Dilute magnetic semiconductors (DMSs) are typically made by doping semiconductors with magnetic transition metal elements. Compared to the well-understood bulk and thin film DMS, the understanding of the magnetic element doping effects in semiconducting quantum dots (QDs) is relatively poor. In particular, the influence of the dopant locations is rarely explored. Here, we present a comprehensive study of the effects of Mn doping on the electronic density of states of PbS QDs. Based on the results observed by scanning tunneling microscopy, X-ray diffraction, electron paramagnetic resonance, and density functional theory calculations, it is found that the Mn doping causes a broadening of the electronic bandgap in the PbS QDs. The sp-d hybridization between the PbS host material and Mn dopants is argued to be responsible for the bandgap broadening. Moreover, the locations of the Mn dopants, i.e., on the surface or inside the QDs, have been found to play an important role in the strength of the sp-d hybridization, which manifests as different degrees of the bandgap change.
Dilute magnetic semiconductors (DMSs) are typically semiconductors doped with transition metal elements, typically Mn, in order to induce magnetic properties in the semiconducting host.1 The presence of the magnetic ions in DMS can lead to a spin-spin exchange interaction between the localized magnetic moments and the band electrons2 and results in the unique properties of DMS materials, such as giant Faraday rotation effects near the band edge3 and change of the bandgap.4 By 1982, most of the DMS materials studied were II-VI semiconductors.2,5 Later on, the field gradually expanded from II-VI to III-V semiconductors6 due to the observed high Curie temperature in III-V based DMS materials.7
Compared to the well understood bulk DMS, transition metal element doping in quantum dots (QDs) is relatively unexplored. By year 2005, doping into semiconducting QDs was still considered one of the major challenges8 in pursuing QD DMS. Many of the failed efforts have unclear reasons. However, it was typically attributed to the “self-purification” process, an allegedly intrinsic mechanism whereby impurities are expelled toward the surfaces of the QDs.9 Despite the difficulty, many studies showed that the magnetic dopants in semiconducting QDs could have significant influences on various applications. For example, mid-gap states due to the magnetic dopants were found to have profound effects on the transport and exciton recombination phenomena.10–12 Mn doping in various semiconducting QDs has been shown to have positive impacts on the quantum dot sensitized solar cell (QDSSC) performance.13–15 It was further revealed that the surface Mn dopants in PbS QDs in contact with Zn2SnO4 nanowires in QDSSCs provide efficient pathways for electron injection, which results in strongly enhanced incident photon-to-current efficiency (IPCE) up to 700%.15
The locations of the Mn dopants in semiconducting QDs have been discussed and reported.8,16–19 However, although recognized by few researchers,16 the influence of the dopant location on the host QD properties is rarely discussed in depth.17,20–22 In existing studies, it has been reported that the luminescent properties of Mn2+ in QDs depend on (1) preparation modes and conditions, e.g., various Mn concentrations;19 (2) aging or annealing treatment;23 (3) surface modification;24,25 or (4) different preparation methods.26,27 These results indicate that the Mn doping configuration plays an important role in affecting Mn2+ luminescence, and yet, not many systematic studies address this effect on the host QD properties. These issues bring up the question regarding the physical properties of the QDs: “How does the dopant's location influence the electronic band structure of the host semiconducting QDs?” This work presents scanning tunneling microscopy and spectroscopy (STM/S) measurements of the effects of Mn doping on the energy band structures of PbS QD thin films, grown by pulsed laser deposition (PLD). In conjunction with the STM/S measurements, theoretical density of states (DOS) calculations are performed to show that Mn dopant locations result in a different level of widening of the electronic bandgap.
PbS is a IV-VI semiconductor, possessing a bulk bandgap of 0.4 eV.28 The exciton Bohr radius in PbS is ∼20 nm,29 making it a promising candidate for tuning the absorption gap in the nanoparticle form for photovoltaic applications.30,31 PbS QDs have been used to produce one of the highest power conversion efficiencies (PCEs) in QDSSC at 10.7% in 201532 and at 9.9% in 2016.33 Mn doping in PbS QDs has been reported experimentally but without much emphasis on the electronic properties of the host PbS QDs.15,30,34–39 A theoretical calculation revealed that the Mn dopant in PbS QDs widens the bandgap of the materials through the sp-d hybridization between the PbS bands and Mn d states.40
The cubic rock salt crystal structure of PbS and Mn:PbS is clearly revealed in the XRD data shown in Fig. 1(a). For PbS QD thin films (black curve), the three most intense peaks correspond to PbS(111) at 2θ = 26.02°, PbS(200) at 2θ = 30.13°, and PbS(220) at 2θ = 43.15°; while the two relatively low intense peaks are associated with Pb(111) at 2θ = 31.26° and Pb(200) at 2θ = 36.24°. The XRD data of the PbS QD films indicate that the synthesized thin films are mostly crystalline PbS with traces of pure Pb aggregates. On the other hand, for the Mn:PbS QD thin films (red curve), as shown in Fig. 1(a), the XRD spectra showed peaks associated with PbS(111), PbS(200), and PbS(220) planes, along with a Ta(110) peak due to the Ta foil strips for mounting samples.
To study the bandgap and the electronic local density of states (LDOS) of the synthesized thin films, UHV STM/S was performed. The STM topographies of PbS and Mn:PbS QD thin films are shown in Figs. 2(a) and 2(c), respectively. The average sizes of both synthesized PbS and Mn:PbS QDs are 9.7 ± 2.3 nm and 9.6 ± 3.6 nm, respectively. dI/dV mappings of the PbS and the Mn:PbS QD thin films taken simultaneously with the topography are shown in Figs. 2(b) and 2(d), respectively. In both samples, two regions with distinct contrast were revealed in the dI/dV mappings. Interestingly, the contrast domains coincide with the topographical QD features, indicating that there are limited electronic interactions between adjacent QDs. By positioning the STM tip over the high and low contrast regions with similar QD sizes, averaged dI/dV spectra near the Fermi energy (±2.0 V) were measured and are shown in Fig. 2(e). For the PbS thin films, the first thing to notice is that the dI/dV spectrum measured in the low contrast regions exhibited a metallic behavior—no bandgap. These metallic domains are believed to be the excess Pb aggregates found in XRD data. The second thing to notice is that the dI/dV spectrum measured in the high contrast regions of the PbS thin films shows a semiconducting feature—a small bandgap of 0.88 eV, which is consistent with the size dependent bandgap for 9 nm size PbS QD.41 On the other hand, the two dI/dV contrast domains found in the Mn:PbS thin films exhibited two distinct bandgaps −0.97 eV and 1.39 eV. The bandgaps are clearly larger compared to those measured in the PbS QD thin films. The question is what is the origin of the two regions with distinct energy bandgaps in the Mn:PbS samples? A few possible origins will be discussed: (a) formation of MnS; (b) quantum confinement effect; (c) variation in doping levels; and (d) locations of the dopants. First of all, the formation of MnS can be ruled out due to the absence of a large bandgap (3.1 eV for α-MnS) and of XRD peaks related to the MnS materials.42 Second, the quantum confinement effect plays a role,43–45 however negligible in this study. With quantum confinement effects, the bandgap of QD shifts to a higher value due to the confinement-induced blueshift. On the other hand, the compression of the exciton radii in small QDs induces an increase in the Coulomb interaction which has been shown experimentally to create a redshift in the bandgap.45 In this study, STM measurements do not induce the excitons. Thus, only blueshift is expected for smaller size QDs. For the scanning tunneling spectroscopy (STS) measurements done in this work, the quantum confinement effects are mitigated due to the similar sizes of QDs [9 ± 1 nm, as seen in Fig. 2(f)] measured. In addition, if the quantum confinement effects are the dominating effects on the bandgaps, one would expect to see a variety of bandgaps and dI/dV contrast as a function of the QD size in one sample. In the data, only two distinct dI/dV contrast and two associated bandgaps are observed, further confirming that the quantum confinement effects are not dominating effects on the observed two bandgaps in the Mn:PbS sample.
The variation in doping levels is also ruled out by the symmetric XRD peak. In particular, if various doping levels exist across the samples, the XRD peak broadening is expected to be a Gaussian like peak shape due to the normal distribution of the doping levels, assuming that different doping levels would result in slightly different lattice constants. In the XRD analysis, the PbS(200) peak in the Mn:PbS samples exhibits a good agreement with a Lorentzian fitting function as shown in Fig. 1(b), indicating that the QDs are uniform in doping levels. Additionally, as the Mn atom concentration is roughly 4%, there is a high likelihood that an interaction between Mn atoms could influence the bandgap. At this point, an obvious conclusion can be made that the Mn doping in PbS causes the bandgap to increase, consistent with previous reports,15,36,40 while the variation in the bandgap is to be discussed in the following paragraphs. A density functional theory (DFT) study explained the widened bandgap in Mn:PbS as originating from the sp-d hybridization between the PbS and the Mn.40 In the same study, it is found that the Mn dopants on the surfaces of QDs have small effects with regard to the altering of the absorption spectrum while the dopants inside the QDs showed a significant influence on the absorption bandgap. Hereafter, the focus shifts to discussing how the dopant locations can affect the electron DOS measured by STS.
Two doping locations are considered: doping on the surface (sur) and doping inside (in) the QDs, respectively. First of all, the electron paramagnetic resonance (EPR) spectrum of the Mn:PbS, as shown in Fig. 3, exhibits six peaks with observable shoulders, indicating that the Mn2+ dopants are located both on the surface and inside the QDs.37 By fitting the EPR data with the combination of two sets of six-first-derivative-Lorentz-peak functions, two hyperfine splitting constants are determined. For the major EPR peak positions in Fig. 3, the hyperfine splitting constant was determined to be A = 96.7 ± 0.1 G, with peak positions indicated by the blue triangles in Fig. 3, which agrees well with the Mn2+ on surfaces of various types of semiconducting materials,37,46 whereas the shoulder peaks exhibit a hyperfine splitting constant of B = 76.0 ± 0.7 G, with peak positions indicated by the green diamonds in Fig. 3, which agrees well with Mn2+ located inside the QDs.37,46 This result indicates that the majority of the Mn dopants are on the surfaces of the QDs while the dopants inside the QDs are not negligible. This result confirms that serious consideration is to be given to both possible doping locations. It is worth noting that although the STM/S is typically considered as a surface sensitive technique, probing Mn dopants with depths up to 2 nm is possible due to the large spatial dopant wave function overlap and has been reported in Mn:GaAs,47,48 allowing the surface sensitive STM/S to probe the Mn dopants located a few nm deep inside the QDs.48
To gain insight into the effects of the two possible doping locations on the DOS, DFT calculations of the PbS and Mn:PbS thin films were carried out. Energy bandgaps of the two doping locations were calculated to be compared with the STM/S results. The calculated DOS near the Fermi energy of the two doping locations is shown in Fig. 4. It is clear that for both locations considered, no observable mid-gap states are observed in the DFT calculated DOS, which is consistent with our STM/S measurements. With a closer look, it is also clear that both the (in) and (sur) doping locations result in a widening of the bandgap. Recall that the bandgaps measured by STM/S (Fig. 2) were found to be increased compared to those measured on undoped PbS QDs. The key finding here is that the two doping locations result in different levels of bandgap broadening. With these results and discussion, it is concluded that both the doping locations in Mn:PbS resulted in the broadening of the bandgaps, while the (in) doping location results in larger bandgap broadening compared to the (sur) doping location. Since the bandgap broadening is due to the sp-d hybridization between the PbS and Mn, this result indicates that the locations of the Mn dopants exhibit different levels of the hybridization in Mn:PbS QDs. In particular, the hybridization between the surface Mn dopants and the host PbS QDs is smaller compared to that between the inner Mn dopants.
In summary, the broadening in the bandgap of PbS QDs upon Mn doping was reproduced using PLD synthesis. The variation in bandgaps measured on different QDs in the synthesized Mn:PbS thin films was observed by STM/S. Combining data analysis of XRD, STM/S, EPR, and DFT calculations, it is concluded that the observed variation in bandgap broadening is the result of the Mn dopant locations, inside versus surface of the QDs. These results provide clear information on how the transition metal element doping locations in semiconducting QDs influence the electronic DOS of the host QDs.
A single crystal substrate of As:Si(100) was placed inside a PLD vacuum chamber with a base pressure of Torr. Then, a Nd:YAG laser, operating at a wavelength of 266 nm and a pulse energy of 185 J, was focused with a fused silica lens of focal length of 30 cm through a quartz window onto a PbS or Mn:PbS target inside the PLD chamber. The pulsed laser creates a plasma plume which is directed toward the As:Si(100) substrate. The deposition process lasts for 30 min without heating the substrate.
The synthesized films are removed from the PLD chamber and loaded into an Omicron low temperature scanning tunneling microscope (LT-STM) system. The samples are annealed at 375 °C for 15 min at a base pressure of mbar before being transferred for STM measurements. The STM measurements are carried out at liquid nitrogen temperature after the thermal drift is minimized. Each sample is then scanned in the constant current mode with the set point of and for PbS and and for Mn:PbS.
X-ray diffraction (XRD) measurements were carried out with a Rigaku Smart Lab diffractometer equipped with a Cu-kα source. All XRD was performed at room temperature in air. Electron paramagnetic resonance (EPR) measurements were performed with a continuous wave electron paramagnetic resonance (CW-EPR) spectroscope. EPR measurements were performed with Bruker EMX in the X-band at room temperature to determine the presence and locations of the Mn dopants. A FEI Quanta FEG 450 field emission scanning electron microscope (SEM) equipped with an Oxford Inca energy dispersive x-ray spectroscopy (EDS) system was used to measure the Mn doping percentage. The EDS measurements indicate that the Mn concentration is ∼4%.
The electronic properties were studied with the Vienna Ab-initio Simulation Package (VASP) using the density functional theory (DFT) method with the PBE exchange correlation functional.49–54 The projector augmented wave (PAW) pseudopotential for DFT was used with an energy cutoff of 500 eV for the plane wave basis functions.55,56 The geometries of PbS films were prepared from a rock salt structure with a (111) surface plane and Pb atoms constituting the termination layers. The Pb28S24 supercell was implemented to construct approximately a 20 Å thin film consisting of 7 and 6 monolayers of Pb and S atoms, respectively, with the 3 core layers frozen at PbS bulk crystal geometry and the 10 surface layers were set to relax to obtain the surface structure. The vacuum between the films was set to 15 Å to avoid interactions. The Mn atoms were symmetrically placed in the surface layers with the substitution of the Pb atom position. Doping the outermost surface layers corresponds to Mn atoms being on the surface, and doping the inner surface layers corresponds to Mn atoms being inside. We used an 8 × 8 × 1 k-point mesh generated according to the Monkhorst–Pack scheme with 0.1 eV Gaussian broadening for the geometry optimization.57,58 The DOS plots were constructed for the surface layers using the same k-point mesh. The spin-orbit coupling effect was also included in the analysis. The reason a thin film supercell, rather than the QD geometry, was used for the DOS calculation is that the largest QDs calculable are only a couple of nm which are not comparable to the sizes of the QDs observed in experiment (∼9 nm). Different sizes of the QDs may further introduce complications due to size effects on sp-d hybridization. The thin film supercell calculations illustrated clearly the Mn atom location influences on the DOS for a surface dopant versus a bulk dopant.
J.T., Y.D., and T.Y.C. acknowledge the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering for financial support (DEFG02-10ER46728) of this research. A.J.Y. acknowledges graduate fellowship support from the National Science Foundation and the University of Wyoming EE-Nanotechnology Program (NSF-DGE-0948027) and from Wyoming NASA Space Grant Consortium, NASA Grant No. NNX15AI08H.