Highly Luminescent Water-Dispersible NIR-Emitting Wurtzite CuInS2/ZnS Core/Shell Colloidal Quantum Dots.

Copper indium sulfide (CIS) quantum dots (QDs) are attractive as labels for biomedical imaging, since they have large absorption coefficients across a broad spectral range, size- and composition-tunable photoluminescence from the visible to the near-infrared, and low toxicity. However, the application of NIR-emitting CIS QDs is still hindered by large size and shape dispersions and low photoluminescence quantum yields (PLQYs). In this work, we develop an efficient pathway to synthesize highly luminescent NIR-emitting wurtzite CIS/ZnS QDs, starting from template Cu2-x S nanocrystals (NCs), which are converted by topotactic partial Cu+ for In3+ exchange into CIS NCs. These NCs are subsequently used as cores for the overgrowth of ZnS shells (≤1 nm thick). The CIS/ZnS core/shell QDs exhibit PL tunability from the first to the second NIR window (750-1100 nm), with PLQYs ranging from 75% (at 820 nm) to 25% (at 1050 nm), and can be readily transferred to water upon exchange of the native ligands for mercaptoundecanoic acid. The resulting water-dispersible CIS/ZnS QDs possess good colloidal stability over at least 6 months and PLQYs ranging from 39% (at 820 nm) to 6% (at 1050 nm). These PLQYs are superior to those of commonly available water-soluble NIR-fluorophores (dyes and QDs), making the hydrophilic CIS/ZnS QDs developed in this work promising candidates for further application as NIR emitters in bioimaging. The hydrophobic CIS/ZnS QDs obtained immediately after the ZnS shelling are also attractive as fluorophores in luminescent solar concentrators.

Introduction

Colloidal semiconductor quantum dots (QDs) have attracted much attention as luminescent probes for bioimaging due to their outstanding optical properties, such as broad absorption spectra, narrow photoluminescence (PL), large absorption cross sections, high PL quantum yields (PLQYs), and high photostability, which make them superior to organic dyes and fluorescent proteins.[1] Moreover, their PL can be tuned throughout the visible to the near-infrared (NIR) spectral window by controlling their composition and size.[1] These properties have been translated into higher sensitivity, multiplexed detection (i.e., multiple PL colors upon single excitation wavelength), and longer observation times.[1] Colloidal QDs can also be used as multimodal probes, allowing two or more imaging techniques (e.g., MRI and optical) to be combined.[2] NIR-emitting QDs are of particular interest, since wavelengths in the first and second biological spectral windows (viz., 650–950 and 1000–1350 nm, respectively)[3] penetrate much deeper in tissue than visible light, while inducing negligible autofluorescence.[3−5] However, most currently used NIR-emitting QDs contain highly toxic elements, such as Cd, Pb, and As (e.g., CdTe,[6−8] CdSe/CdTe,[9] Cd3P2,[10] InAs,[11,12] PbS,[13−15] and PbSe[16]), which severely hinders their application as biolabels. The search for less toxic alternatives is therefore becoming an increasingly relevant topic.

Among the alternatives, copper indium sulfide (CIS) QDs are particularly promising, since they combine low toxicity[17−19] with large absorption coefficients across a broad spectral range and unparalleled PL tunability, spanning a spectral window that covers the PL tunability of CdSe (visible), InP, and CdTe/CdSe (visible and first NIR biological window), and PbS (second NIR biological window).[20] Nevertheless, to date high PLQYs (≥50%) have only been reported for CIS/ZnS and CIS/CdS core/shell QDs with core diameters below 4 nm and PL up to 750 nm.[17−19,21−24] Although luminescent CIS QDs larger than 4 nm have also been reported,[21,25−27] their size and shape dispersion is typically quite large due to the difficulty in balancing the reactivities of multiple precursors (In, Cu, S).[20] Partial topotactic Cu+ for In3+ cation exchange (CE) in template Cu2-S nanocrystals (NCs) has been recently established as an effective strategy to circumvent these limitations,[28,29] allowing the preparation of monodisperse luminescent CIS QDs and NCs of sizes and shapes that would not be easily attainable by direct routes.

Interestingly, CIS QDs and NCs obtained by CE adopt the hexagonal wurtzite (WZ) structure, instead of the cubic chalcopyrite (CP) structure typically observed for CIS QDs synthesized by direct routes.[28,29] This creates new opportunities to expand the spectral tunability of CIS QDs, since WZ CIS QDs emit at lower energies than their CP counterparts.[30] The PLQY of bare CIS QDs is however low (<5–10% for CP CIS[17−20,22−24,26,27,31] and <1% for WZ CIS[30]) due to nonradiative recombination at surface defects. Overgrowth of a zinc blende ZnS shell on CP CIS QDs has been shown to effectively passivate these surface defects, thereby increasing the PLQYs to values as high as 80%, while excellent stability is imparted with preservation of the inherently low toxicity of CIS QDs.[17−19,22,23,27] However, ZnS shelling protocols for WZ CIS QDs are still underdeveloped, as the highest PLQY reported to date for WZ CIS/ZnS QDs is only 1%.[30,32]

Therefore, in order to harness the potential of WZ CIS QDs as efficient NIR-emitters for bioimaging, we developed in this work a sequential approach in which template high-chalcocite Cu2-S NCs (4.5–8.1 nm in diameter, 10% size dispersion) are first converted into WZ CIS QDs with size and shape preservation by topotactic partial Cu+ for In3+ CE. The product CIS QDs are subsequently coated with a ZnS shell. This yields WZ CIS/ZnS core/shell QDs with PL tunable from 750 to 1100 nm and PLQYs as high as 75% (at 820 nm). Finally, the product WZ CIS/ZnS core/shell QDs are phase-transferred to water through exchange of the native ligands by mercaptoundecanoic acid, while relatively high PLQYs in the NIR spectral region (up to 39% at 810 nm) are preserved.

Experimental Section

Materials

Copper(I) acetate (CuAc, 97%), indium acetate [In(Ac)3, 99.99%], indium chloride (InCl3, 99.999%), indium nitrate hydrate [In(NO3)3·H2O, 99.99%], indium acetylacetonate [In(acac)3, 99.99%], 1-dodecanethiol (DDT, ≥ 98%), 1-octadecene (ODE, 90%), trioctylphosphine oxide (TOPO, 99%), trioctylphosphine (TOP, 90%), zinc stearate [Zn(St)2, 10–12% Zn basis], tris(2-carboxyethyl)phosphine hydrochloride solution (TCEP, 0.5 M, pH 7.0), tetramethylammonium hydroxide pentahydrate (TMAH, ≥ 97%), indocyanine green (ICG, United States Pharmacopeia Reference Standard), anhydrous toluene, dimethyl sulfoxide, methanol, and butanol were purchased from Sigma-Aldrich. ODE and TOPO were degassed at 120 °C for 3 h prior to use. All other reagents were used as received. The chemicals were weighed and handled inside a glovebox under N2, while the high-temperature reactions were carried out in standard Schlenk lines.

Synthesis of Cu2-S NC Templates

Colloidal Cu2-S NCs were synthesized following the method reported by Wang and co-workers,[33] with small modifications. In brief, 0.6 mmol of CuAc and 5.6 mmol of TOPO were added to 50 mL of ODE in a three-neck flask combined with a condenser and degassed at 100 °C for 1 h. Subsequently, the reaction flask was purged by N2 and the temperature was set to 210 °C. At 160 °C, 3 mL (12.5 mmol) of DDT was swiftly injected into the flask, causing the solution color to change from dark green to brown. After that, the Cu2-S NCs were allowed to grow at 210 °C for variable amounts of time (20–160 min) and the reaction mixture was cooled down naturally to room temperature. The crude products were washed using isometric butanol and methanol followed by centrifugation at 3000 rpm for 10 min. The washing step was repeated three times. Finally, the purified Cu2-S NCs were dissolved into 6 mL of anhydrous toluene to yield a stock solution. The concentration of this stock-solution is taken to be 0.1 mmol Cu/mL, assuming a 100% reaction yield and no purification losses.

Synthesis of CIS QDs by Partial Cu+ for In3+ CE in Template Cu2-S NCs at Room Temperature (Slow CE)

The room temperature cation exchange reactions were performed using an adaptation of the method reported by van der Stam et al.[28] Typically, 1 mL of the stock solution of purified Cu2-S NCs was diluted in 3 mL of toluene. Then, 0.1 mmol of In(NO3)3·H2O dissolved in 2 mL of methanol in the presence of variable amounts of TOP (0–300 μL; see the Supporting Information, Table S1) was added into the as-prepared Cu2-S NCs solution. The In:Cu molar ratio in the reaction mixtures was ∼1 for all NC sizes. It should be noted that this ratio is a lower limit estimate, since it assumes a 100% reaction yield and no purification losses in the synthesis of the template Cu2-S NCs described above. This assumption leads to an overestimation of the Cu concentration, especially for small NCs.[20] This implies that the CE reactions were carried out under In/Cu > 1. The reaction mixture was maintained at room temperature (19 ± 2 °C) for 3 days. The crude products were washed using the method described above. Finally, the purified CIS QDs were dissolved into 1 mL of toluene and stored in a glovebox under N2.

Synthesis of CIS QDs by Partial Cu+ for In3+ CE in Template Cu2-S NCs at High Temperature (Fast CE)

Fast CE reactions at high temperatures were performed using nearly stoichiometric “In–TOP” complexes to ensure that the CE proceeded as a direct place exchange reaction, as demonstrated by van der Stam et al.[29] Briefly, 1 mL of the stock solution of purified Cu2-S NCs (see above) was degassed under vacuum to remove toluene and then redispersed into a solution of 3 mL of ODE and 500 μL of DDT. Meanwhile, 0.1 mmol of an In source [InCl3, In(NO3)3·H2O, In(Ac)3, or In(acac)3], 40 μL (0.09 mmol) of TOP, and 3 mL of ODE were mixed and degassed at 125 °C for 1 h (these mixtures are often slightly turbid and translucent, but this does not have any observable impact on the outcome of the CE reaction). The final concentration of TOP was 13.8 mM (TOP/In = 0.9). After that, the reaction flask containing the In–TOP complex solution was refilled with dry N2 and kept at 125 °C. The Cu2-S NC solution was then injected into the In–TOP complex solution, and the mixture was maintained at 125 °C for 1 h with stirring and then cooled down to room temperature. The unreacted precursors were removed by centrifuging at 3000 rpm for 1 min. The supernatant was collected and purified by using the same washing procedure described above. Occasionally, especially for reactions involving small NCs, the solution became a gel during the washing, requiring the addition of a few drops of octylamine to redisperse the QDs and proceed with the washing cycles. Finally, the CIS QDs were dispersed into 1 mL of toluene and stored in a glovebox for further use.

Synthesis of CIS/ZnS Core/Shell QDs

ZnS shell overgrowth on CIS QDs was achieved by following a modification of the procedure reported by Li et al.[23] The ZnS precursor solution was prepared by dissolving 0.2 mmol of Zn(St)2 into 5 mL of ODE at 150 °C, followed by addition of 100 μL of a TOP-S solution (0.2 mmol of elemental sulfur in 100 μL of TOP). Subsequently, 1 mL of purified CIS QDs in toluene and 5 mL of ODE were mixed and degassed at room temperature for 60 min to remove the toluene. The CIS QDs solution in ODE was then heated to 210 °C under N2. When the temperature was stable, the ZnS precursor solution (5.1 mL) was added dropwise over 25 min. The reaction mixture was then kept at 210 °C for 60 min, after which the flask was cooled down to room temperature. The product NCs were purified by addition of acetone, followed by centrifugation at 3000 rpm for 10 min and redispersion in toluene. This washing cycle was repeated three times.

Phase Transfer of CIS/ZnS Core/Shell QDs into Water

The purified CIS/ZnS core/shell QDs were transferred into water by adapting a previously reported procedure.[34] Typically, 0.2 mmol of MUA was dispersed in a mixture of 3 mL of deionized water and 400 μL of 0.5 M TCEP, producing a white turbid suspension that became a clear solution upon addition of 0.5 M TMAH under vigorous stirring until pH 11.6 was reached. Meanwhile, 20 mg of CIS/ZnS QDs was dispersed into 4 mL of chloroform. This solution was mixed with the MUA solution, followed by stirring (1000 rpm) overnight at room temperature (∼20 °C). Subsequently, the turbid biphasic emulsion was centrifuged at 2500 rpm for 5 min. The supernatant (water phase containing the QDs, ∼4 mL) was collected in a Millipore centrifugal filter VWR (MWCO 30K) and centrifuged at 10000 rpm for 20 min. The CIS/ZnS core/shell QDs collected in the filter were redispersed in 1 mL of deionized water (pH 7) and stored in a refrigerator (4 °C).

Optical Spectroscopy

Samples for optical measurements were prepared by dissolving the NCs into 3 mL of anhydrous toluene in 10 mm path length sealed quartz cuvettes. Sample preparation was conducted in a glovebox under N2. Absorption spectra were measured using a double-beam PerkinElmer Lambda 950 UV/vis/NIR spectrometer. PL spectra were recorded on an Edinburgh Instruments FLS920 spectrofluorometer equipped with a 450 W Xe lamp as excitation source and double-grating monochromators for both the excitation and the emission. The emission grating was blazed at the NIR (1200 nm). A liquid N2 cooled Hamamatsu R5509-72 photomultiplier tube was used as detector. This setup allows measurements in the 750–1600 nm spectral range without grating or detector changes. The spectra were corrected for the instrumental response. PL decay curves were obtained by time-correlated single-photon counting via time-to-amplitude conversion using a liquid N2 cooled Hamamatsu R5509-72 photomultiplier tube. A pulsed diode laser (EPL-655 Edinburgh Instruments, 656.6 nm, 62 ps pulse width, 0.5 MHz repetition rate) was used as the excitation source.

Photoluminescence Quantum Yields (PLQYs)

The PLQYs were measured using indocyanine green (ICG, PLQY = 12%) in DMSO as a standard[35] (see the Supporting Information, Supporting Method and Figure S1 for details). To avoid inner filter effects, the absorbances of the QD and ICG solutions at and above the excitation wavelength (678 nm) were below 0.1.

Photostability Test

The photostability of a selected CIS/ZnS core/shell QD sample was evaluated using the spectrofluorometer described above. A sample of a NIR-emitting dye (ICQ) was also investigated for comparison. The samples (ICQ dye in DMSO, QDs in toluene) were loaded in 10 mm path length sealed quartz cuvettes. The absorbances of the QD and ICG solutions at 400 nm were below 0.1. The samples were continuously illuminated at 400 nm and their PL intensity was recorded every 5 s over a period of 12 h, while all the instrumental conditions were kept constant. The conditions used for the dye and the QD samples were identical.

Transmission Electron Microscopy (TEM)

TEM measurements were performed using a FEI Tecnai-12 microscope operating at 120 kV. Samples for TEM imaging were prepared by drop-casting a toluene solution of purified NCs onto a carbon-coated 200 mesh copper TEM grid. The excess liquid was removed by blotting using filter paper.

High-Resolution TEM (HRTEM) and Energy-Dispersive X-ray Spectroscopy (EDS)

HRTEM and EDS measurements were carried out on a FEI Talos F200X microscope operating at 200 kV. The lattice spacing of NCs was calculated by Fourier transform (FT) from the area of interest in HRTEM images. EDS measurements were performed using a dedicated low-background holder and Cu-free 300 mesh aluminum or nickel TEM grids with an acquisition time of ∼60 s. To ensure that the elemental concentrations were statistically valid and representative of the whole NC ensemble, EDS analyses were performed on wide areas (∼104–105 nm2), encompassing several hundreds of NCs, and were repeated for three different areas on the TEM grid. The elemental ratios provided in the discussion below were obtained by averaging over the different measurement spots. Samples for HRTEM measurements were prepared by drop-casting a toluene solution of purified NCs onto a superthin carbon-coated 300 mesh copper TEM grid.

X-ray Diffraction (XRD)

XRD patterns were obtained on Bruker D2 Phaser, equipped with a Co Kα X-ray source (1.790 26 Å). Samples were washed at least three times, dried under vacuum overnight, and uniformly dispersed on a Si wafer immediately prior the XRD measurements.

Fourier Transform Infrared Spectroscopy (FTIR)

FTIR spectra were measured from 400 to 4000 cm–1 (2 cm–1 resolution) for 400 scans using a vertex 70 FTIR spectrometer (BRUKER) equipped with KBr/DLaTGS D301 detector. Approximately 0.1–1.0 wt % purified sample (powder) was well-dispersed into 200–250 mg of fine KBr powder, followed by grinding under an infrared lamp, and pressed into a pellet.

ζ-Potential and Dynamic Light Scattering (DLS)

ζ-Potential and DLS measurements were performed on a Malvern instrument Zetasizer Nano ZS using a DTS1070 folded capillary cell. The samples were thoroughly purified by filtration through centrifugal filters from VWR (MWCO 30K) and dispersed in deionized water (pH ∼7). To prevent accidental contamination by dust, this solution was filtered through Millex syringe filter units (pore size 0.2 μm) immediately prior to the DLS measurement. To obtain reliable results, 10 runs were operated with a measurement angle of 173°. The spectra were corrected by the instrument software for viscosity (0.8872 cP at 25 °C), absorption (at 532 nm), solvent (water), refractive index (1.33), and material (CIS) refractive index (2.55). The hydrodynamic sizes are collected in automatic mode and expressed in number percent. The same settings were used to measure the ζ-potential.

Results and Discussion

Size-Controlled Nearly Spherical Cu2-S NC Templates

The method used in this work is very versatile,[33] allowing the diameter of nearly spherical Cu2-S NCs to be tuned from 3 to 8 nm (10% size dispersion) by controlling the reaction time from 20 to 160 min [Supporting Information (SI), Figure S2]. As an illustrative example, Figure a shows the HRTEM image of 6 nm Cu2-S NCs obtained after 90 min at 210 °C. The d-spacing of these NCs is 3.4 Å, which can be assigned to the (002) lattice planes of hexagonal high-chalcocite Cu2S (JCPDS Card 26-1116). This assignment is also supported by the XRD pattern of the sample (SI, Figure S3).

Figure 1

(a) High-resolution transmission electron microscopy (HRTEM) image of 6 nm diameter Cu2-S NCs. The upper panels show the Fourier transform (FT) pattern (left) and a close-up (right) of the NC indicated by the yellow circle. (b) HRTEM image of CIS QDs obtained from the template Cu2-S NCs shown in part a by partial Cu+ for In3+ cation exchange at 20 °C and TOP/Cu = 0.9 (slow CE). The upper panels show the FT pattern (left) and a close-up (right) of the NC indicated by the red circle. The average NC size and polydispersity (∼10%) of the template NCs are preserved in the product QDs (within the uncertainty of the measurement).

Size-Controlled Nearly Spherical Luminescent CIS QDs by Cation Exchange at 20 °C (Slow CE)

Previous work by our group[28] demonstrated that CIS QDs can be obtained by topotactic partial Cu+ for In3+ CE in template Cu2-S NCs at room temperature using TOP as the Cu-extracting agent and an In salt as the In3+ source. Under the conditions used in that study (excess TOP), the Cu+ extraction and In3+ incorporation were shown to proceed by separate chemical pathways, making the CE reaction subject to a very delicate balance between the Cu+-extraction and In3+-incorporation rates.[28] This strict balance was achieved in ref (28) for two different template NC sizes (2.5 and 4 nm) by carrying out the CE reaction at low temperatures (room temperature) and low concentrations. These conditions were however not sufficient to provide a precise balance for larger template sizes (viz., 11 nm), for which a slight deterioration of the size distribution was observed.[28] Under excess TOP, higher temperatures make the balance between the Cu+-extraction and In3+-incorporation rates even harder to achieve and have been shown to lead to hollow CuInS2 nanoplatelets (for template 10 × 50 nm Cu2-S nanoplatelets at 150 °C and using InCl3 as In source)[36] or size reduction (for template 8 nm Cu2-S NCs at 120 °C).[32] A strategy to circumvent these difficulties and render the Cu+ for In3+ CE in Cu2-S NCs more amenable to control has recently been developed by our group and is based on the use of stoichiometric TOP–In complexes as both In source and Cu-extracting agent at high temperatures (100 °C).[29] In this way, the Cu+ for In3+ exchange becomes a direct place exchange reaction, and therefore, the extraction and incorporation rates become inherently coupled.[29]

In the present work, we intend to expand the applicability of this strategy to a wider range of template NC sizes and reaction temperatures. To this end, we first investigated the influence of the TOP/In ratio on the outcome of the partial Cu+ for In3+ CE under slow reaction conditions (20 °C; SI, Table S1 and Figure S4). Cu2-S NCs with 6 nm diameter were chosen as templates, because they are sufficiently large to be readily analyzed with TEM while still yielding CIS QDs with optical properties in a relevant spectral window. It is clear that TOP/In ratios beyond 1.8 induce partial dissolution of the NCs, as evidenced by the steady decrease of the absorbance of the product CIS QDs solutions with increasing TOP concentrations (SI, Figure S4a). This can be attributed to overextraction of Cu+ from the NCs as a consequence of the excess of free TOP for TOP/In ratios larger than 2. Interestingly, the PL intensity of the product CIS QDs decreases dramatically for TOP/In ratios beyond 0.9 (SI, Figure S4b), although the absorption intensity remains constant up to a TOP/In ratio of 1.8. This implies that even a modest excess of free TOP is detrimental to the quality of the product CIS QDs, suggesting that accelerated Cu+ extraction rates induce higher concentration of defects. These observations are consistent with the direct place exchange mechanism proposed in ref (29) and show that the TOP/In ratio is a crucial parameter to control the Cu+ for In3+ CE reaction rates.

The success of the slow partial Cu+ for In3+ CE in template Cu2-S NCs under TOP/In ∼ 1 is evidenced not only by the higher PL intensities of the product CIS QDs (SI, Figure S4b) but also by the preservation of the size and shape of the template NCs after the CE reaction (Figure and SI, Figure S5; the average size and polydispersity remain constant within the uncertainty of the measurement). The d-spacings (3.4 and 2.0 Å) observed in the HRTEM image (Figure b) are consistent with the (100) and (110) lattice planes of the wurtzite CIS crystal structure viewed along the [001] direction. This assignment is confirmed by the XRD pattern of the CIS QDs (SI, Figure S6). Moreover, the EDS spectrum of the product QDs (see SI, Figure S7 for a representative example) confirms the incorporation of In and reveals a Cu/In ratio of 1.4 ± 0.2. It should be noted that these observations are fully consistent with the cation exchange mechanism proposed in our previous work (i.e., the partial Cu+ for In3+ CE in hexagonal high-chalcocite Cu2-S NCs is self-limited and topotactic, thereby leading to wurtzite CIS).[28,29]

Size-Controlled Nearly Spherical Luminescent CIS QDs by Cation Exchange at 125 °C (Fast CE)

Despite its success, the room temperature Cu+ for In3+ CE protocol is very time-consuming (the reaction takes 3 days) and is thus not well suited for the preparation of a large number of different samples. To circumvent this limitation, we have investigated a faster CE protocol, in which a higher reaction temperature (125 °C) is used while the TOP/In is kept close to stoichiometric (0.9). In this way, the Cu+-extraction and In3+-incorporation rates are expected to remain coupled, despite accelerated reaction kinetics, as demonstrated in ref (29) for CE reactions carried out at 100 °C. Previous work by Buhro and co-workers has shown that the reactivity of the In source has a dramatic impact on the balance between the extraction and incorporation rates of Cu+ for In3+ CE reactions carried out under TOP excess at high temperatures (e.g., InCl3 at 150 °C produces hollow CIS nanoplatelets, whereas In(Ac)3 at the same temperature yields intact CIS nanoplatelets).[36] We have therefore investigated the influence of the In source on the outcome of the Cu+ for In3+ exchange in template Cu2-S NCs at 125 °C under stoichiometric TOP/In ratios (SI, Figures S8 and S9). In all cases, the average size and polydispersity of the template Cu2-S NCs are well-preserved in the product CIS QDs, with minor differences between the four different In precursors (SI, Figure S8). However, the product CIS QDs prepared by using In(Ac)3 as the In precursor show the highest PL intensity (SI, Figure S9), compared to InCl3, In(NO3)3, and In(acac)3, suggesting that the cation exchange reaction between the TOP–In(Ac)3 complex and the template Cu2-S NCs leads to fewer nonradiative recombination centers. In(Ac)3 was thus selected as the In source for further optimization of the CE reaction conditions. It is possible that the lower activation energies for incorporation of In3+ from In(Ac)3, as observed by Buhro and co-workers,[36] result in a better balance between the Cu+-extraction and the In3+-incorporation rates, thereby leaving fewer nonradiative defects or, alternatively, more emissive defects.[30] The possible nature of these defects and their impact on the optical properties of CIS QDs will be discussed in more detail later in this paper.

Figure provides a representative example of the CIS QDs produced by partial Cu+ for In3+ exchange in template Cu2-S NCs at 125 °C for 1 h, using the TOP–In(Ac)3 complex as the In source and under nearly stoichiometric conditions (TOP/In = 0.9). The results show that the CE reaction proceeds topotactically, since the average size, shape, and polydispersity of the template Cu2-S NCs are inherited by the product CIS QDs (Figure ). The topotactic nature of the partial Cu+ for In3+ CE reaction is also evidenced by HRTEM (Figure ) and XRD (SI, Figure S10) analysis, which demonstrate that the product CIS QDs are highly crystalline and have the WZ structure, confirming that the hexagonal anionic sublattice of high-chalcocite Cu2S remains largely undisturbed by the CE reaction.[28,29] The fast CE protocol developed here is also highly versatile and was successfully used to obtain WZ CIS QDs in the 3–8 nm size range (polydispersity of ∼10%) in just 1 h reaction time. The optical properties of these NIR-emitting CIS QDS will be discussed in more detail below, together with those of the CIS/ZnS core/shell QDs obtained by overcoating them with ZnS shells.

Figure 2

(a) TEM image and corresponding size histogram of 6.6 nm Cu2-S NC templates. (b) TEM image and corresponding size histogram of 6.7 nm product CIS QDs obtained from the template Cu2-S NCs shown in part a by partial Cu+ for In3+ cation exchange at 125 °C for 1 h. (c) HRTEM image and (d) FT analysis of a single CIS QD from the same sample shown in part b. The FT pattern can be indexed to the axial projection of the WZ CIS structure along the [001] direction.

The high-temperature CE protocol can be successfully carried out also at different temperatures, such as 100 °C (reaction time of 2 h) or 150 °C (reaction time of 30 min). It should be noted that residual TOPO is very detrimental to the CE reaction, deteriorating the size polydispersity (SI, Figure S11). It is therefore of crucial importance to properly wash the template Cu2-S NCs prior to use. The fast CE method is advantageous over its room temperature equivalent not only for its much shorter reaction times but also because it yields WZ CIS QDs with superior PLQYs (1–2%, which are about 1 order of magnitude higher than those obtained by slow CE). To further increase the PLQYs of the CIS QDs prepared at 125 °C, DDT was added to the reaction medium. This enhanced the PLQYs of the as-prepared CIS QDs by a factor 5–10 and improved their colloidal stability, without any observable impact on their size and polydispersity. The fact that the size, shape, and polydispersity of the product CIS QDs are not affected by the addition of DDT implies that the beneficial effects of DDT are solely due to its ability to bind to the surface of the QDs. Although DDT is often used as sulfur source in the synthesis of both Cu2-S (e.g., this work) and CIS (e.g., refs (23) and (31)) NCs, the temperatures required for the thermolysis of the C–S bond (typically above 200 °C for CIS QDs) are higher than those used in our fast CE protocol. DDT is also a well-known soft ligand (chemical hardness η ∼ 6 eV),[37] which strongly binds to soft metal cations such as Cu+ (η = 6.28 eV)37 through its S-donor atom, thereby passivating surface dangling orbitals.[38,39] The enhancement of the PLQYs of WZ CIS QDs by DDT has also been observed by postsynthetic ligand exchange and was attributed to the passivation of surface Cu+ sites.[30] However, we note that although In3+ is a hard Lewis acid (η = 13 eV),[37] we cannot exclude that in the present case DDT may also bind to surface In sites, albeit weakly, since harder ligands are not available and our CE experiments clearly show that TOP (also a soft Lewis base with η ∼ 6 eV)[37] does form complexes with In compounds, even at nearly stoichiometric ratios. The enhancement of the PLQYs upon DDT addition can thus be interpreted as evidence that carrier trapping at surface metal dangling orbitals is an important nonradiative decay pathway. The possible nature of the nonradiative and radiative recombination processes in WZ CIS QDs will be discussed in more detail later. Interestingly, the PLQYs observed in the present work for the DDT-capped CIS QDs obtained by the fast CE method (viz., 20–1% in the 5 to 8 nm size range) are the highest reported to date for WZ CIS QDs and are comparable to the best values reported for CP CIS QDs in the 3–5 nm size range (5–10%).[20] This implies that the PLQYs of WZ CIS QDs are not fundamentally lower than those of CP CIS QDs, in contrast to suggestions made in previous works.[30,40]

ZnS Shell Overgrowth on CIS QDs

In order to enhance the PLQYs and (photo)chemical stability of the WZ CIS QDs prepared by both the slow and the fast CE protocols, a ZnS-shelling protocol was developed. This procedure yielded nearly spherical WZ CIS/ZnS core/shell QDs with shell thicknesses of about 0.5 nm (∼1.5 ZnS monolayer) (Figure ). As expected for heteroepitaxial growth, the ZnS shell adopted the WZ structure of the CIS cores, as clearly shown by both HRTEM (Figure e–h) and XRD (SI, Figures S12 and S13). The presence of Zn in the product CIS/ZnS QDs is confirmed by EDS measurements (SI, Figures S14 and S15). The elemental ratios obtained by quantification of the EDS spectra are roughly consistent with the ZnS shell thicknesses estimated from the TEM measurements and indicate that the Cu/In ratios do not significantly change after the ZnS overgrowth, implying that Cu+ (and/or In3+) for Zn2+ cation exchange during the ZnS shell overgrowth was negligible. In combination with the changes observed in the optical spectra of the product CIS/ZnS QDs with respect to those of the CIS QDs used as cores (Figure b and SI, Figures S16–S18 and Table S2; see discussion below for details), these results suggest that the WZ CIS/ZnS QDs prepared in the present work possess a core/shell structure with an alloyed heterointerface.

Figure 3

TEM images of CIS QDs obtained by (a) slow and (b) fast CE. (c, d) TEM images of CIS/ZnS core/shell QDs obtained by ZnS overgrowth on the CIS QDs shown, respectively, in parts a and b. (e, f) HRTEM images and (g, h) FT analysis of individual CIS/ZnS core/shell QDs selected from parts c and d, respectively. The lattice spacings in parts e and f are 3.4 and 1.95 Å, which correspond well to the {100} and {110} lattice planes of WZ CIS. The FT patterns (g, h) can be indexed to the [001] axial projection of the WZ structure.

Figure 4

(a) Absorption and PL spectra of CIS QDs prepared by fast CE using different sizes (4.5–8.1 nm) of template Cu2-S NCs. (b) Absorption and PL spectra of 5.4 nm CIS QDs obtained by fast CE before and after ZnS shell overgrowth. (c) PL spectra of CIS/ZnS core/shell QDs using different sizes (4.5–8.1 nm) of CIS QDs as cores. (d) PLQYs of CIS QDs obtained by fast CE (black line) and the corresponding CIS/ZnS core/shell QDs (red line) as a function of the QD core size. (e) Photodegradation of CIS/ZnS core/shell QDs (core diameter of 4.5 nm, blue line) and indocyanine green (orange line) under 400 nm illumination.

Optical Properties of WZ CIS and CIS/ZnS QDs

The absorption and PL spectra of the WZ CIS/ZnS QDs prepared in this work are blue-shifted with respect to those of the WZ CIS QDs used as cores (Figure b and SI, Figures S16–S18 and Table S2). The extent of the blue-shift appears to be size-dependent for the CIS/ZnS QDs obtained from fast CE CIS QDs, decreasing with increasing core diameter (SI, Table S2). In contrast, there is no apparent trend for CIS/ZnS QDs obtained by shelling slow CE CIS QDs (SI, Table S2). The shifts observed in the PL spectra were in most cases not significantly different from those observed in the absorption spectra (see SI, Table S2; the difference varies from +20% to −20% depending on the sample) but occasionally deviated from the blue-shift trend observed for the majority of the samples (see, e.g., SI, Figure S17, which shows a red-shift of 176 meV, despite a blue-shift of 63 meV in the absorption spectrum). This anomalous PL red-shift can be understood by considering that the PLQY of the CIS QDs used as cores was very low (0.2%) and, therefore, was not representative of the QD ensemble, probably originating from a small fraction of smaller or otherwise different CIS QDs emitting at higher energies. The dramatic enhancement of the PLQYs induced by the ZnS shelling (2 orders of magnitude; see SI, Table S2) resulted in a PL spectrum that was representative of the ensemble of CIS/ZnS QDs, thereby yielding an apparent red-shift with respect to the very weak and likely biased PL spectrum of the bare CIS QDs used as cores.

This observation implies that one should exercise caution when interpreting spectral shifts derived from comparison of PL spectra prior to and after shelling procedures, since the PLQYs of bare CIS QDs are generally low and therefore the observed PL spectrum may not necessarily reflect the size, shape, and composition polydispersity of the ensemble. The comparison between absorption spectra would in principle be more reliable, since they are determined by the whole QD ensemble. However, absorption spectra may be distorted by the presence of absorbing impurities. Moreover, the band-edge transitions of QDs are strongly size dependent, and therefore, the band gap estimated from absorption spectra is affected by the size polydispersity of the sample.[41] This is a particularly serious issue for CIS and other ternary copper chalcogenides, which typically present essentially featureless absorption spectra without a sharp first absorption transition and often accompanied by a low-energy tail.[20] The difficulties associated with extracting reliable band gap values from the absorption spectra of ternary chalcogenides may be the reason why PL shifts are more frequently reported in the literature than absorption shifts. It should also be noted that these peculiarities may introduce a considerable uncertainty in the spectral shifts deduced from both PL and absorption spectra and may explain the discrepancies between the observations reported in the literature, as we will discuss below.

Spectral blue-shifts have been invariably observed after ZnS shelling of CP CIS QDs and have been attributed to a number of reasons. For example, Li and co-workers reported a 80 meV blue-shift in the PL spectra of 3 nm CP CIS QDs upon overgrowth of a ZnS shell using both Zn and S precursors and attributed it to etching of the cores prior to shell growth.[23] A similar PL blue-shift (90 meV), accompanied by a 62 meV blue-shift in absorption spectrum, has been observed by Berends et al. for 2.5 nm CP CIS QDs overcoated with one monolayer of ZnS.[31] Exposing CP CIS QDs to Zn2+ precursors in the absence of a S-precursor has been observed to induce even larger blue-shifts (as large as 340 meV for a 3.3 nm QD and 200 meV for a 2.4 nm QD) in both the PL and absorption spectra.[42,43] These pronounced blue-shifts were shown to be due to partial Zn2+ for In3+ and Cu+ CE followed by interdiffusion, resulting in gradient (CuInZn)S2 alloy QDs with the same size as the parent CP CIS QDs but larger band gaps.[42] Simultaneous ZnS overgrowth and Zn interdiffusion, resulting in core/shell QDs with a gradient CP (CuInZn)S2 alloy core overcoated by a ZB ZnS shell, has also been shown to occur upon addition of mixed Zn and S precursors to CP CIS QDs (3–5 nm diameter), leading to spectral blue-shifts in both absorption and emission.[43,44] Intriguingly, the PL shifts reported in ref (44) (viz., 100–180 meV, in the 5–3 nm size range) are significantly smaller than those observed in the absorption spectra (160–260 meV, in the same size range). There are also works in which only the PL is observed to blue-shift upon ZnS shelling of CP CIS QDs, while the absorption spectra remain unchanged (e.g., 64 meV for CP/ZB CIS/ZnS QDs with 3 nm cores).[18]

From the above, it is clear that shelling of CP CIS QDs with ZnS leads to spectral blue-shifts and that, despite some dissonant works, the shifts in absorption and photoluminescence are comparable. The remarkable diversity of the trends observed for CP CIS/ZnS QDs demonstrates the chemical and electronic complexity of this system and may be rationalized by considering that the magnitude of the blue-shift depends on the extent of the Zn interdiffusion and alloying into the CIS core, being the largest (e.g., 340 meV)[42] for extensively alloyed CIS/ZnS QDs produced by exposure of CP CIS QDs to only Zn precursors and the smallest (e.g., 60 meV)[31] for CIS/ZnS core/shell QDs with partially alloyed heterointerfaces obtained by shelling CP CIS QDS using both Zn and S precursors. This behavior is reminiscent of that observed for heteronanocrystals (HNCs) of II–VI semiconductors (e.g., ZnS–CdS[38] and ZnSe–CdSe[45]), the elemental distribution profiles of which can be seamlessly tuned from core/shell QDs with a well-defined heterointerface to fully homogeneous alloy QDs, through gradient alloy QDs with increasingly smoother gradients.[45] As a result, the optical properties of HNCs of II–VI semiconductors can be continuously tuned from those of core/shell HNCs with sharp heterointerfaces to those of homogeneous alloy NCs, with preservation of the total volume and composition of the NC.[45] However, it appears that CP CIS/ZnS core/shell QDs with sharp heterointerfaces have yet to be made, since to date only spectral blue-shifts have been reported for these materials (see above),[18,23,31,40,42−44] in striking contrast with the small red-shifts expected for type-I core/shell QDs as a result of the small leakage of the exciton wave function into the shell due to the finite band offsets between the core and shell.[38] Some authors have assumed the formation of a sharp CIS/ZnS heterointerface, attributing the blue-shift observed after ZnS shelling to shrinkage of the core, due to either etching prior to the shell overgrowth[23] or shell ingrowth by CE.[43] However, this assumption has yet to be experimentally validated.

Reports on spectral shifts upon ZnS shelling of WZ CIS QDs are scarce and contradictory,[30,32] since these QDs have only recently been developed.[26,28−30,32] While Leach et al.[30] did not observe any spectral shifts (both in PL and absorption) upon exposure of ∼10 nm diameter WZ CIS platelets to Zn2+, Akkerman et al.[32] reported a 115 meV blue-shift in the PL spectrum of 8 nm diameter gradient alloy WZ CIS/ZnS QDs, without any significant change in the absorption spectrum. These reports are also inconsistent with the behavior observed in the present work (Figure b and SI, Figures S16–S18 and Table S2), which unambiguously shows that the absorption and PL spectra of WZ CIS/ZnS QDs blue-shift with respect to those of the WZ CIS QDs used as cores. This behavior is analogous to that previously reported for CP CIS/ZnS QDs (see above) and can therefore be interpreted as evidence of a small degree of Zn2+ diffusion into the CIS cores, leading to an alloyed CIS/ZnS heterointerface.

The blue-shift in the optical transitions of CIS QDs (both CP and WZ) upon ZnS shelling can be attributed to widening of the band gap as a result of both the partial CIS–ZnS alloying (the band gap of ZnS is larger than that of CIS, viz., 3.5 and 1.53 eV,[38,20] respectively) and the reduction of the effective volume probed by the exciton wave function due to the gradient elemental distribution profiles (Zn concentration increases toward the surface, which translates into an increasing potential barrier for both electron and hole). The increase in band gap will be directly reflected in the absorption spectrum and will be most evident in the lowest energy transition, which is assigned to the 1Se–1Sh transition.[46] The impact of the ZnS shelling on the PL energies of CIS QDs is likely more complex due to the nature of the radiative recombination in these materials.

The PL of CP CIS QDs is characterized by broad bandwidths [full width at half maximum (fwhm) ∼200–300 meV], large “global” Stokes shifts (∼300–400 meV), and multiexponential PL decays with long decay constants (slow component with hundreds of ns).[20,40] These characteristics are dramatically different from those of the 1Sh–1Se band-edge PL observed for the prototypical II–VI and IV–VI QDs (e.g., CdSe and PbSe)[20] and have led to an intense debate concerning their origin.[20,40] Most works attribute the PL in CP CIS to radiative recombination involving defects, but a wide variety of defects and recombination mechanisms have been invoked:[40] donor–acceptor pair, localized electron–valence band (VB) hole, or conduction band (CB) electron-localized hole. An intrinsic recombination mechanism, which ascribes the PL to the recombination of CB electron states with dark and bright VB hole states, has also been proposed.[46] However, recent work has unambiguously demonstrated that the PL in CP CIS QDs originates from the radiative recombination of a delocalized CB electron with a hole localized on a Cu+ ion.[31,47,48] Nevertheless, the nature of the hole localization process (i.e., self-trapping onto a regular Cu+ ion[48] or capture by a native defect, such as CuIn2–) has yet to be unravelled.[31]

The PL of WZ CIS QDs has also been assigned to radiative recombination at defects, but the nature of these defects is thought to be different from those in CP CIS QDs, since WZ CIS NCs typically emit at lower energies than their CP counterparts, despite showing otherwise similar PL characteristics (i.e., fwhm of ∼200–300 meV, “global” Stokes shifts of ∼300–400 meV, and multiexponential PL decays with long decay constants of hundreds of nanoseconds).[30,40] Recent work by Macdonald and co-workers[30] has used DFT calculations to shed light on the radiative recombination mechanism in WZ CIS QDs and concluded that the recombination of an electron localized at Ini2+ with a VB hole was the most likely radiative recombination pathway, since it gave the best match to their experimental PL spectra. However, the authors noted that, among the candidate defects that were studied, the transition energies associated with CuIn and VIn3– also agreed well with their experimental data.[30] This implies that the PL in WZ CIS QDs may also originate from recombination of a CB electron with a localized hole (either in CuIn or VIn3–).

We argue that, considering the strong similarity between the PL characteristics of CP and WZ CIS QDs, recombination of a CB electron with a localized hole is in fact the most plausible mechanism. However, WZ CIS NCs have only recently become available and have yet to reach the same level of study of their CP counterparts. Therefore, a definitive assignment of the radiative decay mechanism in WZ CIS NCs would be premature at this time. Nevertheless, the pronounced size-dependence observed in the optical spectra of the WZ CIS QDs and WZ CIS/ZnS core/shell QDs prepared in our work (Figure a) clearly demonstrates that the radiative recombination involves at least one delocalized carrier. Considering that the CB potential in CIS is more strongly affected by quantum confinement than the VB potential (effective masses of the electron and hole are, respectively, 0.16 m0 and 0.85 m0),[46] we argue that this also suggests that the delocalized carrier is likely the electron. Regardless of the nature of the delocalized carrier, the size dependence of the PL energy allowed us to tune the PL of the WZ CIS QDs from the first to the second NIR biological window (from ∼800 to ∼1050 nm, Figure c) by increasing the QD size from 4.5 to 8.1 nm. PL at shorter wavelengths (650–750 nm) can also be obtained, by using smaller template Cu2-S NCs (2.5–3.5 nm).[28] The observation of efficient PL at 1050 nm is particularly interesting, since reports on CIS QDs emitting in the second NIR window are scarce.[26,30,32]

Overcoating by a ZnS shell leads to a large enhancement of the PLQY for all the investigated sizes (Figure b,d), which is accompanied by an increase of the average exciton lifetimes (SI, Figure S19). This is consistent with previous reports for both CP and WZ CIS QDs[20,23,28,30,31] and indicates that recombination through surface defects is a major nonradiative decay pathway.[23,30,31] However, we also observe that the PLQYs are size-dependent, decreasing for increasing QD size (Figure d), and that the increase of the PLQY upon ZnS shelling is apparently limited by the initial PLQY of the CIS QD cores (i.e., CIS QD cores with lower PLQYs produce CIS/ZnS core/shell QDs with relatively lower PLQYs as well). This suggests that internal defects are also important charge carrier traps, generating additional nonradiative relaxation pathways that cannot be eliminated by shell overcoating. A strategy to further improve the PLQYs of NIR-emitting CIS/ZnS core/shell QDs should thus not only address the ZnS shell thickness and quality but also the underlying reasons for the low PLQYs of the CIS QD cores themselves, which requires a deeper understanding of the radiative and nonradiative decay pathways in WZ CIS QDs. Nevertheless, we note that the PLQYs obtained in our work using CIS QDs produced by fast CE as cores for CIS/ZnS core/shell QDs (e.g., 75% at 820 nm) are, to the best of our knowledge, the highest reported so far for NIR-emitting CIS/ZnS QDs (Speranskaya et al.[19] reported 40% at 750 nm, while Pons et al.[18] reported 30% at 800 nm and 12% at 820 nm). A comparative photodegradation test (Figure e) shows that the photostability of the CIS/ZnS QDs is superior to that of commonly available NIR-emitting dyes (e.g., ICQ).

Phase Transfer of WZ CIS/ZnS Core/Shell QDs into water

The WZ CIS/ZnS core/shell QDs were transferred into water through exchange of the native ligands by 11-mercaptoundecanoic acid (MUA).[34] As the binding energy of Zn–Sthiolate (194.7 kJ/mol) is much higher than those of S–Sthiolate (105.1 kJ/mol) and Zn–Sthiol (31.8 kJ/mol),[34,50] it is of great importance to control the solution pH (>8), so as to deprotonate the thiol group, thereby ensuring a strong bond between the thiolate headgroup and the surface of the CIS/ZnS QDs. However, the high pH will also give rise to the formation of disulfide bonds that may trap the photogenerated hole, resulting in PL quenching.[34] To prevent disulfide formation, reducing agents (e.g., sodium borohydride,[51] dithiothreitol,[52] and TCEP[34]) are commonly added during the phase transfer. In this work, tris(2-carboxyethyl)phosphine hydrochloride (TCEP) was chosen as the reducing agent due to its high efficiency and low contamination during biolabeling.[53] The phase transfer was accomplished at pH ∼11.6, which resulted in a high concentration of deprotonated MUA thiolate groups that exchanged the native ligands at the surface of CIS/ZnS core/shell QDs, thereby making them hydrophilic (Figure ). The average ζ-potential of these water-dispersible CIS/ZnS core/shell QDs is about −60 mV (SI, Figure S20). The negatively charged surface is attributed to the deprotonation of the carboxyl groups of MUA. This assignment is confirmed by the weak stretching mode of the COOH group at 1720 cm–1 and the appearance of a new band at 1398 cm–1, which is related to the symmetric vibration modes of the COO– group (SI, Figure S20).

Figure 5

Schematic illustration of the phase transfer of CIS/ZnS core/shell QDs by ligand exchange using MUA in the presence of TCEP at pH ∼11.6.

The optical properties and morphology of ∼5 nm diameter WZ CIS/ZnS core/shell QDs before and after phase-transfer to water are shown in Figure . The average size and polydispersity of the QDs are not significantly affected by the phase-transfer procedure (Figure b), and the resulting negatively charged QDs are colloidally stable in water for long periods of time (at least 6 months). The hydrodynamic size of the CIS/ZnS core/shell QDs is 14.2 ± 2.7 nm (SI, Figure S20). The absorption spectra of the QDs before and after the transfer are indistinguishable (Figure a), indicating that they remain intact and that no aggregates are formed, which is also clearly demonstrated by the TEM images (Figure b and SI, Figure S21). The PLQY decreases from 75% to 39% after phase-transfer (Figure ). The PL spectrum also remains essentially the same, apart from a small blue-shift (19 meV) from 820 to 810 nm. The origin of this blue-shift is not yet understood but may be related to the effect of the negative surface charges on the amplitude of the electron wave function near the QD surface (assuming that the hole is localized within the core, as discussed above).

Figure 6

(a) Absorption and PL spectra of ∼5 nm CIS/ZnS core/shell QDs before and after phase transfer into water. (b) TEM images and size histograms of CIS/ZnS core/shell QDs before (left) and after (right) ligand exchange (larger area TEM images are given in the SI, Figure S21).

A decrease in PLQY from 26% to 6% upon phase-transfer to water was also observed for 9.2 nm CIS/ZnS core/shell QDs emitting at 1050 nm (SI, Figure S22). Reductions in PLQY upon transfer of core/shell QDs to water are commonly observed[17,18,54] and are attributed to insufficient shell quality in part of the QD ensemble.[54] In the present case, the drop in the PLQY of the CIS/ZnS core/shell QDs in water may be interpreted as an indication that the ZnS shell in some of the QDs is not yet sufficiently robust to withstand the high pHs required for the ligand exchange. Therefore, it is likely that further improvements in the ZnS shell thickness and quality will lead to higher PLQYs after the phase transfer. Nevertheless, it should be noted that the PLQYs observed in the present work (viz., 39% at 810 nm and 6% at 1030 nm) are rather high for NIR emitters in water and are very competitive with respect to the currently available NIR-emitting QDs and dyes (CIS/ZnS QDs, 20% at 800 nm,[18] 8% at 830 nm;[17] indocyanine green, 12% at 835 nm in DMSO;[35] IR-140, 17% at 830 nm;[55] ONITCP, 2.3% at 900 nm;[55] ODNITCP, 1.4% at 930 nm[55]). It is worth noting that the stability of the hydrophilic CIS/ZnS core/shell QDs in water is also competitive, since they can be stored in the dark at 4 °C for at least 6 months with negligible PL loss. Moreover, these QDs are carboxyl-terminated and can thus be easily functionalized with specific proteins or biomacromolecules,[8,12,17,18,51,56−58] which makes them promising as NIR-emitting labels for bioimaging.

Conclusions

In conclusion, we have developed a sequential procedure to synthesize highly luminescent water-dispersible NIR-emitting wurtzite CIS/ZnS core/shell QDs. The procedure starts from nearly spherical template Cu2-S NCs, which are converted into wurtzite CIS QDs with size and shape preservation by topotactic partial Cu+ for In3+ cation exchange at 125 °C. The use of a nearly stoichiometric TOP/In ratio (0.9) ensures that the Cu+-extraction and In3+-incorporation rates remain coupled, despite relatively high reaction temperatures, leading to a highly versatile and fast method to obtain CIS QDs in the 3–8 nm size range (polydispersity of ∼10%). The product CIS QDs can be readily overcoated by a thin (≤1 nm) wurtzite ZnS shell, resulting in CIS/ZnS core/shell QDs with PL tunable from the first to the second NIR window (750–1100 nm) with high PLQYs (e.g., 75% at 820 nm and 25% at 1050 nm). These wurtzite CIS/ZnS core/shell QDs are subsequently transferred to water upon exchange of the native ligands by mercaptoundecanoic acid. The resulting water-dispersible CIS/ZnS core/shell QDs have PLQYs ranging from 39% (at 810 nm) to 6% (at 1030 nm) and good colloidal stability over at least 6 months and are thus promising candidates for in vivo bioimaging applications. It should be noted that the hydrophobic CIS/ZnS QDs obtained immediately after the ZnS shell overgrowth are also attractive as fluorophores in luminescent solar concentrators due to their large absorption cross sections throughout the UV–visible–NIR spectral range, high PLQYs (75% at 820 nm, which is an ideal wavelength in combination with Si solar cells), and large effective Stokes shifts, which would maximize harvesting of the solar spectrum while minimizing reabsorption losses.[59−61]