Quantum defects in carbon nanotubes as single-photon sources

Abstract

Single-photon emitters are essential components of emerging quantum technologies, including secure communication and quantum computing. Single-walled carbon nanotubes (SWCNTs) have emerged as a promising platform for quantum light sources due to their quasi-one-dimensional excitonic host structure and compatibility with telecom photonic systems. Recent advances in deterministic defect engineering—most notably the development of organic color centers (OCCs)—have enabled stable, chemically controllable, and spectrally tunable single-photon emission. OCC-based emitters have demonstrated single-photon purity exceeding 99% and, more recently, room-temperature photon indistinguishability, placing them among the few solid-state systems with quantum-grade performance under ambient conditions. This review surveys progress in the field from three complementary perspectives: chemical synthesis and quantum defect engineering, computational studies of structure-property relationships and excitonic behavior, and experimental investigations of quantum optical properties. We also discuss alternative approaches, including air-suspended SWCNTs and hybrid van der Waals heterostructures, highlighting opportunities and open challenges for scalable integration into quantum photonic platforms.

Introduction

Single-photon sources are fundamental building blocks for next-generation quantum technologies, including secure quantum communication and quantum computing. The first single-photon source, demonstrated in 1986, was based on atomic transitions¹, which require complex and bulky setups that limit scalability and integration into compact and scalable devices. To overcome these constraints, solid-state quantum emitters have been intensively investigated². Platforms such as quantum dots³ and color centers in diamond⁴ have demonstrated single-photon emission, yet challenges persist regarding room-temperature operation, electrical excitation, and emission at telecom-relevant wavelengths (1.3–1.6 µm), which are crucial for compatibility with existing fiber-optic networks.

Single-walled carbon nanotubes (SWCNTs) provide a unique quasi-one-dimensional host for excitonic states⁵. In pristine SWCNTs, each carbon atom is sp² hybridized, forming three σ-bonds with neighboring carbon atoms and contributing one p-orbital perpendicular to the tube surface^6,7,8. These p-orbitals overlap to create an extended π-bond network that supports delocalized electronic states and tightly bound excitons^5,9. The electronic characteristics of SWCNTs—metallic or semiconducting—are governed by their chirality, defined by the (n, m) indices of the chiral vector^6,10,11,12. Several techniques exist for the selective synthesis and enrichment of semiconducting SWCNTs¹³, which are particularly attractive as excitonic host materials due to their well-defined excitonic transitions and narrow-band photoluminescence (PL), despite intrinsically low PL quantum yields in pristine nanotubes^14,15,16,17. Metallic SWCNTs, by contrast, do not contribute to PL and can introduce additional non-radiative decay pathways when present in mixed samples. In the context of quantum light sources, the primary role of SWCNTs is to serve as a one-dimensional scaffold on which excitons can be localized and engineered through defect formation. Accordingly, throughout this review, the term SWCNT refers exclusively to semiconducting nanotubes used as host materials for defect-based quantum emitters.

SWCNTs can be transformed into quantum emitters through exciton localization. In pristine semiconducting SWCNTs, the lowest-energy optical transition (E₁₁) corresponds to a delocalized exciton that can diffuse over tens to hundreds of nanometers along the nanotube axis¹⁸. Introducing localized potential sites, either via intrinsic structural defects or controlled chemical modifications, can spatially confine these mobile excitons, effectively converting a one-dimensional system into zero-dimensional quantum emitters capable of SPE. The first experimental demonstration of SPE from SWCNTs was reported by Högele et al., who observed photon antibunching at cryogenic temperatures arising from naturally occurring exciton localization¹⁹. While this marked an important milestone, the randomness of exciton localization hindered reproducibility and spectral stability. Subsequent studies by Ma et al. and Ishii et al. demonstrated room-temperature antibunching based on exciton–exciton annihilation (EEA), an intrinsic nonlinear process in 1D systems, though with only moderate single-photon purity^20,21.

To achieve deterministic exciton localization and higher single-photon purity, chemically engineered defect states have been developed. Early efforts using oxygen-based defects²² (e.g., epoxide, ether, hydroxyl) showed promise for SPE but suffered from poor control^23,24. A significant breakthrough was achieved with the introduction of organic color centers (OCCs), first demonstrated by Piao et al., which are formed by creating paired sp³ defects within the sp² carbon lattice of semiconducting SWCNTs and proposed as SPEs²⁵. Importantly, OCC-based emitters have demonstrated exceptional quantum optical performance at ambient conditions, exhibiting single-photon purities exceeding 99%, as first shown by He et al.²⁶, and photon indistinguishability at room temperature, most recently demonstrated by Husel et al.²⁷.

Unlike conventional defects that typically act as non-radiative scattering centers, OCCs effectively introduce a localized quantum emitter embedded within the band structure of the semiconductor host. Mobile E₁₁ excitons can be funneled into this lower-energy site, giving rise to stable, redshifted PL (commonly denoted E₁₁* or ${E}_{11}^{-}$) that is spectrally distinct from the native E₁₁ emission²⁸. The resulting emission spans the shortwave-infrared and the telecom bands (> 900 nm to ~1600 nm), making OCC-functionalized SWCNTs highly compatible with silicon-based photonic platforms²⁹. Crucially, OCC formation can be engineered through controlled chemical reactions, building on the original defect-engineering framework introduced by Piao et al.²⁵ and subsequent developments, providing access to tunable defect configurations and chemistry that are difficult to realize in inorganic defect systems²⁸. Together, these features position OCC-functionalized SWCNTs as a compelling platform for scalable quantum light sources compatible with integrated photonic architectures^28,30.

This review provides a comprehensive overview of recent progress in SWCNT-based SPEs, with a focus on controlled OCC engineering. We examine the formation and behavior of localized defect emitters through three complementary perspectives: (1) chemical synthesis strategies for defect creation and control, (2) computational studies that elucidate structure-property relationships and excitonic properties, and (3) experimental investigations of quantum optical measurements. In addition, we discuss alternative approaches to SPEs, including air-suspended SWCNTs and hybrid platforms based on van der Waals heterostructures, as well as advances in photonic integration and electroluminescence. Through this multidisciplinary review, we aim to highlight both the rapid progress and the remaining challenges toward practical SWCNT-based quantum light sources.

Chemistry of organic color centers

OCCs are created on SWCNT surfaces through covalent chemical functionalization, which converts sp² carbons into localized sp³ quantum defects. First introduced via aryl diazonium chemistry by Piao et al.²⁵, OCCs have since emerged as robust quantum emitters, acting as deep exciton traps that produce bright PL across the near-infrared to telecom bands. Their emission wavelength, brightness, and photostability can be tuned by the functional group chemistry, the chirality of the host nanotube, and, more subtly, the specific binding configuration within the sp² carbon lattice²⁸. A wide variety of covalent chemistries have been developed to expand the accessible quantum defects. For a broader overview of the chemistries, the readers are referred to Brozena et al.²⁸, Shiraki et al.³¹, Zaumseil³², and Janas³³.

This section highlights recent advances in chemistries that enable control over defect configuration/morphology, spatial patterning, and quantum emission properties. In the subsections that follow, we examine the molecular-level chemistry that governs these configurations, the resulting spectral features, and emerging strategies for precision control of defect formation.

Spectral diversity and binding configurations

A defining characteristic of OCCs is their spectral tunability, enabled by the chemically tunable covalent defects and their specific binding configurations on the carbon lattice. However, this tunability also introduces spectral heterogeneity, as different binding geometries, even for the same functional group, can produce a broad distribution of defect-induced PL wavelengths. Understanding and controlling this diversity is essential for developing OCCs as reproducible, high-precision quantum light sources.

Origin of spectral diversity

When an aryl group is covalently attached to a SWCNT, it generates a localized state that can trap excitons and induce red-shifted emission, typically labeled as the E₁₁⁻ or E₁₁* transition (hereafter we use E₁₁* as a general label for all red-shifted emissions). The position of the paired substituent—presumably a hydrogen, hydroxyl, or another aryl radical—can vary depending on which carbon atoms are involved³⁴. Due to the hexagonal symmetry of the graphene lattice, there are six distinct divalent binding configurations within a single hexagonal ring, corresponding to ortho and para positions along different bond vectors relative to the tube axis (Fig. 1a)^35,36. The computational basis for these defect configurations will be further discussed in the section “Computational studies of SWCNTs functionalized with molecular defects”.

Fig. 1: Synthetic strategies and structural configurations of organic color centers in SWCNTs.

a Schematic of main binding configurations for an OCC comprised of an aryl dopant and a pairing group. The gray circle indicates the position of the aryl functionalized carbon in the SWCNT sidewall. The blue and red markers denote the carbon bound to the pairing group for the different possible binding configurations, with the blue indicating the ortho positions (pairing group is on an adjacent carbon) and red indicates the para configurations (pairing group is across a benzene ring). The crosses mark the meta positions, which are energetically disfavored. Three ortho and para sites are further classified according to their bonding orientations, marked as L₂₇ (square), L₈₇ (triangle), and L_− 33 (circle). As an example, a 4-bromobenzene group and a hydrogen atom bound to a (6,5) SWCNT are shown. The right plot shows oscillator strength and emission wavelength of the morphologies in (a). Dashed Gaussian curves show the inhomogeneous distribution expected for each transition. b Two-step chlorosulfonic acid-assisted creation of OCC. The SWCNTs are first dissolved in chlorosulfonic acid, along with an aniline derivative and sodium nitrite, and then added dropwise to an aqueous sodium deoxycholate solution. c UV light and strong base-mediated dual reaction pathway for configuration control. The UV light favors the creation of the higher energy E₁₁* defect peak, while the potassium tert-butoxide favors the formation of the lower energy E₁₁*⁻ peak that is associated with a different binding configuration. d DNA-guided creation of highly ordered sp² OCCs. The guanine base pair in DNA can be activated by singlet oxygen to form an sp² defect on the SWCNT surface. By using a DNA sequence with a quinine pitch that matches its wrapping on the SWCNT surface, the defects can be patterned on the SWCNT surface in a highly ordered manner. e Laser-induced reconfiguration of OCCs. When irradiated with high-intensity light, the OCC can facilitate a shift in the relative binding configurations between the aryl group and the pairing group, allowing for the most thermally stable configuration to be achieved. Panel (a) adapted from He et al.³⁵; Panel (b) adapted from Luo et al.³⁸; Panel (c) adapted from Settele et al.⁴²; Panel (d) adapted from Lin et al.⁴⁸; Panel (e) adapted from Qu et al.⁵⁶.

TD-DFT simulations confirm that these six configurations result in unique local potential wells with emission energies spanning more than 300 meV, and varying degrees of exciton localization and oscillator strengths³⁵. This rich chemical diversity explains the broad and often multi-peaked PL spectra seen in ensemble samples. For example, functionalized (6,5) SWCNTs can emit across the 1000–1350 nm range, with OCC peak distributions attributed to specific ortho and para geometries³⁵.

Host–defect decoupling and structure-property maps

Large-scale spectral studies show that while the emission energy of OCCs depends strongly on nanotube diameter, variations associated with chiral angle and mod(n-m,3) family are comparatively modest at the ensemble level. This suggests a degree of electronic decoupling between the trapped exciton and the 1D semiconductor host³⁷. For instance, perfluorohexyl defects introduced into 14 different chiralities produced defect emissions tunable by nearly 400 meV, but with no systematic dependence on mod(n-m,3) classification when averaged across defect configurations³⁷.

This observation suggests that OCCs behave like zero-dimensional emitters embedded in a one-dimensional scaffold, and their emission energy is determined by the electron-withdrawing strength and binding geometry of the attached moiety. Empirical fits to experimental datasets show that the energy shift ($\Delta E$) scales approximately with the inverse square of the nanotube diameter, while local dielectric environment and binding symmetry introduce secondary variations that contribute to spectral scatter³⁷.

Symmetry-induced narrowing in zigzag nanotubes

Remarkably, the symmetry of SWCNTs can be exploited to suppress configurational diversity. In zigzag nanotubes (e.g., (11,0)), the high lattice symmetry renders all ortho or all para binding sites degenerate, leading to a collapse of the emission spectrum into a single dominant, narrow PL band. Saha et al. demonstrated that selective functionalization of (11,0) SWCNTs with 4-methoxybenzene produced OCC PL spanning only 25 meV across multiple tubes, nearly an order of magnitude narrower than that observed in lower-symmetry chiralities such as (6,5) or (10,3) samples³⁶.

This symmetry-enabled narrowing highlights a promising pathway toward monochromatic, potentially indistinguishable SPEs by structurally selecting or enriching specific SWCNT chiralities with high-symmetry lattices. It also suggests that symmetry plays a key role in binding selectivity, complementing what is achievable through chemical control alone.

Synthetic access to specific binding configurations

While the rich spectral landscape of OCCs arises from multiple possible defect geometries, this diversity also poses a challenge for scalable quantum applications that require uniform emission properties. Recent advances in synthetic strategies now provide a set of tools to selectively favor specific binding configurations by controlling steric hindrance, reaction pathways, or lattice symmetry.

Steric and solvent-controlled defect pairing in superacid-mediated diazonium chemistry

A powerful strategy for controlling OCC formation involves a two-step diazonium reaction in superacids^34,38. When SWCNTs are dissolved in chlorosulfonic acid, the aryl diazonium species, which are pre-formed or generated in situ from aniline and NaNO₂, react with the nanotube to attach an aryl group. The system is then neutralized by the addition of water (Fig. 1b) or other solvents^34,38.

Wang et al. further discovered that the pairing group is not introduced while in the superacid, but rather during the neutralization step, where it is provided by the solvent³⁴. Unlike conventional diazonium reactions in aqueous media, which proceed via radical intermediates, this reaction is believed to follow a carbocation pathway. The carbocation forms on a carbon adjacent to the arylated site and reacts with a nucleophile from the quenching solvent to complete the OCC formation.

This mechanism allows for the deliberate tuning of the pairing group: water yields –OH; alcohols, such as methanol, ethanol, and isopropanol, yield –OCH₃, –OC₂H₅, –OC₃H₇, and –i-OC₃H₇, respectively; and liquid ammonia yields –NH₂. Each pairing group subtly modifies the electronic structure of the defect. For instance, OCCs with the –NH₂ pairing group exhibit ∼20 meV less red-shift than those with oxygen-based groups, due to the electron-donating nature of amines³⁴.

Steric control of binding configuration

The size of the aryl group and the pairing group play a critical role in determining the preferred binding configuration of an OCC. Through experimental hyperspectral PL imaging, Wang et al. showed that larger functional pairs shift the OCC spectral population to result in the dominance of defect emission (∼1240–1270 nm), which becomes the majority PL signal (~90%) in systems like 3F-SO₃-Ar/i-OC₃H₇ (ref. ³⁴). DFT modeling suggests that this shift is due to the para binding configurations becoming more energetically favorable than the ortho configuration for larger pairing groups, primarily because of an increase in steric repulsion. This effect has a much stronger impact on the ortho configuration compared to the para configuration. This steric control strategy not only enables spectral control across both NIR-II (900–1880 nm) and telecom (1260–1675 nm) bands but also improves uniformity in ensemble emission by reducing configurational heterogeneity.

These findings from superacid-mediated chemistry unlock a previously underexplored dimension of OCC design: independent control over the aryl and pairing group identities. The decoupling of these two steps, arylation and pairing, enables the orthogonal tuning of optical, steric, and electronic properties of OCCs.

Despite these advances, several open questions remain: Can this method be extended to less acidic or more biocompatible solvents while preserving control over binding configurations? Are additional Lewis-basic pairing groups, such as thiols or phosphines, accessible within this scheme to modulate exciton trap depths or introduce new spin-related properties? Can dual-defect architectures be created through sequential introduction of distinct aryl/pairing pairs?

Photochemical and chemically-triggered OCC creation

Alternative pathways to defect creation involve photoactivation or nucleophilic substitution, offering potentially orthogonal access to binding geometries.

Wu et al. demonstrated that UV light (335–365 nm) can drive radical formation in (6,5) SWCNTs from aryl iodides³⁹ by photoexciting the SWCNT, which enables electron transfer to the aryl iodide. This photoinduced reduction causes the aryl iodide to dissociate into an iodide anion and an aryl radical, which subsequently reacts with the nanotube to generate OCCs. Zheng et al. extended this approach by using higher-energy UV light (280 nm) to directly excite non-halogenated aromatic systems in aqueous media, enabling defect formation from non-halogenated precursors⁴⁰. Notably, when the dissolved oxygen was removed from the SWCNT solution by purging with argon prior to the UV irradiation, a strong E₁₁*⁻ (in this article we use E₁₁*⁻ as a general label for red-shifted emissions showing lower energy than E₁₁*) emission peak emerged alongside the E₁₁* peak typically observed under ambient air conditions, providing a means to tune the emission through atmospheric control.

An advantage of these photochemical approaches is their compatibility with air-suspended SWCNTs. Kozawa et al. showed that aryl iodides, due to their low vapor pressure, can diffuse in the gas phase and react with suspended nanotubes, enabling functionalization without liquid solvents⁴¹. The details of this novel approach will be further discussed in the section “Single photon emission from SWCNTs and OCCs-bearing SWCNTs”. By contrast, diazonium reactions generally require aqueous solvents to dissolve the diazonium salt, but water’s high surface tension disrupts the integrity of suspended SWCNTs, limiting their accessibility to solution-phase chemistry.

Beyond light activation, OCC precursors can be chemically triggered to produce reactive radicals that readily bond to the nanotubes and create OCCs. Settele et al. demonstrated that strong bases such as potassium tert-butoxide can also induce OCC formation in the absence of light⁴². This base-mediated pathway produces a different distribution of binding configurations than those formed by photochemistry, suggesting a dual-channel mechanism in which the chemical environment determines site selectivity. Base activation has also been explored by Kwon et al., who showed that electron-rich substrates can trigger covalent reactions on carbon lattices without the need for photoinitiation⁴³. More recently, Piletsky et al. show that H₂O₂ can effectively mediate the aryl diazonium chemistry, significantly accelerating the OCC formation⁴⁴. These strategies expand the chemical toolbox for OCC creation, offering new routes to functionalize nanotubes under mild or non-photochemical conditions.

Cycloaddition

Cycloaddition reactions provide a structurally constrained route to controlling OCC configurations by forming both covalent bonds in a single concerted step. He et al. reported [2 + 2] photocycloaddition between SWCNTs and various enone compounds can generate OCCs under UV activation⁴⁵. More recently, Qu et al. demonstrated a thermally driven [2 + 2] cycloaddition between SWCNTs and N-methylmaleimide, which resulted in significantly reduced spectral diversity compared to diazonium-based methods⁴⁶.

The key advantage of divalent chemistry, such as [2 + 2] cycloaddition, lies in its inherent geometric constraint: two adjacent carbon atoms on the SWCNT lattice react simultaneously with the enone, which serves as both the functional moiety and its own pairing group. This locks the defects into an ortho binding configuration, excluding para configurations entirely. As a result, the number of possible binding geometries is effectively reduced by half, leading to narrower and more uniform emission spectra⁴⁶. These insights suggest that revisiting other divalent chemistries, such as those reported by Kwon⁴³ and Gifford⁴⁷, may uncover additional routes to geometrically constrained defect formation.

Additionally, among the three ortho isomers, Qu et al. found that the reaction temperature (80-140 °C) modulates the relative distribution of configurations⁴⁶. This temperature-dependent selectivity provides a synthetic handle to further refine emission properties by controlling the OCC at the atomic level.

DNA-patterning of OCCs

While chemical and photochemical methods typically yield statistical distributions of OCC binding configurations, emerging approaches aim to position OCCs deterministically along the SWCNT lattice. Lin et al. demonstrated that DNA strands with designed periodicity can create remarkably ordered arrays of sp² defects, with a helical spacing that matches the underlying nanotube chirality (Fig. 1e)⁴⁸. This method builds on guanine chemistry initially developed by Zheng et al., who found the guanine base in DNA could be activated to react with SWCNTs by singlet oxygen⁴⁹. Interestingly, the resulting covalent attachments form sp²-type defects—not sp³ quantum defects—but still act as shallow exciton traps, with spectral signatures distinct from those of deeper diazonium-derived OCCs^48,50.

In contrast, Wu et al. developed a DNA-guided approach that incorporates modified bases such as 5-iodo-uracil into the DNA strand as the reactive sites^51,52. Upon photoactivation, the aryl iodide moiety initiates sp³ defect formation, enabling the creation of deep exciton traps with well-defined emission features. This strategy holds the potential to allow nanometer precision in OCC placement, leveraging DNA wrapping geometry and steric effects to tune the spacing between OCCs. The ability to position multiple color centers along a single nanotube with spatial regularity provides a platform for engineering coherent multi-emitter arrays or enabling exciton coupling between adjacent OCCs.

Beyond molecular templating, physical patterning techniques may enable spatially defined patterning of OCCs directly on solid substrates. Huang et al. demonstrated a photolithographic method to “write” OCCs on SWCNT thin films using a light-driven diazoether reaction⁵³, producing micron-scale arrays of SPEs with spectral specificity and robust near-infrared emission⁵⁴. Similarly, Dou et al. employed microcontact printing with solvent-tuned diazonium chemistry to pattern OCCs on surfaces⁵⁵. Together, these approaches bridge chemical and physical patterning, expanding the design space for OCC-based optoelectronic and quantum photonic devices.

Post-synthetic reconfiguration of OCCs

Even after the initial formation of an OCC, the defect configuration is not necessarily the correct one. For future quantum technologies, the ability to erase or correct improperly formed color centers will be essential. To this end, Qu et al. demonstrated that laser irradiation (e.g., at 561 nm) can reconfigure kinetically trapped OCCs into thermodynamically more stable structures⁵⁶. The laser annealing narrowed the PL spectrum by eliminating higher-energy defect states, offering a post-synthetic method to transform defect configurations and refine their quantum properties.

Moving toward uniformity

The ability to synthetically select or anneal OCC configurations brings us closer to achieving reproducible single-photon sources, although several open challenges remain: (1) The reproducibility of selective ortho/para defect formation across different nanotube chiralities remains uncertain: current methods often focus on optimizing (6,5) SWCNTs, while the situation of zigzag or large-diameter nanotubes is still underexplored. (2) The extension of chemical reactivity to other moieties, such as amines, boron, or organometallics, for many non-diazotizing functionalization routes remains unclear. (3) The precision of deterministic defect control using DNA templating, and the feasibility of translating this precision to solid-state lithographic or scanning-probe techniques, also remain unknown. The ongoing development of defect placement and configuration control will be pivotal for transforming OCCs into scalable building blocks for photonic and quantum technologies.

Computational studies of SWCNTs functionalized with molecular defects

Theoretical and computational studies have played a pivotal role in elucidating how OCCs modify the intrinsic electronic and excitonic behavior of SWCNTs. By combining ab initio studies and molecular dynamics approaches, researchers have systematically explored how different chemical modifications introduce specific localized electronic states, alter exciton dynamics, and generate new radiative recombination pathways. The results indicate plausible synthetic targets with desired properties for specific applications. In fact, computational insights became crucial for the rational design of SWCNT-based devices for applications ranging from photovoltaics to quantum light sources. This section presents a brief review of the computational investigations of SWCNTs functionalized with OCCs. We discuss the evolution of theoretical approaches, major findings related to defect chemistry, excitonic behavior, dynamics, and emerging directions for future research.

Early investigations: band structure and localized states

Initial computational studies focused on understanding the basic perturbations induced by molecular defects on the electronic structure of SWCNTs. Tight-binding models and effective mass approximations were among the first approaches employed to predict how covalent functionalization disrupts p-conjugation and induces mid-gap electronic states. It was computationally established that defects introduce localized states within the bandgap, influencing both transport properties and optical absorption⁵⁷. These early models, while simplified, brought forward the critical concept that covalent functionalization could serve as a tool for bandgap engineering. Subsequent first-principles density functional theory (DFT) studies refined these predictions. For instance, calculations of oxygen-containing functional groups reveal that defect sites act as electron acceptors, locally distorting the SWCNT lattice and altering the density of states near the Fermi level^22,58. Together with experimental observations⁵⁹, these findings provided early validation that molecular functionalization can be exploited to control electronic behavior in a predictable manner.

Excitonic effects: from band structure to optical signatures

The discovery of strong defect-induced PL in functionalized SWCNTs sparked intense computational efforts aimed at unraveling the mechanisms underpinning these new emission features from an excitonic or quasi-particle perspective. Time-dependent DFT (TD-DFT)^58,60 and many-body perturbation theory (GW-BSE)⁶¹ were employed to explicitly treat exciton properties and their effects on the radiative emission. These and other theoretical investigations found that defects introduce localized potential wells that can trap mobile excitons, confirming conclusions reached in experiments²⁵. Specifically, TD-DFT simulations of finite-size SWCNTs elucidated how aryl and oxygen adducts create optically active trap states below the pristine E₁₁ band, commonly referred to as E₁₁* excitons (to be discussed later in this section). These localized excitonic states result in strong, redshifted emission with an increased quantum yield compared to pristine SWCNTs.

The overall trapping mechanism can be described as follows: In pristine SWCNTs, the extended network of π-orbitals results in large excitons (reaching sizes of ~5–10 nm or larger) that are highly mobile^62,63. The introduction of covalent defects on the SWCNT surface results in potential wells that trap and scatter the excitons. If spatially trapped at the defect site at the E₁₁ energy range, vibrational relaxation redistributes the excitonic population^64,65,66 from the E₁₁ energy to the E₁₁* or E₁₁*⁻ energy region, effectively localizing and immobilizing the exciton to the vicinity of the defect^58,67,68. The degree to which the defect emission is redshifted from the E₁₁ is dependent on the depth of the potential energy well of the defect⁶⁷. The depth of the well is governed by both the local defect morphology/binding configuration⁶⁰ and the chemical identity of the defect moieties. Across reported defect chemistries and chiralities, the resulting redshifts span ~0.1 eV to nearly 0.4 eV, consistent with experimental trends³⁷. In this way, controlling the optical properties of functionalized SWCNTs is a game of controlling both the defect chemistry and the morphologies that form^28,69.

Chemistry and geometry of defect sites

The wide range of synthetic strategies available for decorating SWCNTs with covalent defects enables substantial structural diversity. Because defect structure plays a central role in determining their optical properties, significant efforts have been devoted to understanding the structure-property relationships and identifying strategies for controlling defect formation. A unique strength of atomistic modeling lies in its ability to map how specific functionalization schemes modulate the electronic structure and light-emission properties. Here, we review how computational studies have elucidated these structure-property relationships associated with covalent functionalization.

In this section, we use “defect morphology” to refer to the relative positions and bonding geometry of the sp³-functionalized carbon atoms, and “defect chemistry” to denote the chemical identity of the attached moieties.

Covalent functionalization

Covalent functionalization has emerged as a reproducible and robust approach for introducing stable chemical defects into SWCNTs^28,32,70,71, converting local carbon hybridization from sp² to sp³. Early DFT simulations^58,72 elucidated that the introduction of a single functional group results in a single bond to an adjacent carbon atom that would either become charged or a radical. This condition is quite unstable, and the reactive environment enables the charged or radical carbon atom to quickly capture a second functional group, resulting in two functionalized carbon atoms in close proximity⁶⁰. There are a number of plausible, unique morphologies that can result^60,73. It turns out that adjacent (nearest-neighbor, “ortho”) or third-nearest neighbor carbons (“para”) provide the most experimentally comparable spectra⁷⁴. Further positions exhibit reduced emission intensity⁷⁴. For zigzag SWCNTs (i.e., those where n or m = 0), there are two unique directions with respect to the SWCNT axis. One of these has both functional groups positioned at carbon atoms along the SWCNT axis. The other configuration places them along the chiral vector 30° (or −30°, equivalent by symmetry) compared to the SWCNT axis. For chiral SWCNTs (i.e., where neither n nor m is equal to 0), there are three distinct morphologies. One involves functionalization of two carbon atoms along the chiral angle, and the other two lie at ±30° from the chiral axis. See Fig. 1a for a (6,5) chiral SWCNT example where the ortho and para configurations are labeled as blue and red, respectively, with all three directions labeled. The stability of each configuration is largely dictated by the hybridization of the carbon atoms in the SWCNT network. Since π-orbital overlap lends stability to the systems, those bonds with the strongest π-orbital overlap in the pristine system are also the least reactive⁷⁵. About a decade ago, researchers began to discover the differences in properties between these morphologies and guide efforts targeting specificity^73,76.

The relative predominance of the (6,5) SWCNT in synthetic efforts⁷⁷ makes it a readily available chirality for functionalization and has been the subject of the most intense study over the past decade. The chiral angle for this tube is ~27° ≈ 30°. The morphology that involves functionalization across carbon atoms lying along the chiral angle is labeled L₂₇ ≈ L₃₀. The other morphologies involve functionalization of carbon atoms lying at 87° (L₈₇) ≈ 90° (L₉₀) and −33° (L_-33) ≈ −30° (L_-30) with respect to the tube axis in a left-to-right reference frame. Spectroscopic studies of functionalized (6,5) SWCNTs revealed the emergence of two distinct emission features. The exciton and resulting spectral feature of the pristine system were labeled as E₁₁. The “11” nomenclature arises since the electronic transition can be described as a transition between the first Van Hove singularity from the occupied single-particle orbitals to the Van Hove singularity from the unoccupied single-particle orbitals. The least-redshifted and most-redshifted emission features are labeled E₁₁* and E₁₁*⁻, respectively. The most prominent emission feature in (6,5) SWCNT is E₁₁*, which is redshifted from E₁₁ by 0.13–0.19 eV^25,43. This emission feature appears in most functionalization conditions. The appearance of a second emission feature, E₁₁*⁻ redshifted about 0.25 eV or more, occurs primarily for specific functionalization conditions, particularly in the presence of a high concentration of reagent and with larger functional groups^34,35. This perplexing result was explored using DFT⁷⁸ and TD-DFT⁷⁹. Finite-length unit cells were employed. The termination of dangling bonds at the ends of the tubes is carefully chosen to avoid the introduction of superfluous edge-localized trap states into the electronic structure⁷⁹. The proper capping scheme is dependent on the value of (n-m) for each tube chirality. These computational studies attribute the main E₁₁* emission feature to the moderately redshifted L₈₇ morphology⁸⁰. The most redshifted E₁₁*⁻ feature is attributed to the L₂₇ morphology. The L_-33 morphology introduces an electronic transition only marginally separated from the E₁₁ band by less than 0.1 eV and therefore appears, in experiment, only as a slightly broadened pristine-emission band⁶⁰. With the main emission features attributed to specific defect morphologies⁸¹, an important synthetic question becomes how to control the formation probability of each morphology and, consequently, the dominant emission features (E₁₁* or E₁₁*⁻) of the resulting material. It is important to note, however, that morphology alone does not uniquely fix the emission energy, and that important contributions from nanotube chirality and defect chemistry further modulate the observed spectral positions³⁷.

The emission energies associated with each morphology are dependent on the chirality of the SWCNT, as heavily discussed and visualized by Gifford et al.⁴⁷. For example, as mentioned above, for the (6,5) SWCNT, the most redshifted ortho feature is the one that functionalizes two carbon atoms along the SWCNT’s chiral angle. However, interestingly, for the (7,5) SWCNT, the same ortho configuration results in the least redshifted spectral feature. Thus, a new notation is needed in order to directly compare the anticipated emission energies across different SWCNT chiralities, which is independent of the exact chiral angle. For this purpose, functionalization along the chiral angle is denoted L+ because it forms the smallest positive angle with respect to the tube axis in a left-to-right (L) reference frame. The morphologies that form the greatest positive angle and negative angle are labeled L + + and L-, respectively. Perhaps most notably, the ordering of redshifts of emission features for different morphologies is largely dependent on the value of mod(n-m,3) for an (n,m) chirality SWCNT⁴⁷.

Using DFT, this observation can be understood by examining the nodal structure of the SWCNT’s frontier orbitals. For pristine SWCNTs, the HOMO and LUMO single-particle orbitals are oriented along non-parallel directions with respect to each other, and between SWCNTs where mod(n-m,3) = 1 (“mod1”) or mod(n-m,3) = 2 (“mod2”), the direction of the HOMO and LUMO flips⁴⁷. In Fig. 2, we show an example for two zigzag SWCNTs, (10,0) [left column] and (11,0) [right column]. The like-phase parts of the HOMO in the (11,0) SWCNT (Fig. 2, right) propagate parallel to the SWCNT axis, while the LUMO propagates along the L+ direction (at an angle with respect to the SWCNT axis). The continuous pathway of in-phase orbital alignment enables electron flow, delocalization, and stabilization of the electronic state. For the mod1 (10,0) SWCNT (Fig. 2, left), the opposite is true.

Fig. 2: Defect configuration dictates excitonic redshifts in SWCNTs via orbital alignment.

Calculated electron density of the HOMO and LUMO for functionalized mod1 (red background) and mod2 SWCNTs (blue background). Green traces indicate regions of continuous electron delocalization due to the same orbital phase overlap. Yellow dashed lines indicate where a node disrupts electron delocalization. Green dots connected by a line indicate the carbon atoms that are functionalized and the bond that is converted to a single bond upon functionalization. The yellow ovals highlight the large regions of loss of electron density in that MO due to functionalization. The HOMO in the mod1 system contains nodes parallel to the SWCNT axis, allowing for electron delocalization in the axial direction. Functionalization along L+ preserves this electronic structure everywhere except on a line through the defect. Functionalization along L + + or L−, however, introduces a node into the electronic structure along the remaining lines. Electron delocalization near the defect site is thereby interrupted, resulting in the formation of spatially localized orbitals around the defect site. Reproduced from Gifford et al.⁴⁷.

Functionalization of two adjacent carbon atoms perpendicular to the like-phase orbital propagation direction introduces a node [see Fig. 2, ortho(+) for the mod1 (10,0)], destabilizing the HOMO and stabilizing the LUMO, which closes the gap and introduces a strong redshift. Functionalization along the direction parallel to the like-phase direction, on the other hand, does not lead to large disruption in the orbital propagation and only provides a weak redshift [see Fig. 2, ortho(++) for mod1 (10,0)]. For mod2 SWCNTs (Fig. 2, right), the direction of like-phase orbital direction is reversed. This explains why the ordering of redshifts for the equivalent defect configurations is reversed between the mod1 (L−, L + +, L +, least-to-most redshift) and mod2 (L +, L + +, L−) SWCNTs⁴⁷. Thus, this additional symmetry (beyond, e.g., the chiral angle) defined by the mod(n-m,3) value now uniquely assigns and provides the ordering of all spectral features for all ortho and para configurations in all possible semi-conducting SWCNTs⁴⁷.

In addition to the emergence of the E₁₁* feature that was observed with functionalization of (6,5) with small chemical groups, E₁₁*⁻ emerges when both ortho carbon atoms in the SWCNT bond to a single carbon atom in the functional group (so-called “divalent functionalization”). There is instability in these configurations, compared to those previously discussed, which bind two functional groups to two SWCNT carbons (monovalent), due to the formation of divalent ‘bridges’ that form an unstable three-membered ring between the ortho carbons of the SWCNT and the functional group. In this case, the L_-33 (L-) configuration becomes quite unstable, as formation of the three-membered ring imposes substantial angle strain on the system. On the other hand, formation of the L₈₇ (L + +) morphology results in very small perturbations to the sp²-hybridization of the carbon atoms in the SWCNT and is thus far more thermodynamically stable. Furthermore, the sp³-hybridization of the carbon atom in the bridging functional group is nearly retained. The result is the favoring of the L₈₇ (L + +) morphology⁷⁵. This demonstrates how different classes of the functional group can direct the defect morphology.

While defect-induced redshifts in functionalized SWCNTs are heavily dependent on defect morphology, the properties of the attached functional groups (i.e., the chemical makeup of the adducts) also play an important role. From a theoretical perspective, the magnitude of redshifts is governed by two main chemical factors: (i) the hybridization character of the defect bond (i.e., the degree of s-character), and (ii) the electron-withdrawing strength of the functional groups through inductive effects^43,75,82. Functional groups that retain higher s-character upon bonding, such that the bonding carbon atoms remain nearly sp²-hybridized, tend to produce smaller redshifts (on the order of ~0.1 eV) compared to monovalent or divalent sp³-hybridized adducts.

In addition to bond hybridization, the electron-withdrawing ability of functional groups provides another degree of chemical tuning within a given defect morphology. For example, replacing CH₃ with more strongly electron-withdrawing 3,5-(NO₂)₂-aryl and -CF₃ substituent in (6,5) SWCNTs increases the E₁₁* defect-induced redshifts by 60 meV and 62 meV, respectively, as experimentally demonstrated by Piao et al.²⁵ and Kwon et al.⁴³. This degree of chemical tunability is substantial, yet remains smaller than the larger morphology-driven energy differences that distinguish E₁₁* and E₁₁*⁻, which are typically separated by ~100 meV or more in the same (6,5) SWCNTs^34,43,75,83. To further extend chemical tunability, heteroatom-based substituents (e.g., direct -F or -Cl attachment) have been explored theoretically using DFT as additional routes to modify defect energetics⁷⁶.

Importantly, although theory robustly predicts the energetic ordering and optical signatures associated with defect morphologies, experimentally accessing specific configurations remains probabilistic and depends sensitively on reaction pathways and kinetics^34,46,56.

Interactions between defects

Covalent functionalization of SWCNTs typically produces sparse, statistically distributed defects, with an average axial separation between adjacent OCCs of ~20 nm under reaction conditions optimized at the ensemble level for bright defect photoluminescence²⁵. More recently, control over the axial separation between adjacent defects has been improved by pre-constructing particular strands of single-strand DNA (ssDNA)^{48,51,84,85,86}. In these systems, the ssDNA forms covalent bonds only at the guanine base units⁸⁶ or the halogenated bases⁵¹ within the ssDNA strand. This tunable separation between defects opens up a novel channel to manipulate the emission signals beyond that of single defects by inducing defect-defect interactions^84,85.

Figure 3a shows the standard model for exciton interactions, often referred to as the Kasha model for interacting excitons⁸⁷. Figure 3b presents the linear absorption spectra as a function of the defect-defect separation length between 1 and 4 nm. This can be modulated in an experiment by increasing the distance between adjacent guanines within the ssDNA strand⁸⁴. From one to two defects, the J-aggregate-like splitting occurs, such that the lower state is optically active while the upper state is dark. As the distance increases, the splitting reduces, retaining the lower state’s bright character at all lengths explored. Figure 3c shows the extracted J-coupling values (black line) as a function of the separation length and scales nearly as 1/r (dashed black line). The coupling parameter is often approximated as classical dipole-dipole coupling $J\sim {\mu }_{01}^{2}\,/\,{r}^{3}\,$; however, we note that exchange interactions K likely exist when the defects are closer than 1–2 nm due to large exciton–exciton overlap⁸⁸, which are neglected in the model but are included in the TD-DFT simulations. Figures 3d and 3e explore the absorption spectra of an increasing number of defects, N = 1 to 5. This can be achieved in experiments by increasing the total number of guanine groups in the ssDNA strand. As the number of defects increases, the low-energy bright state decreases in energy. However, as shown in Fig. 3e, the S₁-to-S₂ splitting is largely unchanged due to the fixed defect-defect separation length between adjacent defects, and all states decrease in energy in parallel with increasing numbers of states. Due to the finite possible size of the bound electron-hole, the energy of the S₁ state exhibits a plateau near E₁ ~ 1.3 eV when extrapolating the transition energy to an infinite number of defects.

Fig. 3: Defect-defect interactions and tunable exciton coupling in functionalized SWCNTs.

a Interacting exciton Hamiltonian. b Simulated absorption spectra at varying separation lengths of two defects along the SWCNT axis. c J-coupling and splitting asymmetry extracted from the TD-DFT simulation. d Simulated absorption spectra as a function of the number of adjacent defects. e Singlet exciton manifold as a function of the number of adjacent defects. Panels (a–c) adapted from Weight et al.⁸⁵. Panels (d, e) adapted from Zheng et al.⁸⁴.

Even though our theoretical understanding of the optical properties in all possible defect configurations as well as defect-defect couplings is robust, our understanding of what happens in the experiment is largely limited. Specifically, how to control the formation of specific defect configurations is lacking⁴². Some efforts have been made in examining the resulting structures from resonance Raman spectroscopy⁸¹. To our knowledge, the best theory that exists relies on the π-orbital mismatch in the region close to the defect⁴⁷. This theory suggests that defects that form near others should result in a different configuration compared to the original defect. If functionalization conditions are such that multiple defects end up spatially close together ( < 10 nm), these directing effects play a role in the resulting morphologies and, therefore, the optical properties of the resulting system. One such condition is a high concentration of reagent for dichlorobenzene-functionalized (6,5) SWCNTs. Systems formed under these conditions are decorated with a higher density of defects than for lower concentrations of reagent. Due to this, some defects are inevitably found in the vicinity of others. For low defect densities, the main emission feature is E₁₁*. As the concentration of the reagent increases, the E₁₁*⁻ peak grows in ref. ⁷⁵. This demonstrates the directing effects of defects and enables increased tunability by controlling defect-defect interactions.

Strong coupling with quantized cavity modes enables increased tunability

Exciting new computational results looking at spin dynamics in functionalized SWCNTs reveal spin-polarized transport and chiral-induced spin selectivity (CISS) effects mediated by molecular defects^46,89. Theory enables quantum defect engineering of SWCNT systems to achieve phenomena like superradiance, multi-exciton generation⁶¹, and coherent light-matter interactions⁹⁰. Strong light-matter interactions^91,92, in particular, show great promise for additional control over the emission frequency and complex optical properties in SWCNTs⁹³. Recently, various experimental explorations into SWCNTs^94,95 coupled to a quantized photonic mode have realized the strong photophysical capabilities⁹⁶, such as long-range energy transfer between bundles of two different pristine SWCNT chiralities (i.e., donor-acceptor process)⁹⁷, photophysics of sp³-defected SWCNT exciton-polaritons⁹⁸ and trion-⁹⁹/biexciton¹⁰⁰-polaritons in pristine samples.

Motivated by the experimental work of Graf et al.⁹⁴, where the authors proposed using strong light-matter coupling to tune the emissive properties of SWCNTs, Fig. 4 presents theoretical work¹⁰¹ exploring the strong light-matter interactions between a pristine (6,5) SWCNT system and a Fabry-Pérot optical cavity. Figure 4a schematically shows the polaritonic Hamiltonian^102,103, which couples the cavity and matter degrees of freedom. While all electronic transitions are coupled to the quantized photon, for SWCNT systems, the bright E₁₁ dominates the light-matter interactions due to the strong dipole strength. The common picture of light-matter hybridization to form the so-called upper |+〉 and lower |−〉 polaritonic states is shown in Fig. 4b, where the total splitting is often referred to as the Rabi splitting Ω_R. Figure 4c shows the excitonic linear spectroscopy inside the cavity at four different cavity frequencies ω_c = (i) 1.0, (ii) 1.5, (iii) 2.0 (cavity-E₁₁ resonance), and (iv) 2.5 eV. Importantly, the energy of the bright, dipole-active character can be manipulated to appear below the threshold of the well-known SWCNT dark state manifold (green vertical dashed lines). Outside of the cavity, these dark exciton states reduce the PL efficiency of the system due to non-radiative relaxation from the dark excitons to the ground state. Looking forward to mixing cavity and sp³-defected SWCNT systems⁹⁵, one may be able to further tune the emission features, or other interesting photophysical phenomena, by introducing complex interplays between the bright E₁₁ and E₁₁* states mediated by the quantized photonic mode.

Fig. 4: Strong light-matter coupling in SWCNTs: polariton formation and spectral tuning.

a Polariton Hamiltonian, b Schematic of light-matter hybridization between electronic and photonic degrees of freedom, forming polariton states. c Excitonic spectra of the coupled SWCNT/cavity polaritonic system for four choices of cavity frequencies ω_c = (i) 1.0, (ii) 1.5, (iii) 2.0 (resonance with the bright E₁₁ transition), and (iv) 2.5 eV, exhibiting strong optical emission below the dark state threshold (vertical green dashed lines). The SWCNT E₁₁ transition density outside the cavity is shown in (c). Panel (c) adapted from Weight et al.¹⁰¹.

Electronic and nuclear dynamics: exciton relaxation, trapping, and chemical reactivity

Beyond static electronic structure calculations, recent computationally intensive efforts have focused on time-resolved simulations to capture excitonic dynamics. These simulations directly follow several key processes, including the competition between free exciton migration and defect trapping, as well as the vibronic relaxation at the defect site after trapping. Such pathways lead to radiative recombination and/or non-radiative losses as well as coherent vs. incoherent exciton dynamics induced by defect-induced energy landscapes. For example, a recent nonadiabatic molecular dynamics study using trajectory surface hopping methods has simulated exciton migration and self-trapping at defect sites, revealing ultrafast processes on sub-picosecond timescales^64,104. Other joint theory-experiment studies^105,106,107 demonstrated complex excitonic dynamics due to the presence of multiple, interacting defects, from the recently developed ssDNA functionalization schemes^48,84,86, that lead to complex relaxation behavior, emphasizing the need for multi-defect modeling^84,85 to replicate experimental conditions.

Another important contribution of these dynamical simulations is elucidating chemical reaction pathways and mechanisms for the formation of specific binding configurations. For instance, recent studies demonstrating the successful photochemical substitution of SWCNTs have been carried out¹⁰⁸. In addition to ortho functionalization (1,2-substitution), para functionalization (1,4-substitution) is also plausible, particularly under photochemical conditions that generate longer-lived intermediates or radicals following the addition of the first functional group. The intermediate can exist in the singlet or triplet excited state, leading to distinct spin-selective photochemistries, ultimately resulting in the formation of ortho and/or para-defect morphologies, respectively¹⁰⁸. These experimental results were supported by first-principles simulations, which mapped out various reaction pathways and identified the corresponding transition states, providing a detailed mechanistic understanding of the functionalization process. A similar theoretical framework was applied to identify the formation of three possible bonding configurations of the divalent molecular defects leading to distinct emission spectra⁴⁶. By calculating transition states (Fig. 5a, b), this effort demonstrated kinetic preference of specific defect morphologies, enabling experimental control through reaction temperature (Fig. 5c). More broadly, emerging theoretical frameworks that connect π-electron topology, Clar sextet distribution, and edge-state formation offer new opportunities to rationalize defect-induced localization in carbon nanostructures and may provide additional guiding principles for future OCC design and placement^109,110.

Fig. 5: Kinetic control of divalent OCC configurations.

Transition state search by the CI-NEB method shows that the cycloaddition reaction pathways of the (a) three divalent OCCs on (6,5)-SWCNT are similar, with differences in barrier heights (b), suggesting kinetic control of divalent OCCs using temperature. The top panels show the calculated three transition states, whereas the bottom panel depicts the potential energy along the reaction coordinate. c Emission and relative energy of the three possible bonding configurations of OCCs generated by N-MMI. TD-DFT simulated emission wavelengths of OCCs featuring the PP( − 2/3, 1/3) (red), PP(1/3, 1/ 3) (blue), and PP(1/3, −2/3) (black) bonding configurations (oscillator strength vs simulated wavelength shown in the left and bottom axes in black), superimposed with the experimental PL spectra from (6,5)-SWCNTs-N-MMI synthesized at 80 (green) and 100 °C (yellow) (PL intensity vs experimental wavelength shown in the right and top axes in red). Panels (a)–(c) adapted from Qu et al.⁴⁶.

Single photon emission from SWCNTs and OCCs-bearing SWCNTs

Quantum optical experiments using SWCNTs initially focused on probing excitonic dynamics and photon statistics under cryogenic conditions, where intrinsic disorder led to localized exciton states capable of emitting single photons¹⁹. Over the past decade, the field has witnessed a transition from studying the naturally occurring localized states in pristine nanotubes¹¹¹ to deliberately engineer defects and dopants^28,32. These modifications have aimed to enhance the fundamental—improving spectral stability at room-temperature operation—as well as quantum optical properties such as; purity: the emitted photons being one at a time, indistinguishability: the coherence of the individual photon where identical optical properties for subsequent emitted photons are expected in terms of wavelength, linewidth, lifetime, polarization, with zero spectral fluctuations, brightness: the rate of photon generation and detection, and integration: the ability to precisely place the quantum defects with nanoscale precision, coupled to nanophotonic structures and electronic circuits for controlling the quantum light emission paving the way for on-chip quantum photonic devices.

In the following subsections, we summarize the defect-induced exciton localization, photophysical properties, and photonic integration of SWCNTs for single-photon emission. In addition to the extensively studied covalently functionalized SWCNTs, recent progress of alternative approaches for SPE in SWCNTs and SWCNT-based materials is also reviewed in this section.

Molecular defects and their photophysical properties

Ma et al. chemically introduced isolated oxygen dopants in (6,5) SWCNTs, forming solitary oxygen-related defect sites on the nanotube surface¹¹². Unlike the sp³ OCCs discussed above, these oxygen dopants preserve the sp² lattice and generate qualitatively different excitonic states. The oxygen-related defects created >100 meV trap states that emitted single-photons with g²(0) around ~0.3 (corresponding to ~70% SPE purity) in the 1100–1300 nm range at room temperature. Although operating till 300 K, the authors identified 200 K as the optimum condition with >50% of the doped tubes exhibiting SPE while free from spectral fluctuations. This temperature regime is accessible using commercial thermoelectric coolers, signifying the potential of oxygen-doped SWCNTs for integrated quantum photonic applications.

He et al.²⁶ showed that high-purity single photons can be achieved with the aryl OCCs. The SPE from these defects exhibit g²(0) values < 0.01, indicating 99% of emission events are single photons at room temperature. Moreover, as SWCNTs come in various chiralities and diameters, each with distinct E₁₁ energies, the emission of these OCCs could be tuned across the entire telecom band simply by functionalizing different nanotube species and thereby extending the wavelength range from 1100–1550 nm. The defect-tailored (6,5) SWCNTs with a smaller diameter of 0.76 nm emitted around ~1100 nm, whereas larger-diameter (10,3) tubes of 0.94 nm emitted single photons directly in the telecom C-band at ~1550 nm. The non-polar environment of the localized defect enabled improved photostability, where the emission count rate fluctuations were closer to the shot-noise limit of detection, indicating minimal blinking or bleaching at room temperature. The reported 12% overall SPE quantum efficiency with ~100 s ps lifetime and <30 meV linewidth at room temperature is promising for quantum photonic applications. Count rates of 100 kHz were reported despite a low combined photon collection and detection efficiency at telecom wavelengths, indicating the potential for scaling to MHz rates through direct coupling to optical waveguides or cavity-enhanced architectures.

A central challenge for OCC-based SPEs is spectral inhomogeneity arising from multiple defect configurations generated during functionalization, which can broaden ensemble emission and hinder the production of identical emitters. Recent advances in chemical control, as discussed in detail in the section “Chemistry of organic color centers”, have begun to address this limitation by selectively favoring specific defect geometries. For example, symmetry-guided functionalization of zigzag SWCNTs and reaction-condition-controlled diazonium chemistry, in combination with post-synthesis annealing, have been shown to dramatically narrow emission distributions and stabilize deeper exciton traps^{34,36,42,46,56}. These advances establish a direct link between defect chemistry and quantum optical performance, demonstrating that precise chemical control enables reproducible, spectrally stable, and tunable single-photon emitters operating into the telecom band. Such progress is essential for realizing scalable arrays of nanotube-based quantum light sources.

Ma et al. studied exciton complexes associated with oxygen dopants, identifying dark excitonic states below the neutral excitons as well as red-shifted trap-bound excitons¹¹³. For OCCs, magneto-PL studies by Kim et al. demonstrated magnetic brightening of spin-dependent exciton fine structures⁸¹. PL blinking dynamics of the defects further revealed the sensitivity of emissions around the environment of the nanotubes¹⁰⁷. Kim et al. reported the PL intensity variations arising from the existence of a fluctuating potential barrier in the vicinity of the defect and also identified the role of dark exciton states by analyzing the temperature-dependent PL dynamics¹¹⁴. Hartmann et al. studied how the E₁₁ exciton lifetime responds to varying concentrations of defects, providing additional insight into relaxation dynamics¹¹⁵. Solvent- and wavelength-dependent PL relaxation dynamics further illustrate the sensitivity of defect excitons to their dielectric surroundings¹¹⁶.

At the single defect level, cryogenic optical spectroscopy revealed even richer excitonic behavior. He et al. reported red-shifted emissions below the E₁₁*, which are now identified as charged excitons or trions, and explored how the nanotube’s surroundings can preferentially stabilize these emissions³⁵. Photon-correlation spectroscopy by Nutz et al. showed that individual hexyl group defects can host either neutral defect excitons or trions, but not both simultaneously, indicating mutual exclusivity of these two states¹¹⁷. Additionally, a metastable dark state can intermittently capture the exciton through a shelving process, introducing a brief off-period in the emission with lifetimes on the order of 10–100 ns in doped nanotubes. Finally, Kwon et al. demonstrated that trions can be efficiently generated and localized at the trapping chemical defects, exhibiting brightness enhancements of up to a factor of 16 compared to neutral nanotube excitons¹¹⁸.

Antibunching in air-suspended SWCNTs

As introduced in the previous sections, OCC formation on SWCNTs offers the advantages of chemically tunable optical properties through the use of varied molecular precursors that can covalently bond to the carbon lattice. Through the introduction of dopant states with different emission energies and the achievement of potential traps deeper than thermal energy, single-photon sources with the desired properties can be realized in these covalently functionalized SWCNTs. Nevertheless, functionalization processes are usually solution-based^39,40, leading to undesired results that can degrade the light emission intensity of SWCNTs, such as shorter nanotube lengths and higher defect densities. Furthermore, due to the high sensitivity of SWCNTs to the external dielectric environment, the exciton properties are susceptible to environmental perturbations. For instance, exciton diffusion in SWCNTs will be hindered by dielectric disorder¹¹⁹, and exciton non-radiative recombination predominantly occurs at quenching sites between the SWCNT and the substrate¹²⁰. To overcome the aforementioned issues and achieve stronger PL emission for quantum light emission applications, air-suspended SWCNTs are considered a promising solution. Using chemical vapor deposition (CVD), air-suspended SWCNTs can be grown on substrates to bridge across designed structures, such as pillar posts^{121,122,123,124}, craters¹¹¹, trenches^{21,119,125,126,127}, or TEM grids¹²⁸. Taking advantage of their pristine and defect-free characteristics, suspended SWCNTs boast their bright PL intensity¹²¹, narrow line width¹²², high quantum efficiency¹²³, suppressed PL blinking and spectral wandering^111,124, and long exciton diffusion length^119,125,127. Air-suspended SWCNTs could therefore play a significant role in the progress of SWCNT-based SPEs.

In the first report on the observation of anti-bunching from air-suspended SWCNTs, Hofmann et al. found that spontaneous exciton localization also occurs in air-suspended SWCNTs at a temperature of 4.2 K¹¹¹, similar to SWCNTs placed on substrates¹⁹. Furthermore, clear anti-bunching behavior with g⁽²⁾(0) values reaching 0.3 was observed through photon correlation measurements with pulsed excitation. In a follow-up work, Hofmann et al. showed that spontaneous exciton localization is a general phenomenon at low temperatures and can result from unintentional potential fluctuation due to structural or environmental disorder induced by contamination or gas adsorption¹²⁸. In comparison, continuous wave excitation photon correlation measurements on ultra-clean air-suspended SWCNTs conducted by Sarpkaya et al. showed no anti-bunching behavior at 10 K, despite the observation of enhanced linewidth and prolonged PL decay time in these SWCNTs¹²⁴. The inconsistent results may stem from the use of different laser operation modes, where pulsed excitation offers more advantages in observing anti-bunching than continuous wave excitation due to its better temporal resolution. Furthermore, the time jitter of detectors may also affect the photon correlation results. Following the work on low-temperature photon correlation measurements, Endo et al. investigated the temperature dependence of photon correlation from 6 to 300 K on air-suspended SWCNTs¹²⁶. The anti-bunching behavior was sustained at elevated temperatures and even at room temperature, although spontaneous exciton localization is expected to become weaker¹²⁸. While the observation of anti-bunching behavior at room temperature is encouraging, the g⁽²⁾(0) value of ~0.6 in this experiment is insufficient to claim single photon emission at room temperature.

In another report on room temperature photon correlation characterization of air-suspended SWCNTs, Ishii et al. obtained a low g⁽²⁾(0) value of 0.27, resulting from EEA²¹. In the EEA process, an exciton can recombine nonradiatively by giving its energy to another exciton via an Auger-like process during the scattering of two excitons^129,130. When multiple excitons exist, EEA causes rapid recombination until there is only one exciton remaining, leading to single-photon emission. Since the excitons in air-suspended SWCNTs exhibit long diffusion length and are confined to one dimension, efficient EEA was reported¹³⁰. In the study of anti-bunching in air-suspended SWCNTs by Ishii et al., the dependence on excitation power, SWCNT chirality, and suspended tube length was compared. The g⁽²⁾(0) value was observed to increase with excitation power and eventually saturates, while the saturated g⁽²⁾(0) value and the minimum achievable g⁽²⁾(0) value vary with SWCNT chirality and tube length. Based on the simulation results, it was concluded that low exciton generation rate, long exciton diffusion length, and long nanotube length are beneficial for achieving low g⁽²⁾(0) values. Among these factors, the length of the tubes has the largest influence on obtaining low g⁽²⁾(0) values. Since the only defect sites in pristine air-suspended SWCNTs are the tube ends, the proportion of excitons being quenched is reduced for longer nanotubes. To further enhance the EEA rate in air-suspended SWCNTs, Li et al. employed a non-covalent functionalization method in which pentacene molecules were physically adsorbed onto the SWCNT surface¹³¹. The local bandgap shrinkage caused by the dielectric screening effect led to the formation of a potential trap for excitons, which would collect the free excitons into the trap, as shown in Fig. 6a. If the size of the potential trap is smaller than the exciton diffusion length, improved EEA efficiency and thereby enhanced single photon emission is expected. As shown in the photoluminescence excitation (PLE) results in Fig. 6b, a new PL emission peak with an emission energy ~15 meV lower than the free exciton peak appeared after molecular decoration, indicating the formation of a shallow potential trap and radiative exciton recombination at the decorated sites. Second-order photon correlation measurement of the new PL peak showed clearly improved anti-bunching behavior, with g⁽²⁾(0) values enhanced by up to 27%. The achievement represents a significant step toward realizing room-temperature single-photon emission utilizing only the EEA phenomenon in SWCNTs. Lower g⁽²⁾(0) values are expected if deeper potential traps can be formed using materials with higher dielectric constants.

Fig. 6: Air-suspended SWCNTs as quantum emitters.

a Schematic illustration of quantum well formation and enhanced EEA caused by pentacene decoration on SWCNT. b PLE map of a pentacene-decorated SWCNT. c Schematic of an OCC-functionalized suspended SWCNT. d PL comparison of a SWCNT before and after functionalization. e PL imaging of the PL peaks shown after functionalization in (d). f Schematic illustration of SWCNT/WSe₂ heterostructure. g Band alignment between WSe₂ and SWCNTs with varied SWCNT chirality. h PL spectra of heterostructures formed by WSe₂ and SWCNTs with varied chirality. Additional PL peaks originating from the interface exciton appeared as the band alignment switched from type I to type II. iSecond-order photon correlation measurement result from the interface exciton emission peak of SWCNT/ WSe₂ heterostructure. Panels (a) and (b) adapted from Li et al.¹³¹; Panels (c–e) adapted from Kozawa et al.⁴¹; Panels (f–i) adapted from Nang et al. (2024)¹³³.

Advanced nanostructures for quantum emission

As discussed in the previous sections, OCCs provide significantly deeper potential traps compared to those created by physical molecular decoration, with trap depths exceeding the thermal energy scale. Enhanced anti-bunching behavior is therefore expected in air-suspended SWCNTs by forming OCCs. The conventional solution-based functionalization process, however, easily damages the air-suspended SWCNTs. To functionalize air-suspended SWCNTs, Kozawa et al. introduced a vapor-phase photochemical reaction to create OCCs⁴¹. Air-suspended SWCNTs were first sealed in a chamber filled with iodobenzene vapor. Following UV light illumination, OCCs could be formed on the SWCNT surface, resulting in the appearance of new PL emission peaks, as shown in Fig. 6c and d. The PL images in Fig. 6e clearly reveal that the additional peaks come from the same SWCNT as the E₁₁ peak. Similar to the functionalized SWCNTs introduced in the previous section, the additional PL emission peaks resulting from defect formation in air-suspended SWCNTs also exhibited chirality dependence in PL emission energy and prolonged exciton lifetime. Although photon correlation measurements were not conducted in this study, purer single-photon emission is anticipated if the vapor-phased reaction could be finely controlled to form only one OCC on air-suspended SWCNTs.

Another recent advancement in developing quantum emitters based on air-suspended SWCNTs was achieved through the formation of type II band alignment in mixed-dimensional heterostructures. Nang et al. addressed the first report on the fabrication of air-suspended mixed-dimensional heterostructures consisting of one-dimensional SWCNT and two-dimensional tungsten diselenide (WSe₂) (Fig. 6f), and examined the influence of band alignment types^132,133. Due to the variable band structure of SWCNTs, which depends on chirality, the band alignment of these mixed-dimensional heterostructures can be adjusted between type I and type II band alignment, as shown in Fig. 6g. While no additional PL emission peaks could be observed in heterostructures with type I band alignment, newly emerged PL emission peaks originating from interface excitons could be clearly observed in type II band alignment heterostructures with energies falling within the telecommunication wavelength range (Fig. 6h. The SWCNT E₁₁ and interface exciton PL emission are further compared by excitation power dependence of PL intensity, time-resolved PL, and photon correlation. The much faster saturation of PL intensity compared to E₁₁ emission under low excitation power and the prolonged PL decay time suggest strong localization of interface excitons. This was confirmed by the anti-bunching behavior in photon correlation with g⁽²⁾(0) value reaching 0.467 (Fig. 6i). Since there are almost no defect sites on the SWCNT, the strong localization of these interface excitons was attributed to either tungsten vacancy defect sites in WSe₂ or inhomogeneous local strain in WSe₂. In either case, the achievement in mixed-dimensional heterostructures opens up a new direction for developing room-temperature SPEs at telecommunication wavelengths based on air-suspended SWCNTs.

Photonic and electronic device integration

The integration of quantum emitters into photonic and electronic structures is essential for quantum photonic technologies. On one hand, this integration allows for nanoscale control over the emitter’s optical properties, helping to mitigate the spectral diversity among individual quantum emitters and allowing for deterministic tuning. On the other hand, quantum emitters in free space or on bare substrates emit photons isotropically, making efficient collection and use of those photons highly challenging. Photonic integration offers a solution to enhance light-matter interaction by enabling high Purcell enhancement, which not only boosts the spontaneous emission rate to achieve higher photon count rates but also channels emission into well-defined optical modes, allowing efficient coupling into on-chip waveguides and off-chip fibers. This is critical not only for high collection efficiency but also for achieving photon indistinguishability, a prerequisite for quantum interference and entanglement-based protocols. At the same time, electronic integration enables precise control over emission characteristics, supporting electrically triggered on-demand single-photon generation without relying on optical excitation, and allowing spectral tunability and modulation through controlled exciton injection and transport. Thus, photonic and electronic integration collectively enhance emitter performance, improve photon extraction and routing, enable precise control over quantum emission, and ensure efficient interfacing with photonic circuits and external systems.

He et al. reported dephasing times (T₂) at 4 K ranging from 3 to 12 ps for sp³ defects, with polymer embedding inhibiting charge noise fluctuations, which is an improvement over the ~2 ps typically observed for the band-edge neutral excitons¹³⁴. These longer coherence times are attributed in part to reduced phonon coupling enabled by the deeper confinement of the localized exciton states. T₁ lifetime of 1.5 ns, approaching the intrinsic radiative limit of 2 ns, indicated a high defect-state quantum yield of 50–75%. While promising, the radiative rate must be increased by two orders of magnitude to achieve photon indistinguishability (T₂ = 2T₁). Therefore, integrating functionalized SWCNTs into photonic structures to enhance spontaneous emission via the Purcell effect is crucial for improving the coherence properties.

Reported photonic device architectures include:

1. Dielectric nanocavities and waveguide coupling: Ishii et al. demonstrated that coupling OCC emitters to photonic crystal microcavities (see Fig. 7a for device architecture) resulted in a ~ 50-fold increase in PL intensity with emission rate about ~17 MHz and ~30% reduction in the emission lifetime confirming an increased radiative decay rate due to the cavity coupling while preserving the single-photon purity even as the excitation power was increased to the saturation level¹³⁵. The work showed that doped SWCNTs can be combined with mature silicon photonics technology to achieve a brighter source with a higher photon emission rate on the order of tens of MHz while preserving antibunching without transitioning to multi-photon output—an important demonstration for on-chip integrated photon sources. Closely related nanobeam cavities featuring ultralow mode volumes have previously yielded high coupling to individual SWCNT emitters and efficient coupling to integrated waveguides, offering a clear route to on-chip routing of photons after cavity emission²⁹.

Fig. 7: Photonic and electronic device architectures integrating OCC functionalized SWCNT SPEs, highlighting coupling mechanisms, efficiencies, and fabrication routes.

a Dielectric cavity: OCC-SWCNTs coupled to a silicon photonic crystal cavity fabricated by e-beam lithography and reactive-ion etching, yielding ~50× PL enhancement and ~30% lifetime reduction through Purcell-enhanced coupling to the cavity mode. b Plasmonic antenna: An OCC-SWCNT positioned within the nanoscale plasmonic nanogap of an Au cavity fabricated by e-beam lithography, achieving Purcell factors up to ~415, radiative lifetime shortening to ~10 ps, and cryogenic HOM visibility up to 79%. c Fiber-coupled open Fabry-Perot cavity: A single OCC-SWCNT emitter positioned at the cavity waist formed between a laser-machined concave fiber mirror and a planar DBR, enabling direct single mode fiber coupling with Purcell factors up to ~30 and photon emission rates ~20 MHz. d Room temperature indistinguishability in fiber-coupled open cavity: A linewidth engineered open microcavity in which the cavity linewidth, rather than the emitter dephasing, defines the photon coherence time, enabling room-temperature two-photon interference with visibilities of 50–65% at telecom wavelengths without external spectral post-filtering. e Ambipolar thin film network: OCC-functionalized SWCNT networks incorporated into an ambipolar light emitting FET with lithographically defined Au contacts, demonstrating gate tunable defect state electroluminescence arising from ambipolar carrier injection. f Single nanotube field-effect transistor: A deterministically placed OCC-SWCNT contacted by graphene electrodes and controlled via electrostatic gating, enabling electrically triggered SPE. The accompanying zoom-in highlights the nanotube channel and contact region, demonstrating precise emitter placement and defect state emission from an individual nanotube. Panel (a) adapted from Ishii et al.¹³⁵; Panel (b) adapted from Luo et al.¹³⁶; Panel (c) adapted from Borel et al.¹³⁷; Panel (d) adapted from Husel et al.²⁷; Panel (e) adapted from Zorn et al.¹³⁹; Panel (e) adapted from Li et al.¹⁴⁰.

2. Plasmonic antennas: Luo et al. exploited plasmonic cavities to push SPEs into a regime of ultrafast emission with demonstrated photon indistinguishability at cryogenic temperature while overcoming the short T₂ lifetimes compared to the longer T₁ lifetimes in SWCNTs¹³⁶. The authors showed that a nanotube placed in the gap of a metallic nanoantenna structure (see Fig. 7b) can enhance the Purcell factors up to ~415 with cavity-enhanced quantum yield of 74%—where the OCC’s radiative lifetime was shortened from an initial ~700 ps to as little as 10 ps effectively narrowing the emission linewidth relative to the radiative rate—while maintaining the single-photon purity up to ~99% confirming that the plasmonic structure did not introduce multi-photon artifacts. At a cryogenic temperature of 4 K, the plasmonic-enhanced emitter showed two-photon interference (Hong-Ou-Mandel: HOM) visibilities up to 79% over 150 nm bandwidth, indicating the indistinguishability or enhanced coherence of the emitted photons even without any temporal post-selection or resonant excitation. Moreover, at room temperature, the cavity enabled a > 50-fold enhancement in photon emission rate, with an estimated rate of up to 1.3 GHz into the first lens, indicating a higher brightness factor in these integrated devices.

3. Fiber-coupled open microcavities: Borel et al. reported coupling an individual OCC-functioned SWCNT to a tunable fiber Fabry-Pérot open cavity, achieving a fiber-coupled telecom single-photon source that delivers up to ~20 MHz of single photons into a single-mode fiber with a Purcell factor of up to ~30¹³⁷. The fiber-coupled open cavity used to enhance and directly collect SWCNT emission into a single-mode fiber is depicted in Fig. 7c. The open cavity design enabled frequency tuning and measurement of the phonon sidebands of the emitter under different detunings, where a vacuum Rabi splitting of approximately 25 μeV was inferred, indicating that the system approached the strong-coupling threshold. This fiber-cavity platform is particularly promising because it combines Purcell enhancement with efficient fiber collection—a turnkey single-photon source that can be directly integrated into quantum networks.

Remarkably, Husel et al. demonstrated the first achievement of room-temperature photon indistinguishability by operating the fiber-based open microcavity in a regime termed incoherent good-cavity coupling, where the photon coherence time is determined by the cavity linewidth²⁷ (see Fig. 7d for a schematic depicting the device configuration). Although the emitter’s coherence time (a few ps) is shorter than the cavity photon lifetime (tens of ps) due to room-temperature dephasing, the cavity still filters and enhances the emission effectively with an increase in two-photon interference visibility (50–65%) by over two orders of magnitude ( ~ 217-fold) compared to the free-space uncoupled case. The result was achieved without any external filtering that would reduce the source brightness. Here the cavity itself was used to select a narrow spectral feature of the emission and enhance the PL, demonstrating for the first time that solid-state single photons at telecom wavelengths can show quantum interference properties even at 300 K. This work showed that cavity quantum electrodynamics (QED) could be used to mitigate thermal decoherence for SWCNT SPEs, bringing room-temperature indistinguishable photons closer to reality for a solid-state system.

Reported electrical device architecture: Given that the emission properties of SWCNTs can be enhanced using dielectric resonators or photonic crystal cavities via the Purcell effect, the demonstrations of electroluminescence (EL) through electrically driven OCCs mark significant progress toward on-chip single-photon sources.

1. Thin film network: Xu et al. demonstrated EL emission from aligned thin films of OCC-functionalized SWCNTs using a two-electrode device and observed both neutral excitons and charged trions at OCC sites¹³⁸. A significantly lower threshold electric field for EL generation was observed in OCCs (0.015 MV/cm) compared to pristine SWCNTs (0.95–1.9 MV/cm), suggesting their potential as near-infrared light sources. Zorn et al. investigated the electrical and optical properties of OCC-functionalized SWCNT networks using ambipolar light-emitting field-effect transistors¹³⁹ (see Fig. 7e). The moderate decrease in carrier mobility with increasing defect density implies the formation of shallow charge traps through which EL was generated via ambipolar carrier recombination. In addition, the EL was shown to be tunable with defect concentration and stable over a wide range of current densities. It should be noted that although SWCNT networks enable EL demonstrations, issues such as unaligned nanotubes and bundling hinder their suitability for SPE applications. Unaligned nanotubes limit efficient coupling to photonic structures, while bundling drastically reduces emission intensity through intertube quenching.

2. Single nanotube field-effect transistor: To overcome network-related limitations, Li et al. fabricated a field-effect transistor device with a graphene electrode based on an individual SWCNT and demonstrated defect-induced EL¹⁴⁰ (see Fig. 7f). With gate voltage modulation, the EL emission can be switched from neutral to charged defect state emission. The high spectral purity of charged-state emission is indispensable for the development of on-chip quantum sources. Furthermore, Li et al. conducted the photon correlation measurement on individual OCC-functionalized SWCNTs at 77 K, reporting a minimum g⁽²⁾(0) value of 0.08 for the electroluminescence from the OCC¹⁴¹. Although it was found that the g⁽²⁾(0) value increases with higher applied electrical power, the observation of anti-bunching in EL emission is a significant progress toward realizing room-temperature single-photon electroluminescent emission.

In addition to basic device integration, strain or electric field tuning of emission can be used to align OCC spectra post-fabrication for device matching or wavelength multiplexing^141,142. Compared to other SPEs, OCCs offer a unique blend of synthetic accessibility, spectral tunability, and nanoscale integration. The 1D structure and well-understood chemistry of OCCs also facilitate hybridization with plasmonic antennas, metasurfaces, and nonlinear optical circuits, paving the way for future quantum photonic systems.

Having discussed the key quantum optical characteristics of OCCs, it is instructive to place their performance in the broader context of established solid-state SPEs. Compared to other SPE systems, OCCs in SWCNTs have emerged as a promising class of quantum emitters, particularly due to their telecom-range emission, exceptional room-temperature single-photon purity, and chemical tunability. As summarized in Table 1, SWCNT OCCs offer a distinct combination of performance attributes relative to established solid-state SPE platforms. Their emission in the telecom window directly matches the lowest-loss operating regime of existing fiber infrastructure, while their chemical tunability—arising from control over both defect configuration and substituent chemistry^25,37—provides a level of spectral flexibility that is unavailable in other inorganic defect systems and is advantageous for integration with photonic architectures. Among state-of-the-art emitters, only InAs/GaAs quantum dots exhibit comparable emission wavelengths; however, they require cryogenic cooling to achieve high purity and indistinguishability. By contrast, SWCNT OCCs demonstrate excellent room-temperature antibunching ( > 99%)²⁶ and indistinguishability²⁷, making them strong candidates for practical quantum communication technologies.

Table 1 Comparison of SWCNT OCCs with state-of-the-art solid-state single-photon emitter systems

The room-temperature indistinguishability of OCC emission (0.65 ± 0.24) is currently limited by strong dephasing processes²⁷. Although this performance surpasses that of most other room-temperature solid-state SPEs, it is still below the near-unity values achieved by epitaxial quantum dots at cryogenic temperatures. The lack of established strategies for coherence enhancement represents a key challenge. Progress in identifying dominant dephasing mechanisms, quantifying phonon-limited coherence, and implementing cavity-assisted enhancement will be essential for advancing OCC-based SPEs toward indistinguishable photon generation.

From a fabrication perspective, OCCs offer moderate-to-high scalability due to the precise chemical controllability of defect formation through photochemical and patterned activation strategies^51,54. However, reproducibility remains a significant challenge. Chirality-dependent emission energies, variations in defect binding geometries, multiple accessible bonding configurations, and the low deterministic yields of isolated emitters all contribute to sample-to-sample variability. In addition, sensitivity to the local environment can induce spectral wandering and blinking through interactions with chemical and dielectric fluctuations^28,143. Together, these effects result in spectral inhomogeneity that currently limits large-scale integration.

In terms of spectral linewidth, SWCNT OCCs exhibit values in the range of ~270 µeV at cryogenic temperatures^35,108,144, which are broader than the µeV-scale linewidths observed in NV centers in diamond, V⁻ centers in SiC, and h-BN defects^145,146,147. Their linewidths are nevertheless narrower than, or comparable to, those of perovskite quantum dots (35–65 meV at room temperature; 1–5 meV at cryogenic temperatures)^148,149 and defect-bound excitons in monolayer TMDs (~33 meV at room temperature; ~4 meV at cryogenic temperatures)¹⁵⁰. These linewidths reflect both strong exciton–phonon coupling in 1D materials and local structural variability introduced by covalent functionalization.

The most significant limitation of SWCNT OCCs remains their relatively low quantum yield (~ 16%)²⁵, which is substantially lower than that of NV centers, SiC defects, perovskite quantum dots, and III–V epitaxial quantum dots. This limitation may arise in part from the intrinsic characteristics of one-dimensional nanotubes, where strong non-radiative pathways and exciton diffusion lead to low PL quantum yields, typically around 1% for pristine SWCNTs¹⁵¹. Overcoming this limitation will require advances in suppressing non-radiative recombination, enhancing exciton funneling into defect states, and integrating OCCs with photonic structures that enhance radiative emission.

Conclusions and outlook

For the chemistry of OCC research, the development of techniques that enable the control of binding configuration has progressed rapidly. Through advanced methods, such as steric control, solvent property-mediated chemistry, photochemical reactions, cycloaddition, DNA patterning, and post-synthetic reconfiguration, specific binding configurations of OCCs on SWCNTs can be achieved, which strongly affect the emission energy of OCC PL. Despite rapid progress, several critical questions remain for advancing precision OCC chemistry: (1) Atomic-Level Control: Can we achieve deterministic placement of OCCs at specific lattice sites with single-carbon resolution? (2) Coupled Defects and Quantum Coherence: What is the nature of exciton coupling between closely spaced OCCs? (3) Beyond Aryl Systems: Can new classes of chemistries yield OCCs with unique emission profiles? (4) Spin and Charge Tunability: What roles can OCCs play in spin coherence or trion stabilization? (5) Integrated Device Architectures: How can OCC-SWCNTs be patterned with photonic structures for enhanced readout? Addressing these questions will be pivotal for advancing OCC chemistry toward scalable quantum photonic technologies.

From a theoretical perspective, computational studies have been indispensable in understanding the profound impact of molecular defects on the electronic and optical behavior of SWCNTs. From early band structure modifications to complex excited-state dynamics, theory has not only validated experimental findings but also guided new functionalization strategies. Going beyond atomistic simulations, studies implementing mesoscopic models enable the study of collective mechanical¹⁵² and optical properties¹⁵³ of SWCNT networks and films. The integration of first-principles simulations, nonadiabatic dynamics, and machine learning will be critical to accelerating the design of next-generation functionalized SWCNT systems. As computational methodologies continue to evolve, they promise to further drive innovation in SWCNT-based materials for quantum technologies, optoelectronics, and sustainable energy applications.

In parallel with these advancements, several emerging computational directions are rapidly gaining momentum: (1) Spin dynamics: Investigations of spin-polarized transport and chiral-induced spin selectivity (CISS) effects mediated by molecular defects. (2) Machine learning models: Predicting defect properties, formation energies, and optical responses across large chemical spaces. (3) Quantum defect engineering: Tuning multiple defects to achieve phenomena like super radiance, multi-exciton generation, and coherent light-matter interactions. (4) Multi-scale modeling: Bridging atomistic simulations with mesoscopic models to study collective optical properties of SWCNT networks and films. Emerging and advancing computational methods are enabling this increased momentum. For example, the development of new machine learning models is making prediction of the energetics more feasible for larger systems in a shorter time frame^154,155. Novel physics-informed molecular machine-learning models can be used to compute optical responses¹⁵⁶. Alternatively, much progress has been made in the reduced Hamiltonian formulations, resulting in the semi-empirical extended tight binding (xTB)¹⁵⁷ and xTB-based force fields (GFN-FF)¹⁵⁸, both of which now include most of the periodic table¹⁵⁹ and is compatible with (i) a fully differentiable framework via auto-differentiation¹⁶⁰ replacing the need for analytical nuclear gradients and (ii) simplified response properties (sTD-xTB)¹⁶¹ for the accurate and efficient calculation of excited states, an area in which machine learning in chemistry often fails for broad-aimed machine learning frameworks. These tools enable the exploration of molecular dynamics trajectories across larger chemical spaces with the accuracy of high-level theories at a fraction of the computational resources^162,163.

For experimental research of SPE in SWCNTs, the progress has been rapid with quantum defect-localized excitons in SWCNTs emerging as a promising platform for single-photon generation offering several features: room-temperature operation: localization at room temperature requires deep potential traps much greater than k_BT = 26 meV enabled through covalent functionalization chemistry, telecom wavelengths: the emission wavelength is governed by the defect’s binding configuration and the host nanotube’s bandgap offering wider wavelength tuning across the telecom O, C bands (~1.1–1.5 µm) by selecting appropriate SWCNT chiralities or defect types, and microfabrication platform integration: compatibility with existing photonic and electronic fabrication technologies. These community-wide efforts have now opened promising research directions to fully realize their potential in quantum photonics: (1) Deterministic emitter arrays: nanotubes with defects are often randomly dispersed on substrates today, which is not ideal for scalable circuits. Developing methods to position and align individual defect-emitters at designated locations on a chip is important. Techniques such as DNA-origami placement¹⁶⁴ or templated growth on pre-patterned catalysts¹⁶⁵ or those methods adapted from SWCNT thin films¹⁶⁶ might allow precise emitter arrays in the future. Furthermore, the reported experiments to date have studied one nanotube at a time. For quantum network applications, multiple sources emitting indistinguishable photons are needed. The chemical control of defect configuration will be vital to make photons from different nanotubes identical in spectrum. On-chip spectral tunability of different emitters will also help in aligning multiple emitters to identical spectral properties. (2) Enhanced coherence: even though indistinguishable single photons are obtainable at room temperature using cavity QED structures²⁷, their coherence is still limited by fast dephasing. While improving cavity parameters or using hybrid plasmonic-dielectric structures could be explored, additional controlling strategies, either chemical or physical means, could also be developed. Operating at cryogenic temperatures, such as 4–200 K, can sharpen linewidths, and combining this with cavity enhancement could yield near-unity HOM visibility. (3) Entangled photon generation: Generating entangled photon pairs from nanotubes, for example, via biexciton–exciton cascades as in quantum dots¹⁶⁷, has not yet been observed. Exploring multi-exciton physics in defect-rich SWCNTs could be an exciting direction. (4) Toward fully integrated quantum photonic devices: Future device engineering should aim to unify electrical excitation, photonic routing, and coherence preservation within a scalable platform for various applications in quantum technologies. Achieving this will require combining the most promising aspects of the current device families: the high Purcell enhancement and waveguide compatibility of dielectric nanocavities¹³⁵, the ultrafast emission achievable with plasmonic antennas¹³⁶, and the telecom-band fiber-coupled operation of open microcavities^27,137, together with the electrical addressability already demonstrated in single-nanotube EL FETs^140,141. In addition, emerging fabrication approaches such as two-photon polymerization or direct laser writing enable 3D-printed photonic wire bonds, microcavities, microlenses and mode-selective couplers that can be written directly around solid-state emitters and aligned to low-loss waveguides, as recently demonstrated for telecom quantum dots interfaced with silicon nitride circuits¹⁶⁸. A key frontier lies in hybrid architectures that co-locate electrical contacts and optical confinement without degrading emitter coherence. For example, a hybrid electrical-photonic SWCNT device integrating graphene electrodes with a photonic crystal cavity achieved a coupling efficiency β_int of 99.5% with Purcell factor of 188 with on-chip waveguide out-coupling¹⁶⁹. This outlines a realistic route to conductive-dielectric cavities that can support electrical drive and on-chip photonic routing simultaneously. Together with advances in OCC photopatterning, deterministic SWCNT placement, and CMOS-compatible processing^{54,55,56,164,165,166}, these techniques define a clear pathway toward electrically driven, cavity-enhanced, telecom-band SWCNT single-photon emitters integrated within complex on-chip photonic circuits.

Future progress toward scalable SWCNT OCC-based quantum photonic technologies will depend on improvements in chirality-selective growth, site-specific defect functionalization, and strategies that stabilize emitters against environmental fluctuations. Integration with nanophotonic cavities, plasmonic antennas, and waveguides may enable enhancements in quantum yields, suppression of phonon-mediated dephasing, and improvements in photon indistinguishability. Despite current challenges, SWCNT OCCs remain unique among solid-state SPEs in simultaneously offering telecom-band emission, room-temperature operation, and chemical designability. These attributes position chemically engineered quantum defects as a competitive and highly tunable platform for emerging quantum communication systems and on-chip integrated photonic technologies.