Reduction of Mn³⁺ to Mn²⁺ and near infrared plasmonics from Mn–Sn codoped In₂O₃ nanocrystals

Abstract

Colloidal Sn doped In₂O₃ nanocrystals have gained considerable attention for exhibiting surface plasmon resonance (SPR) which can be easily tuned in the near to mid infrared region by controlling the dopant concentration. Codoping these NCs with magnetic ions such as Fe³⁺, Mn^2+/3+ can help develop interactions between delocalized electrons and localized magnetic spins which are required for spin-based applications. We prepared colloidal Mn–Sn codoped In₂O₃ nanocrystals with a diameter of ∼6–7 nm, starting from a Mn³⁺ precursor for Mn doping. Detailed characterization including Q-band electron paramagnetic resonance spectroscopy show that Mn exists in the 2+ oxidation state in both Mn-doped and Mn–Sn codoped In₂O₃ nanocrystals, in spite of using the Mn³⁺ precursor. This aliovalent Mn²⁺ doping along with charge neutralization through oxygen vacancies are energetically more favorable when compared to isovalent Mn³⁺ substitution in the In₂O₃ lattice. The SPR band decreased both in intensity and energy with increasing Mn content in the Mn–Sn codoped In₂O₃ nanocrystals. This is because Mn²⁺-doping introduces a hole in the lattice promoting electron–hole recombination reducing the free electron density. Also, Mn²⁺ dopants scatter the electrons thereby broadening the SPR band.

---

1. Introduction

Transparent conducting oxides (TCO) like indium tin oxide (ITO) and fluorine doped tin oxide (FTO) are high band gap materials with a significant density of conduction band (CB) electrons (e) which gives them a combination of otherwise mutually exclusive properties like transparency to visible light and high electrical conductivity.¹ Doping magnetic ions in these TCOs can lead to the much required interaction of delocalized electrons with localized magnetic spins. Such interactions can give rise to ferromagnetism at a higher temperature^2–4 along with possible magneto-optic and magneto-electric properties.⁵ Doped oxides showing such properties are popularly known as dilute magnetic semiconductor oxide (DMSO)^6,7 and can act as future spintronic materials.^8–10 However, it has been difficult to observe ferromagnetism from high quality single crystals of magnetically doped TCOs, instead, ferromagnetism have been found in their polycrystalline films.^8,11–13 Other than magnetism, the doping of magnetic transition metal ions in TCOs also affects their optical and electrical behavior.^12–16 Most of the TCO films exhibit nonmetallic behavior in electrical conductivity data which would suggest localization of charge carriers but extensive contribution from insulating grain boundaries makes it difficult to judge whether carriers are localized or delocalized within a given grain.¹⁷

Interestingly, nanocrystals (NCs) of ITO exhibit a strong surface plasmon resonance (SPR) band characterizing delocalized electrons within a NC. Such characterization of delocalized electrons using SPR band do not suffer from the grain boundary problem, unlike electrical measurements.¹⁸ In addition to SPR band, recent reports from our group involved codoping of transition metal ions (Fe³⁺, Mn²⁺) in Sn-doped In₂O₃ NCs^18–20 showing carrier mediated magnetic coupling, where distant magnetic ions interact with each other via the delocalized CB e. Unfortunately, in both the cases, the SPR band generated due to aliovalent Sn⁴⁺ doping was dampened due to scavenging of free electrons by Fe³⁺ (facilitating partial conversion of Fe³⁺ to Fe²⁺), Mn²⁺ (through electron–hole recombination of Mn²⁺-hole with Sn⁴⁺-electron) and scattering of electrons by ionized dopants.^18,20,21

Therefore, we were encouraged to replace Mn²⁺ dopant by Mn³⁺ which would have ideally led to isovalent doping by replacing In³⁺ in In₂O₃ lattice and hence retaining the strong SPR properties in Mn–Sn codoped In₂O₃ NCs. Here we show that even after starting with Mn³⁺ precursor aimed at isovalent doping, our final Mn doped In₂O₃ and Mn–Sn codoped In₂O₃ NCs were aliovalently doped i.e. Mn³⁺ reduces to Mn²⁺. Possible reasons for conversion of Mn³⁺ to Mn²⁺ along with its influence on SPR have been studied.

2. Experimental section

2.1 Chemicals

All the chemicals used were acquired commercially. Indium(III) acetylacetonate (Sigma-Aldrich, purity ≥99.99%), manganese(III) acetylacetonate (Sigma-Aldrich, purity >99.99%) and tin(IV) bis(acetylacetonate)dichloride (Sigma-Aldrich, purity 98%), oleylamine (Sigma-Aldrich, purity 70%), toluene (Rankem, 99.5% purity), methanol (Rankem, 99.9% purity), tetrachloroethylene (Sigma-Aldrich, purity ≥99%).

2.2 Colloidal synthesis of Mn–Sn codoped In₂O₃ NCs

We report here the first synthesis strategy for making colloidal Mn–Sn codoped In₂O₃ NCs using Mn³⁺ precursor. The colloidal Mn–Sn codoped In₂O₃ NCs were synthesized through a simple single pot strategy using standard schlenk line technique, similar to our previous reports.^18–20 Indium(III) acetylacetonate, manganese(III) acetylacetonate, and tin(IV) bis(acetylacetonate) dichloride were taken in stoichiometric amounts to synthesize NCs of desired compositions. For example, to synthesize 10% Mn–10% Sn codoped In₂O₃ NCs, 1.2 mmol In(III), 0.15 mmol Mn(III), and 0.15 mmol Sn(IV) precursors were mixed with 10 mL oleylamine in a 50 mL three-necked round-bottom flask and degassed at room temperature using N₂ atmosphere and vacuum conditions alternatively for 30 min. The temperature was then raised to 100 °C under vacuum conditions for 30 min followed by heating to 220 °C at ∼10 °C min⁻¹. The reaction was then continued at 220 °C for 5 h under N₂ atmosphere along with magnetic stirring. Subsequently, the reaction mixture was allowed to cool to room temperature followed by addition of excess methanol (∼30 mL) as a non-solvent to precipitate the NCs. The precipitated NCs were centrifugated at 6000 rpm for 6 min and the washing process was repeated twice to get rid of excess oleylamine. These oleylamine capped NCs were dispersed in different non-polar solvents like tetrachloroethylene, toluene and chloroform to study various properties.

2.3 Characterization

Energy dispersive X-ray (EDX) analysis was carried out using Zeiss Ultra Plus Scanning Electron Microscope and inductively coupled plasma optical emission spectroscopy (ICP-OES) was carried out using Arcos M/s. Spectro, Germany for determining elemental composition of Mn–Sn codoped In₂O₃ NCs. Perkin Elmer, Lambda-950 UV/vis spectrometer and Shimadzu UV-3600 Plus UV-Vis-NIR spectrophotometer were used to record UV-vis-NIR absorption spectra of NCs dispersed in tetrachloroethylene (TCE). Powder X-ray diffraction (XRD) patterns of NCs were obtained on a Bruker D8 Advance X-ray diffractometer using Cu Kα (1.54 Å) X-ray source. Transmission Electron Microscopy (TEM) was performed on JEOL JEM 2100F and Tecnai G2 FEI F12 TEM microscope being operated at 200 kV. Electron paramagnetic resonance (EPR) spectra were recorded on a JEOL JES-FA200 ESR spectrometer with X and Q band attachment. Electrical conductivity was determined on ∼1.5 mm thick pellets using a Keithley four probe conductivity instrument consisting of Model 6220/6221 Current Source and Model 2182A nanovoltmeter. All the pellets were annealed in N₂ atmosphere at 200 °C for 2 hours prior to conductivity measurements. Magnetic data were obtained from a SQUID magnetometer (Quantum Design MPMS XL-7 Magnetometer). Zero-field cooled (ZFC) and field-cooled (FC) data were acquired by varying the temperature between 2 and 300 K at 100 Oe magnetic field after cooling the samples in zero field or in a 100 Oe field, respectively.

3. Results and discussion

3.1 Structure and morphology

Elemental analysis of Mn–Sn codoped In₂O₃ NCs obtained using EDX has been summarized in Table 1 (representative EDX spectra are shown in Fig. S1 of ESI†). We note here that, EDX analyses were done by recording multiple spectra over different regions of a given sample, and the reported composition was then calculated after averaging out the compositions obtained from multiple spectra. Such averaging is done to reduce the error in the measured composition, particularly at smaller doping levels where EDX analysis can lead to large errors. Furthermore, elemental compositions were obtained using ICP-OES, which agrees with the EDX data, as shown in Table ST1 of the ESI.† All these elemental analyses show that product NCs exhibit compositions similar to the precursor ratios. Therefore, for simplicity, we have used the nominal precursor ratios throughout the manuscript. However, EDX and ICP-OES cannot distinguish between metal ions doped in the core of NCs and those present as secondary impurity phases.

Table 1 Elemental composition (atomic ratio) of Mn–Sn codoped In₂O₃ NCs. EDX data reveals that NC composition is almost same as that expected from the precursor ratios

In : Mn : Sn (atomic ratio)
Precursor ratio	EDX analysis
95 : 5 : 0	95.5 : 4.5 : 0
85 : 10 : 5	87.3 : 8.3 : 4.5
85 : 5 : 10	84.1 : 5.5 : 10.4
89 : 1 : 10	90.2 : 0.8 : 9.0
99 : 1 : 0	98.9 : 1.1 : 0
80 : 10 : 10	83.8 : 8.1 : 8.1
90 : 5 : 5	89.4 : 5.6 : 5.0

---

Fig. 1a compares XRD patterns of different Mn doped In₂O₃ NCs. It can be clearly observed that all doped and undoped In₂O₃ NCs exhibit the same cubic bixbyite structure (space group Ia3, 206) as that of the reference bulk In₂O₃ (JCPDS number 88-2160) with no signs of impurity peaks or secondary phases. Additionally, Mn doping causes a slight shift of diffraction angles to higher values or in other words, a decrease in interplanar distances following Bragg's law of diffraction. Previous literature has attributed this shift extensively to the lattice doping of Mn ions. It is important to note that while Mn³⁺ has a smaller crystal radii than In³⁺ ion (r_Mn³⁺ = 65 pm, r_In³⁺ = 79 pm), Mn²⁺ has a crystal radii greater than that of In³⁺ (r_Mn²⁺ = 82 pm).²² Therefore, ideally Mn³⁺ doping would result in lower interplanar distances in In₂O₃ whereas Mn²⁺ doping would cause them to be greater than that in In₂O₃. However, Mn²⁺ ion is not isovalent with In³⁺ and replaces In³⁺ in the lattice along with an oxygen vacancy to compensate the charge imbalance.²³ An increase in oxygen vacancy concentration would also be reflected in the position of peaks in XRD pattern as loss of atoms would cause the lattice to shrink (lower interplanar distances) and thus higher angles of diffraction. This creates a confusion over the plausible oxidation state of Mn ion and due to this reason, the extent of presence of mixed valency in Mn doped In₂O₃ system has been largely inconclusive as per the literature reports from different groups.^18,23–26

Fig. 1 Powder X-ray diffraction (XRD) patterns of (a) Mn doped In₂O₃ NCs, and (b) Mn–Sn codoped In₂O₃ NCs. Comparison of the XRD patterns of NCs with bulk In₂O₃ reference reveal cubic bixbyite structure for all samples without presence of any secondary phase.

Sn doped In₂O₃ NCs (Fig. 1b) exhibit lattice parameters similar to the undoped sample which is consistent with earlier reports.¹⁹ Also, Mn–Sn codoped In₂O₃ NCs (Fig. 1b) have similar lattice parameters as compared to their only Mn doped counterparts. This observation suggests us that the oxidation of Mn state remains largely the same in both presence and absence of Sn codopant, even though Sn⁴⁺ doping provides an extra electron that can reduce any Mn³⁺ to Mn²⁺ (reduction potential +1.51 V).²⁷ Such a scenario can possibly occur because Mn³⁺ in the system are already reduced to Mn²⁺ by surrounding oxygen vacancies in Mn-doped In₂O₃ NCs, in the absence of Sn codopant. This possibility has been discussed in detail in later part of the manuscript.

Fig. 2 shows transmission electron microscopy (TEM) images for undoped and different Mn–Sn doped In₂O₃ NCs with all NCs being nearly spherical in shape. Undoped In₂O₃ NCs are approximately 9.1 nm in size whereas Mn and/or Sn doped In₂O₃ NCs have an average diameter of 6–7 nm which is in good correlation with that obtained from XRD pattern using Debye–Scherrer equation.²⁸ High resolution TEM (HRTEM) images of undoped and 10% Sn doped In₂O₃ NCs depicted as insets of Fig. 2a and b show highly crystalline single phase NC with lattice fringes. Calculations result in interplanar distances of 2.93 Å and 2.94 Å for undoped and 10% Sn doped In₂O₃ NCs respectively corresponding to {222} lattice planes of In₂O₃ cubic bixbyite structure. Representative HRTEM images from 1% Mn doped and 1% Mn–10% Sn codoped In₂O₃ NCs (insets to Fig. 2c and d) also show lattice fringes with interplanar distances of 2.93 Å and 2.94 Å respectively corresponding to {222} plane. These values of interplanar distances for undoped and doped In₂O₃ NCs obtained from HRTEM match very closely with those calculated from their corresponding XRD patterns which confirms that our NCs are indeed phase pure and possibility of any secondary phase can be ruled out.

Fig. 2 Transmission electron microscopy (TEM) images along with high resolution TEM (HRTEM) images in the inset for (a) undoped In₂O₃ NCs showing spherical NCs with an average diameter ∼9.1 nm. (b) 10% Sn doped In₂O₃ NCs having spherical NCs with ∼6.5 nm average diameter (c) 1% Mn doped In₂O₃ NCs showing nearly spherical NCs with average diameter ∼7.1 nm (d) 1% Mn–10% Sn codoped In₂O₃ NCs showing narrow size distribution of nearly spherical NCs with average diameter ∼7.0 nm.

3.2 Oxidation state and local structure of Mn using EPR spectroscopy

EPR spectroscopy is an excellent tool to probe oxidation state and local structure around a paramagnetic ion.²⁹ We used Q-band EPR spectroscopy to elucidate the oxidation state of Mn in both Mn-doped and Mn–Sn codoped In₂O₃ NCs. Fig. 3 compares Q-band EPR spectra of 1% Mn doped and 1% Mn–10% Sn codoped In₂O₃ NCs. Both the EPR spectra consist of six hyperfine line pattern centered at g = 2.003. This sextet splitting pattern is characteristic of the presence of Mn²⁺ ions which arises as a result of hyperfine coupling between nuclear spin of Mn (I = +5/2) and 3d electrons of Mn²⁺ ion.³⁰ The location of these Mn²⁺ ions can be estimated through the hyperfine coupling constant which is approximately ∼7.1 mT. This value of hyperfine coupling constant is similar to those observed in Mn doped bulk oxides suggesting that Mn–O bonds show significant covalent behavior³¹ and hence most of our Mn²⁺ ions are lattice doped rather than being present on the surface, for which hyperfine coupling constant is expected to be much higher.³²

Fig. 3 Q-Band EPR spectra of 1% Mn and 1% Mn–10% Sn codoped In₂O₃ NCs showing hyperfine structure consisting of six lines centered at g = 2.003 for the presence of Mn²⁺ ions.

On the other hand, Mn³⁺ belongs to a non-kramer category (integral spin) and suffers with high zero field splitting energy which often makes it silent to X-band (∼9.35 GHz) EPR. However, we have employed here Q-band EPR (∼35 GHz) that can probe Mn³⁺ ion. It is known that Mn³⁺ exhibits an EPR signal centered around g = 2.21 under high spin environment of In₂O₃.³³ Therefore, the absence of Mn³⁺ EPR signal for both our Mn-doped and Mn–Sn codoped NC samples suggests that the oxidation state of Mn is 2+ with insignificant amount (or absence) of Mn³⁺ ions. This is interesting as we started off with a Mn³⁺ precursor and ended up with Mn²⁺ in the final NC product.

3.3 Insights on conversion of Mn³⁺ to Mn²⁺

The observed conversion of Mn³⁺ in precursor to Mn²⁺ in product NC might have multiple correlated origins, for example, (i) Mn³⁺ reduces to Mn²⁺ because of applied reaction conditions and/or (ii) Mn³⁺ is not stable after incorporation into the In₂O₃ lattice. We discuss here both the possibilities. A control experiment was performed to study the spontaneity of this conversion under the same reaction conditions but in the absence of In₂O₃ lattice. This control reaction was carried out with Mn³⁺ precursor, in the absence of In(III) and Sn(IV) precursors, and applying rest of the synthesis strategy same as that for Mn–Sn codoped In₂O₃ NCs. XRD pattern of the product obtained from this control experiment (shown in Fig. 4) is compared with reference patterns of bulk MnO, Mn₂O₃ and Mn₃O₄. XRD pattern of the product resembles more to that of the MnO in terms of 2θ values and significantly different when compared to both Mn₂O₃ and Mn₃O₄. These results suggest that Mn³⁺ precursor in the control reaction preferably reduces to Mn²⁺ forming MnO as the preferred composition even in the absence of In₂O₃. Though the energetics involved in the formation of MnO and Mn-doped In₂O₃ would be very different, but the results confirm that Mn²⁺ is more stable compared to Mn³⁺ in both synthesis. In addition to the observed preference for Mn²⁺ oxidation state even in the absence of In₂O₃ lattice, theoretical studies of Zunger et al.¹⁰ also suggested the possibility of reduction of Mn³⁺ to Mn²⁺ after incorporation into In₂O₃ lattice. They calculated the position of defect levels created by different 3d transition metals after substitution of In³⁺ in In₂O₃ lattice.¹⁰ Schematic in Fig. 5 shows the case of Mn³⁺ ion doped in In₂O₃ lattice, after adapting the figure from ref. 10 along with our modifications. Crystal structure of In₂O₃ has two substitution sites b and d, both of which are nearly octahedral in symmetry.¹ Under the nearly octahedral symmetry of lattice sites, the d levels of dopant gets split into sets of t and e levels where the spin configuration (spin up or spin down) further splits them into t₊, e₊, t₋ and e₋ levels. For a Mn³⁺ ion with 4d-electrons in high-spin configuration the highest occupied two degenerate e₊ levels are only half filled and therefore undergoes a Jahn–Teller distortion³⁴ lifting the degeneracy. Both the non-degenerate e₊ level lie in the mid-gap region of the host, as shown in Fig. 5. The e₊ with lower energy is occupied, whereas the other e₊ level is unoccupied. This unoccupied e₊ level can accept an electron from both the conduction band minimum (CBM) of In₂O₃ and oxygen defect states that are energetically closer to CBM. Oxygen vacancies are known to form shallow electron donor levels near the CBM of In₂O₃.³⁵ Energetically favorable electron transfer from oxygen vacancies (donor levels) to unoccupied e₊ (acceptor levels) can reduce Mn³⁺ to Mn²⁺. This possible electron transfer is one of the important reasons for stabilizing Mn²⁺ in In₂O₃ lattice, which is known to form oxygen vacancies.

Fig. 4 Comparison of XRD pattern of the sample obtained from control experiment using manganese(III) acetylacetonate with bulk MnO, Mn₂O₃ and Mn₃O₄ reference. The major peaks in XRD pattern correspond to MnO suggesting dominance of +2 oxidation of Mn in the product in spite of using Mn³⁺ precursor.

Fig. 5 Schematic showing the mechanism for reduction of Mn³⁺ to Mn²⁺ in In₂O₃ lattice. Donor–acceptor transition between 3d levels of Mn³⁺ (acceptor) and those created by oxygen vacancies/Sn⁴⁺ codoping (donor) leads to conversion of Mn³⁺ to Mn²⁺ ions. The figure has been prepared by using the theoretical calculation data reported by Zunger et al. in ref. 10.

It is worthy to note that oxygen vacancy concentration in undoped In₂O₃ NCs is usually low (<1% of the total oxygen concentration),¹ so ideally, in only Mn doped In₂O₃ NCs, the number of electrons released by creation of oxygen vacancies won't suffice for the conversion of all Mn³⁺ to Mn²⁺ with charge imbalance further complexing the scenario. However, previous literature has suggested that Mn²⁺ doping in In₂O₃ is usually assisted with creation of additional oxygen vacancies²³ following eqn (1)

In the previous equation, represents that substitution of In atoms at an In site leaving a neutral charge on the site (denoted by x), means that a Mn²⁺ ion has substituted an In³⁺ ion at an In site and left a negative charge (′) on the In site and represents creation of an oxygen vacancy at an oxygen site leaving +2 positive charge (denoted by ˙) on the oxygen site.

Therefore both the possibilities, (i) converting Mn³⁺ to Mn²⁺ under our reaction conditions even in the absence of In₂O₃, and/or (ii) reduction of Mn³⁺ to Mn²⁺ by accepting electrons from oxygen vacancies in the In₂O₃ lattice can be responsible for formation of Mn²⁺ doped In₂O₃ NCs in spite of using Mn³⁺ precursor. Similarly, Mn²⁺–Sn⁴⁺ codoped In₂O₃ NCs are also formed with Mn³⁺ precursor. In fact, reduction of Mn³⁺ to Mn²⁺ ions can become even more favorable for the codoped samples, where the electron released by Sn⁴⁺ ion can be used for the reduction, in addition to electrons donated by oxygen vacancies. Though some literature have reported possibilities of a mixture of Mn²⁺ with smaller amount of Mn³⁺ in In₂O₃, a careful survey of literature points out that most prior literature qualitatively agree with our observation that Mn²⁺ is the stable dopant species in spite of the fact that it is not isovalent with In³⁺.^{18,23,25,26,36} The charge neutrality in this case is being maintained by easy creation of oxygen vacancy as shown in ref. 23. The quantitative differences, where some literature finds reasonable quantity of Mn³⁺,^25,26 whereas, very small or negligible amount of Mn³⁺ in other literature^18,23,36 is expected to arise due to different sample preparation procedures which would yield a different local structure around a Mn ion and consequently, can fine-tune the formation energies of different oxidation states.

3.4 Plasmonics

Fig. 6a shows UV-vis-NIR spectra for Mn–Sn codoped In₂O₃ NCs with varying doping levels of Mn and constant 10% Sn doping. 10% Sn doped In₂O₃ NCs exhibits an intense SPR band with peak at ∼2200 nm, similar to prior literature.^19,37,38 With increasing doping percentages of Mn in Mn–Sn codoped NCs, this SPR peak gets shifted to a longer wavelength in the mid IR region with a decrease in intensity, similar to our previous report.¹⁸ This tuning of SPR peak by Mn ions serves as an additional proof that both the dopant ions Mn and Sn are part of the same NC and do not indulge in formation of a secondary impurity phase. According to Drude theory, the SPR wavelength is primarily governed by the electron density in the nanostructure and is also dependent upon the size, shape and dielectric constant of the medium.^39,40 We note here that Mn codoping doesn't result in a colossal change in size or shape of NCs (evident from Fig. 2) and doping in such small percentages is unlikely to affect the dielectric constant of the NCs (the dielectric constant for the surrounding medium, tetrachloroethylene being same for all samples). This means that the variation in SPR has more to do with electron trapping around Mn centres than any other factor. We note here that almost all our Mn ions reside in +2 oxidation state as observed from Q-band EPR. The substitution of In³⁺ by Mn²⁺ leaves a hole in the lattice which can be compensated by the free electrons released by Sn⁴⁺ ions, shifting the SPR peak to longer wavelengths. Furthermore, the decrease in intensity of SPR along with broadening can happen because of scattering of charge carriers by ionized impurities.²¹ Such scattering increases with increase in Mn doping percentage. This can be estimated from half width at half maxima (HWHM) for the SPR band which is a function of scattering by the charge carriers. Expectedly, Fig. 6b shows an increase in HWHM with increasing Mn content in Mn–Sn codoped NCs, suggesting an increase in scattering by Mn centers. This decrease in effective electron concentration and increased scattering of electrons by Mn doping is also evident from electrical resistivity data (Fig. 6c) where Mn doping brings down the electrical conductivity from 22 S cm⁻¹ for 10% Sn doped In₂O₃ NCs to 1.8 S cm⁻¹ for 1% Mn–10% Sn codoped In₂O₃ NCs.

Fig. 6 (a) UV-vis-NIR spectra for x% Mn–10% Sn codoped In₂O₃ NCs showing SPR band shifting to longer wavelengths with increase in Mn codoping percentage. (b) Variation of half width at half maxima (HWHM) with Mn codoping showing enhancement in the electron scattering with ionized Mn²⁺ impurity as Mn codoping levels increase. The circles represent experimental data and solid lines are just guide to the eye. (c) V–I curve for Mn–Sn codoped In₂O₃ NCs suggesting decrease in effective electron concentration on codoping with Mn ions.

We also mention here that, preliminary magnetic measurements were carried out with the example of 10% Mn doped In₂O₃ NCs. Both magnetization vs. magnetic field and temperature dependent magnetization curves are shown in Fig. S2 of ESI.† The results show predominant paramagnetic behavior, similar to ref. 18.

4. Conclusions

We prepared only Mn²⁺ doped and Mn²⁺–Sn⁴⁺ codoped In₂O₃ NCs using Mn³⁺ as one of the precursors. The reduction of Mn³⁺ precursor to Mn²⁺ in the product NC has been confirmed by Q-band EPR spectroscopy. Though Mn³⁺ is isovalent to In³⁺, substitution of In³⁺ in the In₂O₃ lattice with Mn²⁺ is more favorable as compared to Mn³⁺ ion. This aliovalent doping of Mn²⁺ in In₂O₃ lattice is probably caused by the interaction between the acceptor d-levels of Mn³⁺ in the mid-gap region with electron donor levels formed by oxygen vacancies. These oxygen vacancies can also maintain the charge neutrality after substitution of In³⁺ with Mn²⁺ ion. In the case of Mn²⁺–Sn⁴⁺ codoped In₂O₃ NCs, Sn⁴⁺ doping provide additional donor electrons. Also, reaction conditions play an important role in converting Mn³⁺ to Mn²⁺. These Mn–Sn codoped NCs exhibit SPR band in the near to mid IR region. Increase in Mn content shifts the SPR band of codoped NCs toward longer wavelengths along with broadening of the SPR peak. This is because of electron–hole recombination between hole provided by Mn²⁺ doping and free electrons released through Sn⁴⁺ doping, and scattering of electrons by the ionized Mn²⁺ impurities which reduces the effective free electron concentration. Accordingly, electrical conductivity also decreases from 22 S cm⁻¹ for 10% Sn doped In₂O₃ NCs to 1.8 S cm⁻¹ for 1% Mn–10% Sn codoped In₂O₃ NCs. This understanding of Mn²⁺ doping mechanism and its influence on plasmonic data will be useful to design and understand dilute magnetic semiconductors using Mn²⁺ doping in different oxide materials.

Acknowledgements

We acknowledge SAIF, IIT-Bombay for Q-band EPR and ICP-OES measurements and IISER Pune for instrumental facilities. We also thank Dr Janardan Kundu from NCL-Pune for providing access to TEM facility and Dr Sunil Nair for SQUID measurements. A. N. acknowledges Science and Engineering Research Board (SERB) for Ramanujan Fellowship (SR/S2/RJN-61/2012) and DST-Nano Mission grant (SR/NM/NS-1474/2014) Govt. of India. B. T. and A. Y. thank CSIR and INSPIRE respectively for fellowships.

References

1. G. B. González, Mater., 2012, 5, 818–850.
2. G. K. Liang Goh, K. Y. Seng Chan and T. Liu, CrystEngComm, 2011, 13, 524–529.
3. W. Wan, J. Huang, L. Zhu, L. Hu, Z. Wen, L. Sun and Z. Ye, CrystEngComm, 2013, 15, 7887–7894.
4. R. Viswanatha, D. Naveh, J. R. Chelikowsky, L. Kronik and D. D. Sarma, J. Phys. Chem. Lett., 2012, 3, 2009–2014.
5. Q. Sun, Y. Zeng and D. Jiang, CrystEngComm, 2012, 14, 713–718.
6. S. A. Chambers, T. C. Droubay, C. M. Wang, K. M. Rosso, S. M. Heald, D. A. Schwartz, K. R. Kittilstved and D. R. Gamelin, Mater. Today, 2006, 9, 28–35.
7. S. B. Ogale, Adv. Mater., 2010, 22, 3125–3155.
8. J. M. D. Coey, M. Venkatesan and C. B. Fitzgerald, Nat. Mater., 2005, 4, 173–179.
9. J. Philip, A. Punnoose, B. I. Kim, K. M. Reddy, S. Layne, J. O. Holmes, B. Satpati, P. R. Leclair, T. S. Santos and J. S. Moodera, Nat. Mater., 2006, 5, 298–304.
10. H. Raebiger, S. Lany and A. Zunger, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 165202.
11. C. N. R. Rao and F. L. Deepak, J. Mater. Chem., 2005, 15, 573–578.
12. J. Philip, N. Theodoropoulou, G. Berera, J. S. Moodera and B. Satpati, Appl. Phys. Lett., 2004, 85, 777–779.
13. T. Ohno, T. Kawahara, M. Murasugi, H. Tanaka, T. Kawai and S. Kohiki, J. Magn. Magn. Mater., 2007, 310, e717–e719.
14. A. P. Caricato, M. Cesaria, A. Luches, M. Martino, G. Maruccio, D. Valerini, M. Catalano, A. Cola, M. G. Manera, M. Lomascolo, A. Taurino and R. Rella, Appl. Phys. A, 2010, 101, 753–758.
15. M. Cesaria, C. Anna Paola, G. Leggieri, A. Luches, M. Martino, G. Maruccio, M. Catalano, M. Maria Grazia, R. Rella and A. Taurino, J. Phys. D: Appl. Phys., 2011, 44, 365403–365410.
16. K. Nomura, J. Sakuma, T. Ooki and M. Takeda, Hyperfine Interact., 2008, 184, 117–121.
17. A. Hakimi, M. G. Blamire, S. M. Heald, M. S. Alshammari, M. S. Alqahtani, D. S. Score, H. J. Blythe, A. M. Fox and G. A. Gehring, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84, 085201.
18. B. Tandon, A. Yadav and A. Nag, Chem. Mater., 2016, 28, 3620–3624.
19. B. Tandon, G. S. Shanker and A. Nag, J. Phys. Chem. Lett., 2014, 5, 2306–2311.
20. G. S. Shanker, B. Tandon, T. Shibata, S. Chattopadhyay and A. Nag, Chem. Mater., 2015, 27, 892–900.
21. S. D. Lounis, E. L. Runnerstrom, A. Llordés and D. J. Milliron, J. Phys. Chem. Lett., 2014, 5, 1564–1574.
22. R. D. Shannon and C. T. Prewitt, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1969, 25, 925–946.
23. Y. An, S. Wang, L. Duan, J. Liu and Z. Wu, Appl. Phys. Lett., 2013, 102, 212411.
24. T. Okazaki, T. Yoshioka, Y. Kusakabe, T. Yamamoto and A. Kitada, Solid State Commun., 2011, 151, 1749–1752.
25. S. S. Farvid, N. Dave, T. Wang and P. V. Radovanovic, J. Phys. Chem. C, 2009, 113, 15928–15933.
26. S. S. Farvid, T. Sabergharesou, L. N. Hutfluss, M. Hegde, E. Prouzet and P. V. Radovanovic, J. Am. Chem. Soc., 2014, 136, 7669–7679.
27. J. Zimmerman, Biochem. Mol. Biol. Educ., 2005, 33, 382.
28. A. L. Patterson, Phys. Rev., 1939, 56, 978–982.
29. D. M. Murphy, in Metal Oxide Catalysis, Wiley-VCH Verlag GmbH & Co. KGaA, 2009, pp. 1–50.
30. B. Bleaney, Discuss. Faraday Soc., 1955, 19, 112–118.
31. O. Matumura, J. Phys. Soc. Jpn., 1959, 14, 108.
32. N. S. Norberg, K. R. Kittilstved, J. E. Amonette, R. K. Kukkadapu, D. A. Schwartz and D. R. Gamelin, J. Am. Chem. Soc., 2004, 126, 9387–9398.
33. A. R. Vazquez-Olmos, J. I. Gomez-Peralta, R. Y. Sato-Berru and A. L. Fernandez-Osorio, J. Alloys Compd., 2014, 615, S522–S525.
34. A. D. Liehr, in Progress in Inorganic Chemistry, John Wiley & Sons, Inc., 2007, pp. 281–314f.
35. S. Lany and A. Zunger, Phys. Rev. Lett., 2007, 98, 045501.
36. G. Peleckis, X. L. Wang and S. X. Dou, J. Magn. Magn. Mater., 2006, 301, 308–311.
37. T. Wang and P. V. Radovanovic, J. Phys. Chem. C, 2011, 115, 406–413.
38. A. M. Schimpf, S. D. Lounis, E. L. Runnerstrom, D. J. Milliron and D. R. Gamelin, J. Am. Chem. Soc., 2015, 137, 518–524.
39. J. Pérez-Juste, I. Pastoriza-Santos, L. M. Liz-Marzán and P. Mulvaney, Coord. Chem. Rev., 2005, 249, 1870–1901.
40. M. M. Alvarez, J. T. Khoury, T. G. Schaaff, M. N. Shafigullin, I. Vezmar and R. L. Whetten, J. Phys. Chem. B, 1997, 101, 3706–3712.

---

Footnote

† Electronic supplementary information (ESI) available: EDX, ICP-OES, and magnetic data. See DOI: 10.1039/c6ra16676h

---