Introduction

The electron-phonon coupling is, in addition to the Coulomb interaction, one of the fundamental interactions between quasiparticles in solids1. It plays an important role in a variety of physical phenomena, in particular, low-energy electronic excitations can be strongly modified by coupling to lattice vibrations, which influences e.g. their transport2 and thermodynamic3 properties.

Semiconducting transition metal dichalcogenides (S-TMDs) based on molybdenum and tungsten, i.e. MoS2, MoSe2, MoTe2, WS2, and WSe2, are the most well-known representatives of van der Waals (vdW) materials4,5. Their most distinguished hallmark is the transition from indirect- to direct-band gap, when thinned down from a bulk to a monolayer (ML)6,7,8,9. Due to the very strong absorption and direct energy band gap in the ML limit, in recent years this class of materials has become of great interest from both research4,10,11,12 and development point of view13,14. The photoluminescence (PL) signal of S-TMDs is caused mainly by excitonic effects, even at room temperature, due to the large excitonic binding energy at the level of hundreds of meV15,16,17. It arises from the reduced dimensionality of the material and limited dielectric screening of the environment18. Constant progress in sample preparation, particularly the encapsulation of MLs in thin layers of hexagonal BN (hBN)11, leads to narrowing of the observed emission lines to the limit of a few meV, which opens the possibility of studying a variety of individual excitonic complexes associated with both bright and dark states19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37.

Raman scattering excitation (RSE) experiment performed on S-TMD MLs was proposed a few years ago in ref. 38 as a powerful technique to investigate the interaction between different excitonic complexes and phonons, i.e. exciton-phonon coupling (EPC). This approach is analogous to the PL excitation (PLE) method, in which the detected spectra are measured as a function of the excitation energy, while the detection energy/window is fixed. In the RSE experiment, the detection window can be set to cover the emission of different excitonic complexes, while the excitation energy is tuned in the energy region only slightly above (a few dozen of meV) these excitonic resonances. This results in the outgoing resonant conditions of Raman scattering (RS)39. The analysis of the outgoing phonon modes crossing the excitonic emission allows one to reveal the details of the exciton-phonon interaction in a given material. The RSE spectroscopy might be seen as the extension of the resonant Raman scattering method (RRS), which was extensively used to study the characteristics of phonon modes in S-TMDs40,41,42,43,44,45,46. However, a typical RRS experiment is carried out with a few selected excitation points42,43,44 offering only limited access to investigate the resonant changes in the shape/intensities of the phonon modes. This restriction is lifted with the RSE measurements performed when practically continuously tuning the excitation energy38,47,48. RSE spectra were previously reported for bare MLs of MoSe247 and WS238 exfoliated on Si/SiO2 substrates, as well as for the MoSe2 ML encapsulated in hBN flakes48. In the case of the MoSe2 MLs47,48, the RSE spectrum was dominated by several phonon replicas of the LA mode. In contrast, for the WS2 ML38, a very rich Raman spectrum was presented with numerous phonon modes originating from the edge of the Brillouin zone (BZ) as well as multiphonon modes. Knowing that S-TMD MLs are organised into two subgroups, i.e. bright and darkish, due to the type of the ground exciton state (bright and dark, respectively)28,32,34,49, their low-temperature (T ~ 4.2–20 K) PL spectra display completely different complexity. The high quality of the MLs embedded between the hBN flakes, accompanied by the division of the MLs into two subgroups, has motivated us to conduct a comprehensive study devoted to the EPC in such S-TMD MLs with the aid of the RSE technique.

In this work, we use the RSE technique to investigate the exciton-phonon interaction in four high-quality samples consisting of the MoS2, MoSe2, WS2, and WSe2 MLs encapsulated in hBN flakes. We observe the intensity enhancements of Raman modes, while their emission energies match the emission energy of the corresponding neutral exciton (X). The measured RSE spectra are composed of many peaks that can be attributed not only from the BZ centre (Γ points) but also from other points in the BZ (e.g. M points), which are followed by lines arising from multiphonon processes. In the case of outgoing resonance conditions with the X emission, the intensity profiles of the A\({}_{1}^{{\prime} }\) modes in the studied MLs resemble the emission lines of the corresponding X emission. The A\({}_{1}^{{\prime} }\) modes’ enhancements are extraordinarily strong for the WS2 and WSe2 MLs, but the corresponding increases of the A\({}_{1}^{{\prime} }\) peaks in the MoS2 and MoSe2 MLs are significantly smaller. Instead, for the Mo-based MLs, the modes involving LA phonons from the edge of the BZ, such as 2LA, 3LA, and so on, are greatly enhanced. Moreover, we observe that the A\({}_{1}^{{\prime} }\) intensity is also significantly enhanced when the excitation energy is in the vicinity of the X emission, which leads to the incoming resonance condition. Our experiments demonstrate that the landscape of the exciton-phonon interaction, as traced with our RSE experiments, appears to be qualitatively different in two distinct subgroups, of bright and darkish, S-TMD monolayers. The specific alignment of single-particle energy bands and the related characteristic scattering processes are speculated to account for the observed differences.

Results

Raman scattering excitation spectra

Figure 1 presents false colour maps of the low-temperature (T = 5 K) emission intensities collected for the MoS2, MoSe2, WS2 and WSe2 MLs encapsulated in hBN flakes. The vertical position of a given map has been shifted to reflect the relative energy of the neutral exciton (X) emission. The assignment of the X line as arising from the neutral A exciton is straightforward and consistent with many other studies on these MLs11,17,28,34,36. We focus on the X lines, but we are aware that the low-temperature PL spectra of the investigated MLs are composed of several additional emission lines (see, for example, Fig. 5), particularly due to charged excitons, dark complexes, and the corresponding phonon replicas22,23,29,31,36,50,51,52,53,54.

Fig. 1: Raman scattering excitation.
figure 1

False-colour maps of optical response of the a MoS2, b MoSe2, c WS2, and d WSe2 MLs encapsulated in hBN measured at low temperature (T = 5 K) under excitation of tuneable lasers (excitation power ~ 150 μW). The maps are vertically aligned with respect to the energy distance from the energy of the neutral exciton (X) emission in these materials, which is marked on the red right-hand energy scales. The colour scales have been normalised to the maximum intensity. The detection energies of the maps are centred around the X emission. The phonon modes are visible as several narrow parallel resonances passing diagonally across the maps. For the MoS2 ML, white arrows point out lines which are investigated in the following.

Full size image

It can be seen that the line shape of the detected signal significantly depends on the excitation energy. Several parallel narrow lines superimposed on neutral exciton emissions are clearly observed in Fig. 1. These sharp lines follow the tuned excitation energy, which points out the Raman scattering as their origin and are examined in detail in the following. Moreover, for the MoS2 and WS2 MLs, the X emissions are strongly enhanced with decreasing excitation energies towards the respective X energies. The observed enhancement of the X intensity can be described in terms of extremely efficient formation of neutral excitons at larger k-vectors due to the near-resonant excitation. Note that the laser line used in experiments is narrow enough to investigate individual narrow peaks, but the excitation range of measurements is limited by the spectral range of the lasers. Moreover, the dye laser applied for MoS2 measurements has a limited tuning accuracy. This leads to the observation of the narrow lines only at specific excitation and detection energies, in contrast to almost continuous linear evolutions present for other MLs, see Fig. 1. As is seen in the figure, the intensities of the Raman modes strongly depend on the material. For Mo-based MLs, i.e. MoS2 and MoS2, the intensity enhancement of different phonon modes is similar and relatively small compared to the X intensity. In contrast, extraordinary increases in the selected Raman peaks accompanied by a large number of smaller modes are observed in the spectra measured on the WS2 and WSe2 MLs.

In order to investigate in detail the apparent phonon modes in the aforementioned optical responses of the MLs, the RSE spectra are plotted in Fig. 2. The RSE spectra were obtained by fixing the detection energy at the emission energy of the X line, while the excitation energy was being tuned. Note that the horizontal energy scale in the figure corresponds to the relative distance between the excitation and X energies, and is given in cm−1, which is a typical unit for RS experiments and represents the so-called Raman shift. As can be seen in the inset of Fig. 2, the intensity of the A\({}_{1}^{{\prime} }\) modes in the W-based MLs is extraordinarily high and exceeds the X emission several times (compare with Fig. 1c and d). In contrast, the phonon modes related to the longitudinal acoustic branch, i.e. 2LA, 3LA, …, are enhanced for the Mo-based MLs, while the corresponding intensities of the A\({}_{1}^{{\prime} }\) modes are much smaller. All of the presented RSE spectra are very rich, consisting of many phonon modes. This confirms the resonant excitation conditions of RS, since the corresponding non-resonant Raman spectra are composed of only two Raman-active modes in the backscattering geometry of the experiment, i.e. A\({}_{1}^{{\prime} }\) and E\({}^{{\prime} }\)55, whereas RS spectra become especially rich under resonant excitation conditions40. Using the previous results presented in the literature38,40,43,56,57,58, we have identified almost all observed peaks, marked in Fig. 2 with vertical black and grey arrows. Their energies and assignments to the respective phonons are summarised in the Supporting Information (SI). Besides the two Raman-active modes from the Γ point of the BZ, the majority of the observed peaks were ascribed to phonons from the M points, which are located at the edge of the hexagonal BZ of the MLs in the middle between the K+ and K− points. For example, we have identified peaks related to combinations of all three branches of acoustic phonons (e.g. LA, ZA, TA) as well as their higher orders (e.g. 2LA, 3LA,…); see the SI for details. Their observation can be associated with the resonant excitation of RS, which results in the appearance of Raman inactive momentum-conserving combinations of acoustic modes from the edge of the BZ40,59,60.

Fig. 2: RSE spectra.
figure 2

Raman scattering excitation (RSE) spectra detected at the energies of the X line measured on the MoS2, MoSe2, WS2, and WSe2 MLs encapsulated in hBN. The vertical scale has been adjusted to make the low-intensity peaks visible, while the inset shows the RSE spectra with the most pronounced peaks. Note that the RSE spectrum of the MoSe2 ML was multiplied by a factor of 2.5 for clarity. The black arrows denote the labelled peaks in the Figure, while the grey arrows point to the identified peaks with their assignment presented in Supplementary Table 1 in the SI. The A\({}_{1}^{{\prime} }\) peaks are marked with their energy in cm−1.

Full size image

We have also investigated the RSE spectrum measured on the MoSe2 ML grown on hBN flake using the molecular beam epitaxy (MBE) technique61. The ML was also covered with a thin hBN flake using mechanical exfoliation and dry transfer technique. The spectra measured for epitaxial layers exhibit peaks at energies similar to those in the case of exfoliated layers, but the relative intensity of various peaks differs. Particularly pronounced for epitaxial layers are LA replicas, which are observed from 2nd to 12th order, see SI for details.

Enhancement profiles of A\({}_{1}^{{\prime} }\) modes due to outgoing resonance

The resonant enhancement of the phonon modes presented above originates from the so-called outgoing resonance, which requires that the resonant excitation energy must be equal to the sum of the phonon and exciton energies58. To verify this hypothesis we compare the PL emissions of the X lines and the integrated intensities of the A\({}_{1}^{{\prime} }\) peaks in Fig. 3. The choice of A\({}_{1}^{{\prime} }\) modes is motivated by their substantial intensities compared to the remaining phonon modes, see Fig. 2. Note that we also analyse some other selected modes, e.g. 2ZA, 2LA, etc., in the SI, which show a similar behaviour as the one found for A\({}_{1}^{{\prime} }\). As can be seen in Fig. 3, the data were fitted using Lorentzian functions, which nicely reproduce the profiles shown for both the X line and the A\({}_{1}^{{\prime} }\) peak. Furthermore, the parameters of the Lorentzian profiles fitted for the X lines and the A\({}_{1}^{{\prime} }\) peaks are summarised in Table 1. Except for the results obtained for the MoS2 ML, the extracted profiles for the A\({}_{1}^{{\prime} }\) peaks, described by their energies and linewidths, follow the shape of the X emission line. For the MoS2 ML, the observed small discrepancy in the energies of the X and A\({}_{1}^{{\prime} }\) peaks of about 1 meV may be explained in terms of the small energy resolution of the laser used in this case. Nevertheless, the results obtained confirm that the enhancements of the A\({}_{1}^{{\prime} }\) peaks in the S-TMD MLs are due to the outgoing resonance.

Fig. 3: PL emission versus RSE response.
figure 3

Comparison of (orange points) PL emissions of the X lines and (blue points) integrated intensities of the A\({}_{1}^{{\prime} }\) peaks measured on the a MoS2, b MoSe2, c WS2, and d WSe2 MLs encapsulated in hBN. Solid black curves represent the corresponding fitting using Lorentzian function.

Full size image
Table 1 Summary of the obtained fitting parameters of the X emission lines and A\({}_{1}^{{\prime} }\) profiles using Lorentzian functions, shown in Fig. 3.
Full size table

Outgoing versus incoming resonance

Figure 4a and b shows correspondingly a schematic illustration of the incoming and outgoing excitation resonance with the excitonic transition (X). The resonance type depends on the resonant conditions between the incident or scattered light with the X energy, while the energy difference between the excitation and the emission amounts to the phonon energy. The results presented so far are associated with the outgoing X resonance, which occurs when the scattered-photon energy is equal to that of the optical X transition. In contrast, the incoming X resonance takes place when the incident photon energy equals the energy of an optical X transition.

Fig. 4: Resonant Raman scattering.
figure 4

Schematic illustration of the a incoming and b outgoing excitation resonance with the excitonic transition, denoted as X.

Full size image

To study the possibility of achieving the incoming resonant conditions in S-TMD MLs, we measured the optical response of the WSe2 ML in both the outgoing and incoming resonance conditions, see Fig. 5. The low-temperature PL spectrum of the WSe2 ML encapsulated in hBN flakes, presented in panel (a) of the figure, displays several emission lines with a characteristic pattern similar to that previously reported in several works on WSe2 MLs embedded in between hBN flakes19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37. Panel (b) and (c) display correspondingly the optical response of the WSe2 ML in the outgoing and incoming resonances with the neutral exciton emission. As for the outgoing conditions, the observed results are analogous to those shown in Fig. 1, the incoming resonance leads to the two prominent effects. The first one, seen as a significant enhancement of all the emission lines, is associated with the resonant excitation to the X transition. Due to these conditions, the shape of the PL spectrum is strongly modified by an enormous increase in the emission intensity of the negatively charged biexciton (XX−). The outstanding XX− increment can be explained in terms of the considerable n-type doping of the studied WSe2 ML, which results in a large reservoir of negative dark trions21,22,23,36. Due to the fact that a negatively charged biexciton is formed by a negative dark trion and a neutral bright exciton (X)21,22,23,36, the resonant excitation of the X complex should cause the creation of a large population of XX−, as is seen in Fig. 5c. The second effect is associated with the observation of several narrow peaks, parallel to the excitation laser, which are superimposed on the enhanced emission lines below the X peak. These peaks can be ascribed to phonon modes, whose intensities are enlarged due to the incoming resonance. As in the outgoing case, the increase in the intensity depends on the phonon symmetry, resulting in the most pronounced A\({}_{1}^{{\prime} }\) peak.

Fig. 5: Outgoing versus incoming resonant conditions.
figure 5

a Low-temperature (T = 5 K) PL spectrum of the WSe2 ML encapsulated in hBN flakes measured using excitation energy of 2.4 eV and power of 150 μW. The lines' assignments are as follows: X—neutral exciton; XX—neutral biexciton; TS and TT—singlet (intravalley) and triplet (intervalley) negatively charged excitons, respectively; XI and XD—momentum- and spin-forbidden neutral dark excitons, respectively; XX−—negatively charged biexciton; TD—negatively charged dark exciton (dark trion); X\({}_{{{{{\rm{E}}}}}^{{\prime\prime} }({{\Gamma }})}^{{{{\rm{D}}}}}\)—phonon replica of the XD complex. b, c False-colour maps of low-temperature (T = 5 K) optical response of the WSe2 MLs encapsulated in hBN under excitation of tuneable-energy lasers (excitation power ~ 10 μW) measured in the vicinity of the outgoing and incoming resonance conditions of Raman scattering with the X line, respectively. The colour scales in panels (b) and (c) have been normalised to the maximum intensity. The vertical red dashed lines denote the edge of the used long-pass filters.

Full size image

Alike in Fig. 3, we analyse the intensity profiles of the A\({}_{1}^{{\prime} }\) mode measured on the WSe2 ML under the incoming and outgoing resonance conditions of the Raman scattering with the X line, see Fig. 6. It can be seen that the A\({}_{1}^{{\prime} }\) evolutions can be nicely described by Lorentzian functions with similar linewidths, while their intensities differ considerably. The enhancement of the A\({}_{1}^{{\prime} }\) intensity is about 3 times larger in the outgoing resonance as compared to the incoming one, which can be because of several differences, such as: the strength of the exciton-phonon coupling in these two regimes, involvement of other excitonic states (e.g. dark trion or dark exciton) and of emission/absorption subprocesses, etc. For a given phonon mode, the energy separation between the outgoing and incoming resonances with a particular transition should be equal to this phonon energy. The obtained energy separation for the A\({}_{1}^{{\prime} }\) mode presented in Fig. 6 is of about 30 meV. This value is equivalent of 242 cm−1, which is very close to the A\({}_{1}^{{\prime} }\) wavenumber from Fig. 2 (~251 cm−1).

Fig. 6: RSE profiles of the A\({}_{1}^{{\prime} }\) mode.
figure 6

Intensity profiles of the A\({}_{1}^{{\prime} }\) mode measured on the WSe2 ML encapsulated in hBN flakes. Blue and orange circles correspond to the incoming and outgoing resonance conditions of the Raman scattering with the X line.

Full size image

Discussion

In the literature on the subject, a physical picture has been so far drawn that the exciton-phonon coupling in S-TMD MLs can be understood considering the symmetries of phonon modes with respect to the symmetries of orbitals associated with the involved transitions (excitons)42. For example, it has been established that the A\({}_{1}^{{\prime} }\) mode is enhanced when the excitation laser is in resonance with A and B excitons in S-TMDs, while the E\({}^{{\prime} }\) intensity is increased under resonanant conditions matching the energy of C excitons42,43. Our results, shown in Fig. 2, demonstrate a substantial difference in the intensity and complexity of the RSE spectra measured on the Mo- and W-based MLs. While the A\({}_{1}^{{\prime} }\) modes are extraordinarily enhanced for the WS2 and WSe2 MLs leading to the extremely rich RSE spectra (up to 16 identified phonon modes in the WS2 ML), the RSE spectra of the MoS2 and MoSe2 MLs are dominated by the modes involving acoustic phonons, e.g. 2LA, 3LA, …. In our opinion, this difference can be understood in terms of the excitonic states in the vicinity of their band gaps. Although S-TMD MLs share a very similar band structure, i.e. they are direct band-gap semiconductors with the minima (maxima) of the conduction (valence) band located at the K+ and K− points of the BZ4,5, the arrangement of the optically active transitions in the vicinity of their band gaps is different. A strong spin-orbit coupling results in spin-split and spin-polarised subbands in both the valence band (VB) and the conduction band (CB). Consequently, MLs of S-TMD are organised into two subgroups, i.e. bright and darkish, due to the type of the ground exciton state (bright and dark, respectively)49. In bright MLs, the excitonic recombination between the top VB and the bottom CB is optically active (bright), while the opposite happens darkish MLs. Nowadays, it is well established that for MLs encapsulated in hBN flakes, MoSe2 is bright, while MoS2, WS2, and WSe2 MLs are darkish (see refs. 12,28,32,34). It means that the observed difference in the intensity of the phonon modes in the RSE spectra, shown in Fig. 2, may coincide with a particular alignment of bright and dark states in S-TMD MLs.

The substantial enhancement of the LA phonons from the M points in the MoSe2 has been explained by the efficient phonon-assisted scattering occurring between the bright X exciton and the dark indirect exciton (IX) formed by an electron and a hole located at the Q and K points, respectively (see ref. 47 for details). Recently, similar description of the EPC, which occurs between the acoustic phonons from the M point and the continuum dark exciton states related to the optically forbidden transition at K and Q valleys, was demonstrated in thin layers of WS262. This confirms that for the acoustic phonons there is no significant difference between the MoSe2 (bright) and WS2 (darkish) MLs. Simultaneously, the small increase of the A\({}_{1}^{{\prime} }\)(Γ) intensity in the MoS2 ML suggests that the strength of the EPC is not remarkable. In contrast, the extraordinarily high intensity of the Raman-active A\({}_{1}^{{\prime} }\)(Γ) phonon modes (does not require any additional scattering between different points in the BZ) in the W-based MLs resulting in the very rich Raman spectra can not be described using the same approach, see Fig. 2. Note that the similar difference of one order of magnitude in increment of the A\({}_{1}^{{\prime} }\) peaks in outgoing conditions were reported for the MoSe2 and WS2 MLs44, which corroborate our results.

Particularly, similar results obtained for the WS2 and WSe2 MLs may suggest the involvement of the ground dark excitons in the efficient EPC in these materials. In the following, we speculate on the possible mechanisms, which may affect the EPC in these darkish materials. As the ground exciton state in W-based MLs is dark, the intensity of the bright exciton emission might be controlled by concurrent processes of the radiative recombination (emission of photons) and the relaxation process to the dark exciton and dark trion (depending on the doping level of the ML). It is known that the relative intensity of the bright and dark excitons in WSe2 is a function of temperature63 with a significant quenching of the X emission at low temperature. Consequently, large reservoirs of dark excitons and dark trions in W-based MLs are formed at low temperature12,28,34, while the analogous ones will be absent in bright MLs (ground exciton state is bright). These long-lived reservoirs may increase the probability of intravalley scattering between bright excitons and dark complexes, facilitating the emission of phonons from Γ points in the WS2 and WSe2 MLs. It suggests that the strength of the EPC in S-TMD MLs is not only associated with the symmetries of both phonons and electronic bands42, but may be affected by other phenomena, such as the relative order of excitonic states in the vicinity of direct transitions in K points of their BZ. The most complex are the results obtained for the MoS2 ML, which are very similar to the ones of the MoSe2 ML, i.e. the LA phonons from the M points are significantly enhanced. It is known that the MoS2 MLs encapsulated in hBN flakes exhibit dual character. The theoretically predicted structure of the CB and VB yields optical activity of the energetically lowest transition, while including the excitonic effects leads to the dark ground excitonic state32,50. This may indicate that the EPC in the MoS2 ML is more associated with its electronic band structure than with the excitonic one. The aforementioned phenomena, which may be responsible for the EPC in S-TMD MLs, requires solid theoretical analysis, which is beyond the scope of our experimental work.

Summarising, we have presented the investigation of the exciton-phonon interaction in four high-quality samples consisting of MoS2, MoSe2, WS2, and WSe2 MLs encapsulated in hBN flakes. By sweeping the excitation energy for a fixed value of the detection energy, we have observed an astonishing amplification of phonon-modes’ intensity, while the detection energy was in resonance with neutral exciton emission. Due to this phenomenon, the measured RSE spectra are composed of many phonon modes originating not only from the BZ centre (Γ points) but also from others (e.g. M points), which are followed by lines due to multiphonon processes. For the outgoing X resonance, we have found that the intensity profiles of the A\({}_{1}^{{\prime} }\) modes in the studied MLs resemble the emission lines of the corresponding X emission. The A\({}_{1}^{{\prime} }\) enhancements are extraordinarily strong for the WS2 and WSe2 MLs, but an analogous effect in the MoS2 and MoSe2 MLs are significantly smaller. For the Mo-based MLs, the modes involving acoustic phonons, such as 2LA and 3LA, are significantly enhanced. Moreover, we have observed that the A\({}_{1}^{{\prime} }\) intensity is also greatly strengthened since the excitation energy is in the vicinity of the X emission, leading to the incoming resonance condition. We have proposed that the difference in the obtained RSE measured on S-TMD MLs at low temperature can be understood in terms of the division of MLs into bright and darkish subgroups. These results shine new light on the exciton-phonon interaction in the MLs of S-TMDs, pointing out that not only the symmetries of phonons and excitons play an important role in this process, but also the type of their ground excitonic states.

Methods

Sample preparation

The studied samples were composed of four S-TMD MLs, i.e. MoS2, MoSe2, WS2 and WSe2, encapsulated in hBN flakes and supported by a bare Si substrate. The structures were obtained by two-stage polydimethylsiloxane (PDMS)-based64 mechanical exfoliation of bulk crystals of S-TMDs and hBN. A bottom layer of hBN in heterostructures was created in the course of non-deterministic exfoliation to achieve the highest quality. The assembly of heterostructures was realised via successive dry transfers of a ML and capping hBN flake from PDMS stamps onto the bottom hBN layer.

Experimental setup

Optical measurements were performed at low temperature (T = 5 K) using typical setups for the PL and PLE experiments. The investigated sample was placed on a cold finger in a continuous-flow cryostat mounted on x–y manual positioners. The non-resonant PL measurements were carried out using 514.5 nm (2.41 eV) and 532 nm (2.33 eV) radiations from the continuous wave Ar+ and Nd:YAG lasers, respectively. To study the optical response of a ML as a function of excitation energy, i.e. to measure PLE spectra, two types of tuneable lasers were used: dye lasers based on Rhodamine 6G and DCM, and a Ti:Sapphire laser. The excitation light was focused by means of a 50 × long-working distance objective that produced a spot of about 1 μm diameter. The signal was collected via the same microscope objective, sent through a monochromator, and then detected by a charge-coupled device (CCD) camera. Measurements of the low-energy part of the RSE spectra, i.e. from around 15 meV from the laser line, were carried out using ultra steep edge long-pass filters mounted in front of the spectrometer.