Nanoscale Transformations in Covellite (CuS) Nanocrystals in the Presence of Divalent Metal Cations in a Mild Reducing Environment.

We studied the structural and compositional transformations of colloidal covellite (CuS) nanocrystals (and of djurleite (Cu1.94S) nanocrystals as a control) when exposed to divalent cations, as Cd2+ and Hg2+, at room temperature in organic solvents. All the experiments were run in the absence of phosphines, which are a necessary ingredient for cation exchange reactions involving copper chalcogenides, as they strongly bind to the expelled Cu+ ions. Under these experimental conditions, no remarkable reactivity was indeed seen for both CuS and Cu1.94S nanocrystals. On the other hand, in the covellite structure 2/3 of sulfur atoms form covalent S-S bonds. This peculiarity suggests that the combined presence of electron donors and of foreign metal cations can trigger the entry of both electrons and cations in the covellite lattice, causing reorganization of the anion framework due to the rupture of the S-S bonds. In Cu1.94S, which lacks S-S bonds, this mechanism should not be accessible. This hypothesis was proven by the experimental evidence that adding ascorbic acid increased the fraction of metal ions incorporated in the covellite nanocrystals, while it had no noticeable effect on the Cu1.94S ones. Once inside the covellite particles, Cd2+ and Hg2+ cations engaged in exchange reactions, pushing the expelled Cu+ ions toward the not-yet exchanged regions in the same particles, or out to the solution, from where they could be recaptured by other covellite nanoparticles/domains. Because no good solvating agent for Cu ions was present in solution, they essentially remained in the nanocrystals.

In recent years, there has been growing interest in colloidal nanocrystals (NCs) of copper chalcogenides, mainly because of their tunable near-infrared (NIR) plasmonic properties,[1−22] and consequently their potential applications as alternatives to noble metal nanoparticles, for example in photothermal therapy for biomedicine or in sensing.[23−32] Copper chalcogenide NC have also been investigated extensively in cation exchange (CE) reactions. In CE, the cation sublattice is partially or completely replaced by a sublattice of new types of cations, with (in principle) minor modification of the anion sublattice.[33−37] CE reactions are regulated by various factors, such as the relative stabilities of the initial and final NC phases, as well as valency, ionic radius, and solvation energy of the entering and exiting cations.[38−41] Solvation energies are important terms in the overall energy balance in CE: in all reported exchange reactions involving copper chalcogenide NCs, the use of phosphines (such as tributylphosphine or trioctylphosphine) is critical in order to achieve extraction of Cu+ ions from the NCs, as these molecules are soft Lewis bases and have strong affinity with soft Lewis acids (the Cu+ ions in this case).

In this work, we investigated the conditions under which copper chalcogenide NCs, in the specific case CuS (covellite) NCs (and Cu1.94S (djurleite) NCs as a control), could engage in reactions with metal cations, in the absence of phosphines. Covellite is a structure in which the S ions have on average −1 oxidation state, due to the presence of S–S covalent bonds. This is markedly different from the most common −2 state for sulfur in the vast majority of ionic compounds involving sulfur and one or more metal cations, including Cu1.94S.[42] We explored whether the possibility to break these S–S bonds in covellite (by the intake of electrons from a reducing agent) could act as a driving force for the concomitant entry of metal cations in the NCs, a mechanism that would not be possible for Cu1.94S. We chose Hg2+ and Cd2+ as test cations, based on the fact that Hg2+ and Cd2+ are classical cases of ions that, in the presence of phosphines, can be engaged in quantitative CE reactions with Cu2–S NCs.[43−45]

In all tests, carried out at room temperature for 24 h, we found that, if the particles (both Cu1.94S and CuS NCs) were simply mixed with solutions of the Hg2+ or Cd2+ cations (at a metal to Cu feed ratio equal to 1:1), with no reducing agent added, they exhibited a limited reactivity toward the cations, with a fraction of incorporated metal cations around 1–7% with respect to the total amount of Cu present in the NCs (Table , Supporting Information Table S2). Then, in the presence of ascorbic acid acting as a mild reducing agent, the reactivity of covellite NCs significantly increased, with the incorporation fraction being around 20–22% (relative to Cu) and reaching values over 30% in the presence of a large excess of metal cations (in the case of Hg2+, a large excess of cations induced the concomitant formation of CuHgSBr). At the same time, sizes and shapes of the initial particles were essentially preserved during the reaction. On the other hand, the addition of ascorbic acid to Cu1.94S NCs did not boost their reactivity toward Cd2+ or Hg2+ cations any further. These results suggest that the possibility for covellite NCs to be reduced is the main drive for their increased reactivity toward the metal cations.

Table 1

M:Cu Ratio (M = Cd, Hg), as Determined by ICP Analysis, in the NC Samples Collected after Reacting Covellite CuS NCs with Cd2+ and Hg2+ Cations (for 24 h), Either in the Presence or in the Absence of Ascorbic Acid

cation	M:Cu ratio (no ascorbic acid)	M:Cu ratio (with ascorbic acid)
Cd2+	0.01:1	0.22:1
Hg2+	0.05:1	0.20:1

We also studied in detail the structural and compositional evolution of the covellite NCs in the two cases. The reaction of Cd2+ and Hg2+ with CuS NCs led to mixtures of NCs of mainly three types: (i) NC heterostructures made of local domains of CdS (HgS) and domains of Cu2–S (0 < x < 1) composition; (ii) totally exchanged CdS (HgS) NCs; and (iii) NCs with Cu2–S composition. The Cu2–S NCs/domains often embedded a fraction of Cd2+ (Hg2+) ions. These reactions can be considered as being initiated by an incorporation of the metal cations in the NCs, made possible by the concomitant inflow of electrons from the reducing agent. Once in the NCs, the electrons enable the local reduction of the S anion framework via rupture of the S–S bonds, whereas the cations are locally engaged in CE reactions. Notably, the Cu+ ions released from the exchanged regions/NCs were not found in solution, but were rather taken up by the unexchanged covellite domains and NCs, which then acquired a Cu2–S (0 < x < 1) composition. Unlike the previously reported CE studies of other copper chalcogenides with Cd2+ and Hg2+, one of the notable observations in the present study is that covellite (CuS) NCs can engulf metal cations and retain copper, since a negligible release of copper species in solution is seen.

Experimental Section

Materials

Copper chloride (CuCl, anhydrous, 99.99%), copper(II) acetate monohydrate (C4H6CuO4·H2O, 99.99% trace metals basis), cadmium iodide (CdI2, 99.990%), mercury bromide (HgBr2, 99.9%), oleylamine (OM, >70%), and octadecene (ODE, 90%) were purchased from Sigma-Aldrich. Elemental sulfur (99%) was obtained from Strem Chemicals, and methanol (anhydrous, 99.9%) and toluene (anhydrous, 99.8%) were from Carlo Erba reagents. All chemicals were used as received without further purification.

Synthesis of Covellite NCs

Covellite CuS nanoplates were synthesized by a modified procedure previously reported by us.[42] Briefly, a sulfur solution was prepared first by degassing a mixture of 0.160 g (5 mmol) of sulfur, 25 mL of ODE, and 25 mL of OM in a 100-mL three-neck flask at 130 °C under vacuum for 30 min. The as-formed clear yellow solution was cooled to room temperature (RT) under N2 atmosphere, followed by addition of 0.248 g (2.5 mmol) of CuCl and degassing at RT for an additional 60 min. The as-obtained dark green solution was then heated to 200 °C with a ramp of 8 °C/min, and it was then kept at this temperature for an additional 30 min, after which it was cooled to RT. To precipitate the NCs, methanol was added, and followed by centrifugation. The precipitate was finally dispersed in toluene for further characterizations and reactions.

Reaction of the As-Synthesized Covellite CuS NCs with Cd2+ and Hg2+ Cations

All these reactions were performed at RT in a N2-filled glovebox. As a typical example of a reaction of CuS with Cd2+, and Cd2+ and ascorbic acid, 0.1 M solutions were prepared by respectively dissolving CdI2 (0.0146 g) and ascorbic acid (0.0352 g) in methanol (0.4 mL for CdI2 and 2 mL for ascorbic acid) at RT. Then, 0.4 mL of Cd2+ solution and 2 mL of ascorbic acid solution were added in a vial containing 4.0 mL of covellite NC dispersion (0.01 M in Cu+ ions, in toluene) under magnetic stirring. Aliquots (1.5 mL) of the NC solution were collected at different reaction times to monitor the evolution of morphologies, optical spectra, and compositions. The aliquots and the final sample collected after 24 h were precipitated by addition of 1.5 mL of methanol followed by centrifugation at 3000 rpm for 20 min. For all aliquots, the supernatant was then carefully collected for inductively coupled plasma–optical emission spectroscopy (ICP–OES) measurements, while the precipitate was dispersed in toluene for one additional cleaning cycle. The final precipitate was redispersed in toluene (0.5 mL) for subsequent characterization. The reactions with Hg2+ cations followed a similar procedure. The molar ratio of the guest cations (Cd2+ and Hg2+) introduced in the reaction environment over the Cu+ cations (from covellite CuS), that is, the feed ratio, was equal to 1:1 in all the reactions, except for the cases in which a large excess of guest cations was probed.

Transmission Electron Microscopy (TEM)

Low-resolution TEM measurements were carried out on a JEOL JEM-1011 microscope operating at 100 kV. The samples were prepared by drop-casting NCs solutions on carbon-coated 200-mesh copper grids. High-resolution TEM (HRTEM) images were acquired on a JEOL JEM-2200FS microscope equipped with a Schottky gun working at 200 kV, a CEOS spherical aberration corrector in the objective lens allowing for a spatial resolution <1 Å, and an in-column Ω energy filter. The chemical composition of the NCs was determined by energy dispersive X-ray spectroscopy (EDS) performed in scanning mode (STEM) and by acquiring the spectra on a Bruker Quantax 400 system with a 60-mm XFlash 6T silicon drift detector (SDD). The quantification of the elements in the spectra was performed with the Cliff–Lorimer method. For HRTEM and STEM-EDS analyses the samples were prepared by drop-casting NC solutions onto ultrathin carbon-coated 300-mesh gold grids and inserted with a beryllium cup holder to avoid spurious copper signal.

X-ray Diffraction (XRD)

XRD analysis was performed on a PANalytical Empyrean X-ray diffractometer equipped with a 1.8 kW Cu Kα ceramic X-ray tube and PIXcel3D 2 × 2 area detector, and operating at 45 kV and 40 mA. The diffraction patterns were collected in air at room temperature using parallel-beam (PB) geometry and symmetric reflection mode. XRD data analysis was carried out using HighScore 4.1 software from PANalytical.

Elemental Analysis

ICP–OES was performed on an iCAP 6000 spectrometer (ThermoScientific) for quantification of elemental composition of NC samples and of the supernatant solutions. The samples were decomposed in aqua regia (HCl/HNO3 equal to 3:1 (v/v)) overnight prior to ICP–OES measurements.

X-ray Photoelectron Spectroscopy (XPS)

The samples were prepared in the glovebox by drop-casting a few microliters of NC solutions onto a graphite substrate (HOPG, ZYB quality, NTMDT), which was then transferred to the XPS setup in an ad hoc transfer vessel to avoid air exposure. Measurements were performed on a Kratos Axis Ultra DLD spectrometer, using a monochromatic Al Kα source (15 kV, 20 mA). Wide scans were acquired at an analyzer pass energy of 160 eV. High-resolution narrow scans were performed at constant pass energy of 10 eV and steps of 0.1 eV. Photoelectrons were detected at a takeoff angle of Φ = 0° with respect to the surface normal. The pressure in the analysis chamber was maintained below 7 × 10–9 Torr for data acquisition. Data were converted to VAMAS format and processed using CasaXPS 2.3.16 software. The binding energy scale was internally referenced to the C 1s peak (BE for C–C = 285 eV).

Optical Spectroscopy

Raman spectra were recorded on a Jobin Yvon HR800 spectrometer. The NCs were deposited on silicon substrates under N2. Data were acquired at λ = 632.8 nm with a 50× objective using a nominal power of 25 mW and integration time of 120 s. UV–vis-NIR extinction spectra on the NC solutions were performed on a Varian Cary 5000 UV–vis-NIR spectrophotometer in the 350–2000 nm range.

Results and Discussion

The covellite structure is peculiar, since it is formed by a repetition, along the c axis, of a trilayer motif made of a layer of triangular CuS3 units between two layers of CuS4 tetrahedra.[46,47] The trilayers are stitched together by sulfur–sulfur covalent bonds. There is an equivalent structure for CuSe, named klockmannite.[48] On the other hand, no other metal ions are known to crystallize with sulfur (or with selenium) in the same structure, or in any structure that resembles covellite. In a previous work of ours, we evidenced how covellite NCs could easily incorporate Cu+ cations, at room temperature, up to a stoichiometry around Cu2S.[42] During this incorporation, despite that some XRD reflections of covellite were maintained (albeit with a shift to lower angles), the overall structure underwent partial amorphization. We also demonstrated that the insertion of Cu+ ions was charge-compensated by the acquisition of an equimolar amount of electrons, made possible by the oxidation of a fraction of Cu+ ions remaining in the solution phase to Cu2+. A logical extension of those findings was then to explore whether a similar reaction scheme could work with other metal ions too, and, if so, under what conditions such ions could react with covellite NCs and what would be the products of the reaction. For Hg2+ and Cd2+ cations studied here, the mechanism operative for the Cu+ insertion case discussed above (part of Cu+ ions in solution is oxidized to Cu2+ and provides electrons to charge-balance the entry of Cu+ ions in the NCs) cannot work (Hg and Cd are already at their highest oxidation state). This left room to other various possibilities: (i) the guest ions may be capable of replacing the Cu+ ions in the covellite lattice, as in a standard CE reaction, but then the anion framework would have to reorganize considerably; (ii) incorporation of the ions in the NCs would be feasible, provided it could be supported by a source of electrons in the reaction environment; or (iii) a combinations of the two former mechanisms could be operative, for example there could be both incorporation and CE.

Our experiments, reported here, supported a combination of the two mechanisms. First, for both cations tested, when the reactions were performed in the absence of ascorbic acid, limited reactivity was found (we recall that the molar ratio of the guest cations (Cd2+ or Hg2+) to Cu+ ions from covellite CuS, in all the reactions discussed in this work was 1:1, unless otherwise stated). Table , second column, lists the M:Cu ratios (M = Cd, Hg) in the CuS NCs as found by elemental analysis (according to ICP–OES) on the samples purified after 24 h of reaction. In all samples, the fraction of cations incorporated in the CuS NCs was low, around 5% for Hg and 1% for Cd. Because of such low amount of cations incorporated, chemical quantification by ICP–OES analysis on a bulk sample of NCs was preferred to EDS analysis, which is instead a local analysis and is performed on groups of NCs in the electron microscope. The limited reactivity of covellite NCs in the absence of ascorbic acid could be evidenced by the little change in the XRD patterns (which can still be indexed as covellite, see Figure a), morphology (Figure S1), and high-resolution XPS (S 2p) (Figure S6b,c) of NCs collected after 24 h of reaction with Cd2+ and Hg2+ cations without ascorbic acid.

Figure 1

XRD patterns (a) and optical spectra (b) of as-synthesized covellite (CuS) NCs, and NCs collected after mixing CuS with Cd2+ (red) and Hg2+ (blue) for 24 h, in the absence of ascorbic acid.

Despite the low degree of incorporation of cations (which perhaps in some cases was just a surface adsorption phenomenon), damping and red-shift of the NIR absorption band (the latter due to a localized surface plasmon resonance, LSPR) was seen in the optical absorption spectra of the solutions of NCs at 24 h, with varying degree from cation to cation (Figure b). This indicates that the NIR LSPR of covellite NCs is strongly sensitive to various environmental factors. Similarly, djurleite (Cu1.94S) NCs (composition close to Cu2S) exhibited limited reactivity with Cd2+ and Hg2+ cations in the absence of reducing agent (Figure S4a, Table S2), except that their weak NIR plasmon absorbance (due to the presence of Cu vacancies) was further damped and red-shifted upon mixing the NCs with cations (Figure S4b).

For reactions of covellite (CuS) carried out in the presence of ascorbic acid, however, much higher M:Cu ratios (M = Cd, Hg) were found in the NCs (see Table , third column and Table S1, the latter reporting also the compositions as found by EDS analysis). In contrast, the reactivity of djurleite (Cu1.94S) NCs was not significantly increased by the presence of ascorbic acid for both cations (Figure S4c and Table S2), although the formation of additional Hg2Br2, in the Hg case, complicated the XRD analysis: only 7% of Hg and 0.3% of Cd were detected in the resulting NCs after 24 h (Table S2). The higher reactivity of the covellite NCs was also confirmed by XPS analysis of the purified NCs (Figure S6d,e). Here, the spectral features in the S 2p region, which allows discrimination between sulfides and disulfides,[42] were clearly supporting the almost complete absence of signal from disulfides in the final NCs, which are instead present in the initial covellite NCs due to the S–S covalent bonds in their structure. Therefore, as the vast majority of S–S bonds were broken in the reacted NCs, we expect a significant reorganization of the crystal lattice for the reactions tested. The discussion that follows examines in detail the individual cation cases and reports, unless otherwise stated, the results of reactions on covellite NCs performed in the presence of ascorbic acid.

Reactions with Cd2+ Ions

Panels a and b of Figure compare low-magnification TEM images of the initial CuS sample and of the same sample after 24 h incubation in a mixture of Cd2+ ions and ascorbic acid. The overall size and shape of the initial sample (and their relative distributions) were essentially preserved. Size-distribution histograms of the initial and the final NCs, as well as TEM images of initial and intermediate samples, are shown in Figures S5a,b and S7a–e, respectively. Experimental XRD patterns of aliquots taken at 1, 2, 4, and 24 h evidenced the emergence of peaks belonging to the CdS phase (Figure S7f). The XRD pattern of the sample at 24 h, reported in Figure c (top, to be compared to the pattern of the initial covellite NC, shown in c, bottom), carried strong signatures of CdS, while the assignment of the remaining peaks was less straightforward. HRTEM and EDS analyses of the 4-h (intermediate) and 24-h samples revealed a mixture of NC structures. The predominant types of particles were heterostructures composed of domains that could be indexed as CdS and Cu2–S (0 < x < 1), respectively (Figure f,g). In addition, there were entirely exchanged NCs, i.e. CdS nanoplatelets, and platelets with phases possibly matching various Cu2–S bulk phases, but which, according to EDS analysis, contained a fraction of Cd too (Figure h–m). In the 24-h sample even chalcocite Cu2S NCs were detected. An increasing degree of incorporation of Cd was found in the NCs over reaction time, from 5% (with respect to Cu) at 1 h, to 22% at 24 h, as evidenced by the ICP analyses (see Table and Table S1).

Figure 2

TEM images of the as-synthesized covellite CuS NCs (a) and Cd-incorporated NCs achieved by reacting covellite NCs with Cd2+ cations in the presence of ascorbic acid for 24 h (b). The scale bar is 20 nm. XRD patterns (c), Raman spectra (d), and optical spectra (e) of samples shown in a and b. (f) HRTEM image and (g) corresponding FFT of a representative single CdS/Cu2–S nanoparticle collected at 4 h by reacting the as-synthesized covellite CuS NCs with Cd2+ cations in the presence of ascorbic acid. (h) HRTEM image and (i,j) corresponding EDS mapping. (k–m) FFT of the single NPs shown in h. (n) Schematic illustration of the reaction of CuS NCs with Cd2+ cations in the presence of ascorbic acid.

We can hypothesize that the incorporation of Cd2+ ions in the covellite NCs, aided by the overall reducing environment, can locally initiate CE, with reorganization of the anion (sulfur) framework and the consequent formation of CdS domains, a process that eventually can spread to the whole NC. The Cu+ ions that are released from the CdS-exchanged domains can diffuse to the not exchanged domains within the same NC. This would explain the formation of CdS–Cu2–S heterostructures, such as the one displayed in Figure f. On the other hand, the presence of Cu2–S platelets can be rationalized by the incorporation of Cu+ ions that had been previously released from the exchanged NCs to the solution phase. This is similar to what was found previously by us[42] when treating covellite NCs with Cu+ ions. These hypotheses were additionally backed by the experimental evidence that elemental analysis (by ICP–-OES) of the solution phase revealed no appreciable Cu in the supernatant after NCs precipitated from the solution: therefore, even if Cu+ ions were released in solution, they were unstable there and tended to react with the CuS NCs/domains. Partial support to the CuS → Cu2–S transformation came also from the red-shift and strong damping of the NIR absorbance as the reaction proceeded (Figure e, see also Figure S7h). This is in line with our previous work on the evolution of CuS NCs to Cu2–S,[42] and is due to the decrease in the density of free carriers in the NCs that accompanies such transformation.

The transformations were also corroborated by Raman analysis (Figure d). The Raman spectrum of the pristine CuS NCs exhibited vibration modes at 472 and 266 cm–1 (see also Figure S7g), which are in good agreement with the Raman spectra of CuS reported by Ishii et al. and Peiris et al.[47,49] The peak at 266 cm–1 corresponds to a Cu–S vibrational mode,[50] while the intensity of the initially intense peak at 472 cm–1, attributed to a S–S stretching mode, decreased (relative to the peak at around 266 cm–1) as the stoichiometry evolved from CuS to Cu2–S (0 < x < 1). A broad peak around 300–330 cm–1 might be due to the silicon substrate, as its contribution was seen in all the samples analyzed (see Figure S11 for the Raman spectrum from the substrate only). After 24 h reaction with Cd2+, a broad feature ranging from 220 to 350 cm–1 emerged, which might be due to the overlapping modes of Cu–S and Cd–S (the longitudinal optical phonon mode of CdS peaks at around 300 cm–1).[51] In comparison, the initial Raman peak at around 472 cm–1 had lost considerable intensity, an indication that part of the copper sulfide phase was transformed to CdS and most likely the remaining part had evolved to a copper sulfide phase richer in Cu than the initial covellite NCs.

We also carried out reactions using a large excess of Cd2+ ions (with a Cd:Cu feed ratio of 10:1). These experiments produced a sample in which the Cd:Cu:S ratios in the NCs were 0.39:1.1:1, that is, no complete cation exchange could be achieved (see Figure S9c,d). This makes sense, if one considers that the concomitant introduction of Cd2+ ions and of electrons (provided by ascorbic acid) in the NCs can proceed as long as there are S–S bonds that can be reduced. The S atoms involved in this bond type represent 66% of the overall number S atoms in covellite, and the rupture of each bond operated by two electrons permits the entry of one Cd2+ ion. Therefore, the theoretical limit of incorporation of Cd2+ atoms would correspond to 33% of the total number of S and Cu atoms in covellite, which is very close to the ratio found experimentally. Clearly, in this case too, the incorporated Cd ions trigger cation exchange reactions, with release of Cu ions that could diffuse then to the not exchanged domains/NCs. It is noteworthy that even when working with such high Cd:Cu feed ratios (10:1) no Cu could be detected in solution (that is, all Cu+ ions stayed or were reinserted in the NCs).

Reactions with Hg2+ Ions

The reactivity of CuS NCs toward Hg2+ cations in the absence of ascorbic acid was limited, although the incorporation of Hg after 24 h could reach 5% (Table ), higher than that of Cd. On the other hand, in analogy with the case of reaction with Cd2+ cations discussed earlier, a more pronounced reactivity with Hg2+ was seen in the presence of ascorbic acid. Here again, both size and shape of the NCs were basically preserved (see Figure a, Figure S8a–d, and S5c portraying additional images of intermediate samples and the size distribution histogram of the 24 h sample). However, the XRD pattern of the sample after 24 h reaction time (Figure b), and those of the intermediate samples (see Figure S8e), evidenced the formation of the HgS phase over time. Other peaks were present, which made the assignment more difficult. These peaks could be due to a mixture of Cu2–S phases (0 < x < 1), although the diffraction peaks from HgS became predominant over time. In analogy with the reactions involving Cd2+ ions, negligible amounts of Cu species were released to the solution, as revealed by elemental analysis (by ICP–OES), and can be interpreted along the same lines as for the reactions with Cd2+. The incorporated amount of Hg was higher here than for the Cd2+ case: it ranged from around 10% (relative to Cu) at 1 h, to 20% at 24 h (see Table and Table S1, the latter reporting also compositions as found by EDS analysis). The phase transformation from covellite to Cu2–S (0 < x < 1) and HgS was concomitantly accompanied by a marked red-shift and strong damping of NIR absorption (Figure d and Figure S8f). According to HRTEM and EDS analysis, the sample was composed of a mixture of HgS/Cu2–S heterostructures, often containing more than one HgS domain per NC (Figure e,f) and fully exchanged HgS NCs, as well as Cu2–S NCs (Figure g–j), the latter particles this time containing only a minor fraction of Hg. The Raman spectra of the 24-h sample was in line with the results discussed for the Cd2+ reaction, that is, with the most notable and interpretable change being the damping of the strong peak initially at 472 cm–1 (Figure c). Overall, the reaction scheme (Figure k) was therefore not much different from that involving the Cd2+ ions (Figure n).

Figure 3

(a) TEM image of NCs collected at 24 h by reacting the covellite CuS NCs with Hg2+ cations in the presence of ascorbic acid. The scale bar is 20 nm. (XRD patterns (b), Raman spectra (c), and optical spectra (d) of as-synthesized covellite CuS NCs and Hg-incorporated samples shown in a. (e) HRTEM image and (f) FFT of a representative single HgS/Cu2S nanoparticle collected at 4 h by reacting the as-synthesized covellite CuS NCs with Hg2+ cations in the presence of ascorbic acid. (g,h) HRTEM images of individual NCs collected at 24 h. (i,j) Corresponding FFT of the single NCs shown in h and g, respectively. (k) Schematic illustration of the reaction of CuS NCs with Hg2+ cations in the presence of ascorbic acid.

In the reaction with excess Hg2+ cations (with a starting Hg:Cu feed ratio equal to 10:1), the final Hg:Cu:S molar ratios, as measured by ICP, were around 0.6:1.1:1. The Hg:Cu ratio (0.6:1.1 = 0.55:1) was higher than the theoretical limit set by the fact that the Hg incorporation should stop when all S–S bonds are broken (0.33, see the Cd case discussed earlier). This apparent inconsistency was clarified by XRD analysis (see Figure S9c), according to which this sample contained an important contribution from a CuHgSBr phase.[52] This is interesting, since this compound is usually synthesized by reacting HgS with CuBr,[52] while here it was formed by reacting CuS with HgBr. This compound is, however, unstable under the electron beam, as by HRTEM we could only unequivocally identify HgS, chalcocite, and covellite-like Cu2–S NCs (Figure S10). The Hg case, however, will require further scrutiny.

Conclusions

We have studied the incorporation of heavy metal Cd2+ and Hg2+ cations in covellite (CuS) NCs at room temperature, in the presence of a mild reducing agent such as ascorbic acid. The incorporation of guest metal cations is promoted by the breaking of the S–S bonds, operated by the reducing agent. A negligible fraction of Cu species could be detected in the solution phase, due to the recapture of released Cu by the not exchanged CuS domains or by the remaining CuS NCs, upon the incorporation of Cd2+ and Hg2+ cations. This is much different from the previously reported CE studies on various other copper chalcogenide NCs. This notable feature and the mechanism proposed in this study extend the knowledge on nanoscale chemical transformations.