Doping Independent Work Function and Stable Band Gap of Spinel Ferrites with Tunable Plasmonic and Magnetic Properties

: Tuning optical or magnetic properties of nano-particles, by addition of impurities, for speci ﬁ c applications is usually achieved at the cost of band gap and work function reduction. Additionally, conventional strategies to develop nano-particles with a large band gap also encounter problems of phase separation and poor crystallinity at high alloying degree. Addressing the aforementioned trade-o ﬀ s, here we report Ni − Zn nanoferrites with energy band gap ( E g ) of ≈ 3.20 eV and a work function of ≈ 5.88 eV. While changes in the magnetoplasmonic properties of the Ni − Zn ferrite were successfully achieved with the incorporation of bismuth ions at di ﬀ erent concentrations, there was no alteration of the band gap and work function in the developed Ni − Zn ferrite. This suggests that with the addition of minute impurities to ferrites, independent of their changes in the band gap and work function, one can tune their magnetic and optical properties, which is desired in a wide range of applications such as nanobiosensing, nanoparticle based catalysis, and renewable energy generation using nanotechnology.

W ide band gap semiconductor nanoparticles, with energy band gaps (E g ), enhance the performance of nanoscale devices in a wide range of applications ranging from deep-UV optoelectronics, 1 quantum computing, 2 biosensing, 3 highpower electronics, 4 to the use of nanoparticles in extreme environment applications. Engineering atoms to form nanoscale alloys is one of the main routes for achieving E g modulation in nanoparticles. 5,6 For instance, several nanoalloys such as ZnO, 7 ZnGa 2 O 4 , 8 Ga 2 O 3 /SnO 2 , 9 CuO, 10 GaOOH, 11 and many others 12 have been synthesized recently with unique electrical and optical properties. Among these materials the spinel compounds (such as ZnGa 2 O 4 ) possess structural features which are suitable for tuning the optoelectronic properties of the material at nanoscale. 13 This is due to the inherent nature of the spinel configuration which allows a deserved choice of cation coordination within the structure of these materials. 14,15 This flexibility in the choice of cation also allows tuning of the dielectric and magnetic (in the case of spinel ferrites) properties of the spinel materials. 16−19 More specifically, the spinel ferrites have a general structural formula AB 2 O 4 where A and B are divalent and trivalent cations, respectively. 20 Depending on the choice of cations and external conditions such as temperature or pressure, the distribution of the cations within the tetrahedral and octahedral positions in the structure can be changed. 21,22 This distribution of the cations is further defined by the degree of inversion (x) which is represented as ( 4 where the value of x varies from 0 to 1. Note that the material structure for x = 0 is called a normal spinel, for x = 1, it is an inverse spinel, while for x = 2/3, the structure is known as a random spinel structure. 23 We can consider this process of changing the degree of inversion as the nanoparticle doping process. To change the degree of inversion, various physical and chemical techniques such as sol−gel, combustion, hydrothermal, and coprecipitation methods can be utilized to achieve doped nanoparticles with uncommon optoelectronic properties. 24 For example, increase in conductivity, enhancement in the magnetic moment, and nuclear magnetic properties, including hyperfine field distributions, isomer shifts, and quadrupole interactions, 25 can be achieved by doping the nanoparticle. Most often such changes are necessary for fine-tuning the application of the developed nanoparticle for a specific purpose. For instance within application of semiconductor nanoparticles (such as our spinel nanoparticles) associated with the photocatalysis, it is of utmost importance that the free electron hole pairs are generated. 26,27 Therefore, the necessary condition for it to be a photocatalyst is the low recombination rate which can be easily tuned by adding dopant or changing its concentration through modulation in the degree of inversion. Similarly, for nanoplasmonic biosensing, such as surface enhanced Raman spectroscopy (SERS) or localized surface plasmon resonance (LSPR), application of the degree in inversion can lead to enhanced sensitivity, surface stability, and reproducibility in multimetallic or semiconductor like LSPR platforms. 28−30 In this context, we have developed bismuth (Bi) doped nickel zinc (Ni−Zn) spinel ferrite nanopartciles, Ni 0.5 Zn 0.5 Bi x Fe 2−x O 4 , where x is the degree of inversion, varying from 0.00 to 0.1 in steps of 0.02. We characterized the developed material with physical, optical, magnetic, and atomic resolution tools to understand the features of the developed material. We find that the optical band gap and work function of the material do not change with the degree of inversion while the magnetic and plasmonic properties' (optical properties in UV−vis) regions change with changes in the degree of inversion. This suggests that the optomagnetic properties of the developed material are tunable without including the effects of dopant which also generally changes the optical band gap of the material. Achieving the optical gap (E g ) is one of the fundamental issues in achieving controllable and purposeful optoelectronic properties in the wide wavelength range of semiconductors. 31 Therefore, the results discussed here serve as a guide in the development and discovery of materials with unique properties. The Ni 0.5 Zn 0.5 Bi x Fe 2−x O 4 nanoparticles were synthesized by the citrate precursor method where aqueous solutions of zinc, nickel, bismuth, and iron salts were dissolved with citric acid to create a homogneous solution. This solution was then sintered at low temperatures to yield nanoparticles in the powered form. The schematic in Figure 1a shows the process of synthesis, and detailed steps are discussed in the methodology section. For simplicity, hereafter we will use x to refer to the relative concentrations of the precursor used in the synthesis. It should be noted that the process of nanoparticle preparation is simple, is cost-effective, and can generate particles in large scale within a short duration of time. The amount of particle generation is only limited by the size of the instrumentation involvedsuch as the size of the furnace, hot plate, magnetic stirrer, and beakers. After development of the nanoparticles, we carried out X-ray photoelectron spectroscopy (XPS); Figure 1b shows the low-resolution full scan XPS for x = 0.06, while XPS for all other dopant concentrations is shared in the Supporting Information. Essentially, the peaks correspond to the binding energies (BEs) of Bi 4f, Ni 2p, Zn 2p, Fe 2p, and O 1s. 32−35 The BEs of Bi 4f, Ni 2p, Zn 2p, and Fe 2p are ascribed to the BEs of the Bi 3+ , Ni 2+ , Zn 2+ , and Fe 3+ ionic states in the samples. As observed in Figure 1b, the bismuth ion is found to have an oxidation state of +3 as the core level spectra for Bi 4f are in the ranges 158−160 eV and 164−166 eV for the 7/2 and 5/2 spin orbit doublet components, respectively. In the observation of the element iron, the Fe 2p states of the Fe atom could be used to distinguish the variation between Fe 2+ and Fe 3+ . The pure Fe 2p narrow scan (see the Supporting Information) shows two wide doublet peaks with spin−orbit splitting energies of 3.88 and 4.43 eV and positioned in the ranges 710−713 eV and 722−727 eV for Fe 2p 3/2 and Fe 2p 1/2 , respectively. The peak with the range 710−713 eV is attributed to the Fe 3+ cation located at the octahedral site in the spinel structure. The peak with the range 722−727 eV is due to the presence of Fe 2+ cation located at the tetrahedral site in the spinel structure. In addition, we observe the nickel element in the ranges 854−856 eV and 863−874 eV, which correspond to the Ni 2p 3/2 and Ni 2p 1/2 electrons in the Ni 2+ oxidation state, respectively. The presence of the element zinc is also confirmed by observation of the peak energies in the ranges 1020−1023 eV and 1044−1049 eV, which correspond to Zn 2p 3/2 and Zn 2p 1/2 electrons in the Zn 2+ oxidation state, respectively. Note that the element bismuth also exists in the pentavalent state (Bi 5+ ) which can replace the few Fe 2+ ions in the spinel nanoferrite structure that can further influence the magnetic and electrical properties of the nanoparticles. Furthermore, the spinel crystalline structure (AB 2 O 4 ) with tetrahedral and octahedral sites for metal atoms is also confirmed by correlating the identified elements in the XPS with X-ray diffraction (XRD). Using XRD we also extract the lattice parameter, X-ray density, bulk material density, porosity, and crystalline size of the nanoparticles; see Figure 1c. We observe a decrease in the lattice parameter with addition of the dopant. In addition, the lattice parameter, bulk density, and Xray density increase with the addition of Bi (at x = 0.02) and then decrease with further increase in the Bi concentration from x = 0.02 to 0.10. See the Supporting Information for the detailed XPS high-resolution scan and XRD of all nanoparticles. The changes in the aforementioned atomic properties identified using XRD are attributed to the changes in the phase of the material. 36  Figure 2 shows the transmission electron microscopy (TEM) images for all six nanoparticles (S1 to S6). As shown in the images, the particles have dimensions in the nanorange. The particle size distribution extracted from the TEM characterization will be discussed later in Figure 3a. Figure 2 also shows the selected area electron diffraction patterns (SAED) for the samples. The d-spacing calculated from the SAED patterns matched well with the XRD patterns (see the Supporting Information for the XRD related data analysis). The SAED patterns with plane labels (marked in yellow) have been added in the figures. From SAED analysis, the spacings for all six samples are very similar and show no measurable differences. Within these high-resolution images, the lattice planes can also be observed. The calculated interplanar spacing was utilized for determination of the plane type, and based on the analysis of multiple HRTEM images from each of the samples (here only one HRTEM image is reported from each sample), a variety of planes are observed. For reporting purposes, HRTEM images with various identified planes are shown here. For instance, the interplanar spacing of 0.30 nm is measured, which corresponds to the 220 plane of the spinel cubic crystal in sample S1. For sample S2, lattice spacings of 0.26 nm (corresponding to the 311 plane) and 0.22 nm (corresponding to the 400 plane) of the spinel cubic crystal are observed. The interplanar angle of 72 degrees further confirms the identified planes. Similarly, for samples S3−6 the lattice spacings corresponding to cubic spinel planes are identified as 311 and 220 (marked in yellow within the images). Additionally, the fast Fourier transform (FFT) corresponding to the planar area indicates sharp patterns asserting the highly crystalline nature of the prepared nanoparticles.
The work function of the nanoparticles was calculated using ultraviolet photoelectron spectroscopy (UPS). UPS allows us to investigate materials within 3.10−124 eV energy, which is ideal for the investigation of valence and conduction band electrons. Empirically the work function (φ) can be calculated from the UPS measurements according to the relation φ = hν − E cutoff , where hν is the excitation energy source (He I 21.22 eV) and E cutoff is the energy cutoff measured from the UPS spectra (Figure 3b). UPS measurements produced the value ≈5.88 eV for all our samples. From Figure 3b, the energy cutoff is measured as 15.34 eV (W) for all six samples. Therefore, the difference between excitation energy sources was measured to be 5.88 eV. It should be noted that despite the change in concentration of the dopant from 0.00 to 0.10 and differences in morphology (different sizes of the nanoparticles as seen in Figure 3a) this value does not change, validating our observation that the developed nanoparticles have a work function which is independent of the dopant concentration. This is in contrast with the reports in the literature where adding metallic dopant either increases 37 or decreases 38 the work function of a given nanomaterial. To further validate the independence of the band structure of the material on the dopant concentration, we calculate the band gap of the developed nanoparticles. To determine the band gap of the samples, we performed UV−vis spectroscopy of all samples. The absorbance spectra of all six samples are found to be identical, and minute differences in the absorption, which are not easy to distinguish, are observed. Therefore, we only show samples S1 and S6, between which the absorbances of the S2−S5 samples exist, in Figure 3c−e. To calculate the band gaps of the samples, the experimental absorbance spectra were reduced to the form (a*E) 2 = f(E), where the frequency dependence of the absorption coefficient in the vicinity of the absorption edge is described by a = [C(E − E g ) 1/2 ]/E. In this equation, a is the absorption coefficient (here we use aborbance A, proportional to the absorption coefficient), E is the photon energy, E g is the optical band gap, and C is a constant. 39 The band gap values were determined by extrapolating the straight sections of the obtained curves in the short-wavelength region of the spectrum. From the analysis, we observe that the band gap of the sample containing no Bi dopant is 3.20 eV, and by increasing the Bi concentration, the band gap increases to a maximum of ≈3.22 eV (in sample S6). As the observed change in the band gap is less than 0.5%, we can consider it to be relatively independent of the concentration of the Bi dopant. A general band diagram of the nanoparticles is shown in Figure  3f where the band gap is shown as 3.20 eV and the work function is 5.88 eV. Figure 4 a shows the magnetic and optical characterization of the developed spinel nanoparticles. Prior to the discussion in Figure 4, we would like to mention that using the unpolarized light, we observe a wide absorbance region in the UV−vis spectrum; see Figure 3c−e. In addition, the kinetic energy measured in UPS, Figure 3b, can also be attributed to the molecular electronic absorptions within 2−20 eVprimarily due to the plasmon excitation, both bulk and surface plasmon. The bulk plasmon consists of longitudinal waves and in the To confirm that these plasmonic effects were contributed by the Bi dopant, we study the inner shell ionization in the reflection mode using reflection electron energy loss spectroscopy (REELS). REELS provides information on the electron transport phenomena in the surface of a solid, where incident electrons are elastically scattered through an interaction with the material's valence and core electrons, providing direct insight into the elemental information. The REELS spectra, in Figure 4a, were measured at TOAs (take off angles) of 90°for different inversion values of the spinel nanoparticle. The spectra are magnified 2 times relative to the no-loss peaks. The analyzer transfer function, which is proportional to the kinetic energy of the electrons, has also been removed, and therefore the lines represent the experimental spectra. Analyzing these peaks, the REELS has generated complementary results to XPS in Figure 1b, where the peak at 946.63 eV can be correlated with the precursor concentration, the x = 0.00−0.10 value of bismuth. These peaks are commonly observed as satellite peaks for atoms in ferrite based alloys. 40 To study the effect of polarized light, we investigate differential absorption of left and right circular polarized light using magnetic circular dichroism (MCD); see Figure 4b. Here we observe that the shapes of the MCD spectra for all the samples are similar (less than 10% variation in the peak locations), except for sample 2 (x = 0.02), which has a large deviation in its MCD properties. This is an anomaly which we attribute to the fact that the size of the particles in the x = 0.2 sample was much larger than the other doped nanoparticles, as seen from the TEM characterization. Moreover, a quantitative comparison of the signal intensity for all samples is presumptive because it is nontrivial to accurately determine the total thickness of the nanoparticles along the path of the light beam through the sample. However, the constancy in the shape of the MCD spectrum indicates that the substitution of Bi for Fe ions in the concentrations used does not affect the electronic structure of Ni−Zn-ferrite nanocrystals. Therefore, here we only analyze sample 3 (S3) in detail to represent and discuss the features of the developed nanoparticles. The obtained MCD for S3 is shown in Figure 4c and d. The MCD spectrum shape is similar to that of the maghemite (γ-Fe 2 O 3 ) thin film in our previous work. 41 This is not surprising since both compounds contain Fe 3+ ions of the 3d 5 ion configuration in the octahedral and tetrahedral positions. The experimental spectrum is also decomposed into individual components of Gaussian shape in Figure 4d. The best fit of the calculated spectrum to the experimental one is obtained taking into account four components with energies E1−E4 = 2.023, 2.288, 2.590, and 2.839 eV, respectively. By analogy with maghemite, the strongest peaks E2 and E3 can be related to the electronic transitions from the ground states of the Fe 3+ ion 6 A 1g ( 6 S g ) in the octahedral and 6 A 1 ( 6 S) in the tetrahedral sites to the excited states: 4 A 1g , 4 E g ( 4 G) and 4 A 1 , 4 E( 4 G). Since the characterization is performed in the UV−vis range, the absorbance properties of the material toward polarized light are attributed to the surface and bulk plasmons as indicated earlier in our discussion.
Magnetic properties are reported here in Figure 4e−i. The saturation magnetization (M s ) and coercivity (H c ) depend on both Bi concentration and nanoparticle average size; see Figure  4e for the dopant concentration, which affects the particle as seen from Figure 3a. Indeed, in samples 3−5, the nanoparticle size is approximately the same or less as compared to samples Nano Letters pubs.acs.org/NanoLett Letter 1, 2, and 6. In the developed nanoparticle series (3−5), a decrease in the H c is observed with a linear trend that changes with increase in the Bi concentration. The minimal H c value (14 Oe) observed in our sample 5 is rather attractive for a wide range of high-frequency applications. Detailed curves for extraction of these M s and H c are shared in the Supporting Information for all 6 samples. We also investigate the effect of temperature on M s and H c ; see Figure 4f−i, which indicates a linear drop in the values when temperature is varied from 77 K to 445 K. We found a linear relationship between magnetization and coercivity; see Figure 5 a. While the relationship between magnetization and coercivity might be linear, it should not be generalized for all nanomaterials as we have only found this linear relationship for our developed ferrite. Figure 5 b correlates the nanoparticle properties with the dopant concentration. Note that within Figure 5b, the data sets used for ploting the curves are normalized with respect to the largest data value in that particular data set. This leads to a dimensionless number which is represented as a normalized value with arbitrary units in the y-axis, further elucidating the comparison between the various properties of the materials with a change in the degree of the inversion. Clearly, Figure 5b shows that the optical band gap and work function do not vary with the change in the dopant concentration (less than 1% changes observed). Variations up to 17% are observed in the coercivity, and more than 70% variations are in the magnetization, plasmonic strength, and particle size when the dopant concentration is changed. Note that these variations are the maximum changes of the values with a given data set, i.e. the difference between the highest and lowest values, which are calculated as a percentage change with respect to the lowest value in the data set. We further find the correlation between the normalized values that change within the given data set, and these are represented in the form of a heat map in Figure  5c. This heat map again suggests that there are insignificant changes within the band gap and work function with a change in the dopant concentration. The fabrication process is optimized for uniform, homogeneous, and stoichiometric nanoferrites. The different parameters such as quantity of distilled water used, citric acid as fuel, presintering, sintering temperature, and heating and cooling rates were all optimized for the reproducible results discussed here.
In summary, we have proposed a state-of-the-art design concept for the spinel nanoferrites which have tunable magnetic and plasmonic properties but stable optical band gap and work function. These ferrites were made up of Ni and Zn elements, and to achieve the tunable optomagnetic properties, Bi was inserted within the compound. To reveal the aforementioned unique properties, we performed several physiochemical, magnetic, and optical characterizations including REELS and MCD performed for the first time in Ni− Zn spinel ferrites. This work would open up a new avenue to develop wide band gap nanomaterials possessing specifically tailored optomagnetic properties, which have been facing the challenge of band gap reduction with increase in property enhancing dopants.
Detailed information on XPS, XRD, vibrating sample magnetometer (VSM) measurements, MCD, and methods used for nanoparticle preparation and characterization (PDF) Incubation grant from the Ministry of Electronics and Information Technology (MeitY) under Technology Incubation and Development of Entrepreneurs (TIDE 2.0) program and the startup nanoLatticeX.