275 (2016).28. P. Kolchin et al., Nano Lett. 15, 464–468 (2015).29. M. Kahl et al., Nano Lett. 7, 2897–2900 (2007).30. B. Ji et al., Nat. Nanotechnol. 10, 170–175 (2015).

ACKNOWLEDGMENTS

We gratefully acknowledge helpful discussions with W. Tisdaleand D. Englund. We thank F. Krumeich for the electronmicroscopy measurements. Funding: The lead authors of thisstudy (H.U., W.S., and A.E.K.K.) were funded by the U.S.Department of Energy, Office of Basic Energy Sciences, Divisionof Materials Sciences and Engineering (award no. DE-FG02-07ER46454). K.E.S. was supported by the Center for Excitonics,an Energy Frontier Research Center funded by the U.S.Department of Energy, Office of Science, Basic Energy Sciencesunder award no. DE-SC0001088. B.S. was funded through theInstitute for Soldier Nanotechnologies. M.G. was funded by theNSF under award EECS-1449291. C.F.P. and N.D.K. were supported by

NSF GRFP fellowships. M.V.K. and F.K. acknowledge financial supportfrom the Swiss Federal Commission for Technology and Innovation(CTI-No. 18614.1 PFNM-NM). Author contributions: H.U. conceivedof the study and experiments, built the experimental setup, performedthe single-emitter characterization, PCFS measurements, anddata modeling and interpretation. W.S and A.E.K.K. assisted withsingle-particle spectroscopy and data analysis. B.S. and K.E.S. helpedwith developing data analysis software. F.K., M.G., C.F.P., andM.V.K. synthesized perovskite quantum dots and performedensemble characterization (absorption, PL, quantum yield, andtransmission electron microscopy). N.D.K. developed a lasersystem for acquisition of fast single-emitter lifetimes. H.U. andall authors interpreted the data under the supervision of M.G.B.H.U. wrote the manuscript with input from all authors. Competinginterests: The authors declare no competing interests. Dataand materials availability: All data are available in the manuscriptor the supplementary materials.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/363/6431/1068/suppl/DC1Materials and MethodsSupplementary TextFigs. S1 to S9References

17 July 2018; accepted 7 February 2019Published online 21 February 201910.1126/science.aau7392

Utzat et al., Science 363, 1068–1072 (2019) 8 March 2019 5 of 5

RESEARCH | REPORTon N

ovember 5, 2020

http://science.sciencem

ag.org/D

ownloaded from

Coherent single-photon emission from colloidal lead halide perovskite quantum dots

Katherine E. Shulenberger, Collin F. Perkinson, Maksym V. Kovalenko and Moungi G. BawendiHendrik Utzat, Weiwei Sun, Alexander E. K. Kaplan, Franziska Krieg, Matthias Ginterseder, Boris Spokoyny, Nathan D. Klein,

originally published online February 21, 2019DOI: 10.1126/science.aau7392 (6431), 1068-1072.363Science 

, this issue p. 1068Sciencephotons or entangled photon pairs for quantum information processing.can overcome these limitations and provide unprecedented versatility for the generation of indistinguishable single

now show that perovskite quantum dotset al.production scalability and reproducibility between individual emitters. Utzat molecular beam epitaxy have demonstrated transform-limited emission linewidths. However, they are limited in terms ofquantum emitters with near-perfect optical coherence. Light-emitting defects in diamond and quantum dots grown by

The development of many optical quantum technologies is dependent on the availability of solid-state singlePerovskite quantum emitters

ARTICLE TOOLS http://science.sciencemag.org/content/363/6431/1068

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2019/02/20/science.aau7392.DC1

REFERENCES

http://science.sciencemag.org/content/363/6431/1068#BIBLThis article cites 30 articles, 0 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on Novem

ber 5, 2020

http://science.sciencemag.org/

Dow

nloaded from