Materials that exhibit plasmonic response in the UV region can be advantageous for many applications, such as biological photodegradation, photocatalysis, disinfection, and bioimaging. Transition metal nitrides have recently emerged as chemically and thermally stable alternatives to metal-based plasmonic materials. However, most free-standing nitride nanostructures explored so far have plasmonic responses in the visible and near-IR regions. Herein, we report the synthesis of UV-plasmonic Cr2N nanoparticles using a solid-state nitridation reaction. The nanoparticles had an average diameter of 9 ± 5 nm and a positively charged surface that yields stable colloidal suspension. The particles were composed of a crystalline nitride core and an amorphous oxide/oxynitride shell whose thickness varied between 1 and 7 nm. Calculations performed using the finite element method predicted the localized surface plasmon resonance (LSPR) for these nanoparticles to be in the UV-C region (100–280 nm). While a distinctive LSPR peak could not be observed using absorbance measurements, low-loss electron energy loss spectroscopy showed the presence of surface plasmons between 80 and 250 nm (or ∼5 to 15 eV) and bulk plasmons centered around 50–62 nm (or ∼20 to 25 eV). Plasmonic coupling was also observed between the nanoparticles, resulting in resonances between 250 and 400 nm (or ∼2.5 to 5 eV).
I. INTRODUCTION
Plasmonic metal nitrides have recently received significant interest as reports outlining multiple synthetic methods to prepare free-standing nanoparticles (NPs),1–3 demonstrating their thermal stability4 and high photothermal efficiency,4,5 have emerged. Many of these studies have predominantly focused on TiN, which exhibits broad absorption with a localized surface plasmon resonance (LSPR) maximum, typically in the near-IR region.6 As such, TiN has been explored for solar light-driven processes, such as water evaporation4,5 and photocatalysis,7–9 and applications that require absorption in the biological transparency window, such as photothermal therapy10 and photoacoustic tomography.11 Moving down the group, ZrN and HfN have also been explored, which possess LSPR in the visible region of the electromagnetic spectrum.12,13 They have been shown to have higher photothermal efficiencies compared to TiN due to stronger electron–phonon coupling.14–16 Detailed experimental studies on plasmonic metal nitride NPs beyond those containing group 4 transition metals are scarce. Therefore, most of the plasmonic applications involving nitrides have predominantly focused on light–material interactions in the visible and near-IR region.
Materials with plasmonic responses in the UV region (100–400 nm) are of interest for disinfection, biological imaging, sensing, and developing metamaterials.17 Metals such as Al, Mg, Ga, In, Rh, Pd, Pt, Sn, Tl, Sb, Pb, Bi, and Ru have a bulk plasma frequency below 400 nm and have been explored for UV applications.18–24 However, these metals can be either expensive, oxidize under ambient conditions, or have low melting points, making them incompatible with high-temperature fabrication techniques.22,23 Computational studies have predicted that chromium metal nitrides (CrN and Cr2N) can possess LSPR in the blue-to-UV region of the electromagnetic spectrum.25,26 The high melting points of these nitrides (>1500 °C) make them an attractive alternative to metals such as Al and Mg. However, synthetic methods to prepare these nitride NPs and a better understanding of their plasmonic behavior are required to truly assess their applicability.
Recently, plasmonic Cr2N NPs with a size ranging between 0.8 and 30 nm were synthesized using a single step pulse laser irradiation.26 The absorption spectra of these particles showed multiple LSPR between 320 and 420 nm with a maximum at 372 nm. This broad absorption was hypothesized to be due to hotspot formation from particle–particle interaction. While the resonances originating from interparticle coupling have been observed, the bulk and surface plasmon resonances of individual Cr2N NPs have not been characterized experimentally. This study focuses on exploring the plasmonic properties of Cr2N NPs using UV–vis absorbance spectroscopy, electron energy loss spectroscopy (EELS), and numerical methods. The NPs were prepared using a solid-state reaction between Cr2O3 and Mg3N2, which resulted in water-dispersible NPs and were characterized using various analytical and microscopy techniques. The NPs had bulk and surface plasmon resonances below 200 nm (6.0 eV), which is corroborated by the calculations performed using the finite element method.
II. EXPERIMENTAL METHODS
A. Materials
Chromium oxide (Cr2O3, 99.9%, 18 nm), and magnesium nitride (Mg3N2, 99%, −325 mesh) were purchased from U.S. Research Nanomaterials Inc., and hydrochloric acid (HCl) was purchased from Anachemica (37%). Sodium hydroxide (NaOH) was purchased from Alfa Aesar (98%). Deionized water (DI-water, 18.2 MΩ) was obtained from a Sartorius Arium water purification system. All glassware was cleaned thoroughly with DI-water and then with acetone, and placed in an oven to dry prior to use.
B. Synthesis of Cr2N NPs
The synthesis procedure to make Cr2N particles was adapted from a previously reported method.12 Briefly, Cr2O3 (0.152 g, 1.0 mmol) nanopowder was mixed with Mg3N2 (0.403 g, 4.0 mmol) using a spatula. This step was performed in a nitrogen-filled glovebox to avoid Mg3N2 decomposition in the presence of moisture. The powder mixture was transferred to a CoorsTM high alumina combustion boat, and the boat was immediately placed in a quartz tube. The reaction mixture was then heated to 1000 °C at a ramp rate of 10 °C/min and held at that temperature for 3 h under argon flow in a Lindberg Blue M™ furnace. The reaction mixture was cooled to room temperature, and the resulting product was transferred to a 100 ml beaker. Then, 25 ml of 1.0M aqueous HCl solution was slowly added to the reaction product and stirred together for 1 h to remove magnesium oxide (MgO) and any unreacted Mg3N2. The reaction solution was centrifuged for 15 min at 3300 rpm and the supernatant containing acid and soluble by-products was discarded (after the first centrifugation step, the supernatant is colorless). The solid was resuspended in 5.0 ml of DI-water by sonication and centrifuged for 15 min at 3300 rpm. The colored supernatant containing Cr2N NPs was collected, and the precipitate was resuspended in 5 ml DI-water by sonication and centrifuged again for 15 min at 3300 rpm. The colored supernatant was collected and mixed with the supernatant from the previous step. To collect the powder XRD pattern of the dispersed NPs, the solution pH was adjusted to ∼6 using NaOH solution (1.0M), and the particles were allowed to settle overnight and centrifuged to isolate the solid.
C. Characterization techniques
Powder X-ray diffraction (XRD) patterns were collected using a Rigaku Ultima IV X-Ray diffractometer with Cu Kα radiation (λ = 1.54 Å). The samples were placed on a zero-background Si wafer and the spectra were collected at 3 counts/s. X-ray photoelectron spectroscopy (XPS) measurements were performed on a PHI Versaprobe instrument with an Al Kα source (1486.6 eV). Powders of each sample were dispersed in water and then dried as a near-coverage layer onto a diced Si wafer at 70 °C. CasaXPS software (VAMAS) was used to interpret high-resolution (HR) XP spectra. The background was subtracted using a Shirley-type background to remove most of the extrinsic loss structure. Absorption spectra were collected using an Agilent CARY 5000 spectrophotometer with a spectral range of 190–3300 nm and a measurement step of 1 nm. NP suspensions were placed in a quartz cuvette, and DI-water was used as the blank. Dynamic light scattering (DLS) measurements were conducted using a Zetasizer Nano Series Nano-ZS (Malvern Panalytical) instrument with a built-in 623 nm laser. For the size measurements, dilute NP solutions were placed in a plastic cuvette, whereas for the zeta potential measurements, a DTS1070 folded capillary cell was used. Zetasizer software was used for data analysis. Transmission electron microscopy (TEM) samples were prepared by drop-casting nanoparticle suspensions onto a copper grid coated with holey carbon film, and the images were collected on either a Thermo Scientific Talos 200 X microscope or a Hitachi-9500 electron microscope at an operating voltage of 200 kV. The selected area electron diffractograms (SAED) were collected on the Hitachi-9500 electron microscope. The core-loss electron energy loss spectroscopy (EELS) data were acquired in Digital Micrograph environment using a Gatan CMOS detector. Image J was used to measure particle size and aspect ratios.27 High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images were collected on a Thermo Scientific Talos 200 X microscope. Low-loss EELS data were acquired on a double-corrected FEI Titan scanning transmission electron microscope operated at 80 kV and equipped with a monochromator and Gatan Quantum GIF spectrometer. The beam was monochromated to achieve an energy resolution of ∼100 meV, using a dispersion of 25 meV/channel (bulk plasmon spectra) or 10 meV/channel (surface plasmon spectra) on a CCD camera with a convergence angle of 14.2 mrad and a collection angle of 25.4 mrad. Data processing on the low-loss EELS was done using custom Python software.28
D. Computational studies
The optical properties of Cr2N NPs were computed using a finite element method (FEM) solver for Maxwell’s equations in COMSOL MultiPhysics. To simulate the optical characteristics of the materials, Lorentz oscillator parameters of Cr2N were obtained from Aouadi et al.29 to determine the real and imaginary parts of the dielectric function of the material. In simulations, particles were modeled as three-dimensional nanospheres suspended in water. To ensure the accuracy of the results, rigorous convergence analysis and perfectly matched layers (PML) boundary conditions. Three-dimensional models were meshed via the built-in meshing algorithm in COMSOL with a maximum element size of 0.1r, where r is the radius of nanospheres. The extinction cross sections of Cr2N NPs were calculated as the sum of absorption and scattering cross sections at a given NP size.30 The extinction cross section values of TiN, ZrN, HfN, and Au were obtained from calculations performed by Monfared et al.31
III. RESULTS AND DISCUSSION
The solid-state nitridation of metal oxide nanopowders with Mg3N2 has been previously used by our group to successfully synthesize free-standing group 4 plasmonic transition metal nitride nanostructures.12 A similar procedure was utilized here to synthesize Cr2N NPs. Cr2O3 nanopowder with an average particle size of 45 ± 10 nm and an ellipsoidal shape (Fig. S1) was commercially purchased. The powder XRD pattern of the oxide precursor showed reflections characteristic of cubic phase Cr2O3 [Fig. 1(a)].32 This was reacted with Mg3N2 at 1000 °C for 3 h and the resulting reaction product was treated with a 1M aqueous HCl solution to remove MgO and any unreacted Mg3N2. The nitride product was isolated through a series of washes as outlined in the experimental section. The powder XRD pattern of the dried NP dispersion showed reflections characteristic of Cr2N with a hexagonal crystal structure,33 and no peaks corresponding to the starting materials or reaction by-products were observed. The TEM analysis showed most of the NPs to be pseudospherical with an average diameter of 9 ± 5 nm [Figs. 1(b) and 1(c)]. While most particles were between 5 and 15 nm, a small portion of larger NPs ranging between 15 and 25 nm were also formed (Fig. S2). The SAED pattern showed the formation of single-crystalline Cr2N NPs (Fig. S3). Additionally, the chemical composition of the Cr2N was investigated using EELS elemental mapping. The EELS analysis showed good overlap between Cr [Fig. 1(e)] and N [Fig. 1(f)], confirming the presence of a chromium nitride phase, whereas O [Fig. 1(g)] was found to be predominantly in N deficient regions. The nitride NPs prepared using this method have been shown to oxidize after the HCl workup to form surface oxynitride and oxide species, which is likely the origin of the O species in the EELS map.12
Zeta potential measurements were conducted on the Cr2N NP colloidal suspensions to determine surface charge and colloidal stability. The Cr2N NPs dispersed in water exhibited a zeta potential of 45.4 ± 0.7 mV, indicating that the NP surface is positively charged. The positive charge most likely originates from the protonated hydroxyl functional groups (–OH2+) groups on the surface that result from the acid workup.34,35 Zeta potential values higher than 30 mV (or below −30 mV) also indicate the formation of a stable colloidal suspension.36,37 The NP suspension had a hydrodynamic diameter of 47.0 ± 0.5 nm as determined by the DLS. To further probe the surface chemical state of the Cr2N, the NPs were analyzed using XPS technique. The high-resolution XP spectrum of the Cr 2p region [Fig. 2(a)] for the precursor had 2p3/2 peak fits at binding energies of 576.5, 578.9, and 580.6 eV corresponding to Cr2O3, Cr(OH)3, and CrO3, respectively.38,39 The combined 2p1/2 fit and the satellite peak maxima were at 586.7 and 597.5 eV, respectively.40 No distinctive peak was observed in the N 1 s region in the starting material [Fig. 2(b)]. The Cr 2p spectrum of Cr2N NPs had 2p3/2 peak fits at binding energies of 573.8, 575.7, and 576.8 eV corresponding to Cr2N, CrOxNy, and Cr2O3, respectively [Fig. 2(c)].41,42 The chemical identity of the small peak located at 579.7 eV is currently unclear but could be due to Cr(V) species.43 The data indicate the presence of an oxide and oxynitride shell around the Cr2N NP core, similar to previous reports.12,44 This is further confirmed using TEM analysis, which showed the presence of an oxide shell around the NPs (Fig. S4). The thickness of this shell varied between 1 and 7 nm for various NPs and was found to be non-uniform even within the same particle. The N 1s spectrum of Cr2N NPs showed a weak signal at ∼395.5 eV corresponding to the nitride [Fig. 2(d)].41 Given the oxide shell thickness, it is likely that not many electrons from the Cr2N core are detected.
The real (εr) and imaginary (εi) components of the dielectric function of Cr2N between 100 and 700 nm are shown in Fig. 3(a). A negative real permittivity is observed in this wavelength range, indicating free electron behavior.25 The magnitude of the imaginary component of the dielectric function increases with decreasing frequency, indicating higher optical losses at longer wavelengths.45 The absorption spectrum of Cr2N NPs with a diameter of 9 and 25 nm was calculated using the finite element method [Fig. 3(b)]. The NP size chosen for the calculations is equal to the average and large particle size in the sample. The data suggest that the LSPR maximum of Cr2N NPs is in the deep UV region [around 60 nm (20.66 eV) for 9 nm NPs and at 145 nm (8.55 eV) for 25 nm particle size] and the LSPR peak of the experimental sample should be relatively broad covering all of the UV and parts of the visible spectrum. The Cr2N NP solution has a brown color [Fig. 3(b) inset] and shows a broad absorbance [Fig. 3(b), solid line] when suspended in DI-water. The characteristic LSPR peak was not observed in the measured spectrum; however, the calculations show this to be present below the spectrometer’s wavelength cutoff (190 nm). The absorbance spectrum showed a shoulder peak at ∼310 nm (4.0 eV), which could be due to plasmon coupling between NPs as observed by Gubert et al.26 The computational analysis on the absorption and scattering of Cr2N NPs was performed to evaluate the extinction behavior of the NPs of different sizes. The normalized absorption and scattering coefficients (Qabs and Qsca) are defined by the calculated absorption or scattering cross sections of NPs (Cabs and Csca) using FEM normalized to the geometrical cross section of the NPs. The extinction coefficient of the NPs can be then defined as Qext = Qabs + Qsca. As shown in Fig. 3(c), for smaller NPs (≤50 nm) the absorbance is higher in the UV-C region (∼100 to 280 nm or 12.4–4.43 eV) whereas for the larger NPs (70–100 nm), the absorbance is higher in the UV-B and UV-A regions (280–400 nm or 4.43–3.10 eV). It is also clear that the absorption and scattering maxima of particles shift to longer wavelengths and become broader with an increase in NP size [Figs. 3(c) and 3(d)]. To understand the relative magnitude of absorption and scattering in Cr2N NPs, the extinction cross section of Cr2N was compared to Au, TiN, ZrN, and HfN at their respective LSPR maxima for three different particle sizes (10, 20, and 30 nm) as shown in Fig. S5. The extinction represents the collective effect of absorption and scattering of incident light by the NPs. For particles ≤20 nm, the extinction cross section of Cr2N was larger than Au, TiN, ZrN, and HfN at their plasmon resonance maxima, whereas for 30 nm particles, the extinction cross section of Cr2N was lower than ZrN and HfN but larger than Au and TiN.
While a distinctive LSPR could not be observed in the absorbance spectrum limited by the instrument wavelength cutoff, EELS was used as an alternative technique to map both bulk and surface plasmon intensity and energy with high spatial resolution. Since the EELS experiments are conducted under a high vacuum, some variations can be expected compared to the NPs suspended in a solvent due to differences in refractive indices of the surrounding media.46 Further, the EELS technique probes the near-field response, whereas the optical excitation explores the far-field.47 It has been previously reported that the nearfield properties appear at a lower energy compared to far-field phenomena.48,49 Figure 4(a) inset shows regions from which the EELS spectra were acquired. All the spectra obtained in these studies were normalized to the zero-loss peak. The large signal with peak centered around 20–25 eV (i.e., ∼50 to 62 nm) is indicative of the bulk plasmon [Fig. 4(a)]. The shoulder peaks observed in spots 3–5 spectra are indicative of surface plasmons; however, these are difficult to differentiate from the bulk plasmon signal. To further confirm the presence of surface plasmons, aloof EELS experiments were conducted where the electron beam does not interact with the sample directly but does so through long-range Coulomb interactions with the evanescent field of the surface plasmon resonances.50 As shown in Fig. 4(b), inset, four different areas were chosen that are close to the NPs but not directly on the sample. The bulk plasmon peaks that appear at higher energies are no longer present in the aloof mode, which is expected since the electron beam is not directly interacting with the particles [Fig. 4(b)]. It is possible that in spot 1, the interaction with the electron beam might be too weak to observe a surface plasmon peak from Cr2N. In spots 2–4, the aloof EELS spectra indicate the presence of a surface plasmon signal between 5 and 15 eV (i.e., ∼80 to 250 nm), which aligns with the calculations that indicate LSPR to be present in this energy region for NPs with a diameter below 50 nm.
Figure 5 shows EELS loss probability maps with the energy ranges that were chosen based on the peak positions in the low-loss EELS spectra. Between the energies of 3.06 and 4.9 eV, signal is mostly arising in between the particles indicating plasmonic coupling, which was also observed in the solution absorbance measurements. At energies between 6.96 and 8.60 eV, the map shows higher intensities around the NPs, likely originating from surface plasmons. At higher energies (>9 eV), most of the sample appears bright, which can be attributed to both the surface and bulk plasmons.
IV. CONCLUSIONS
Plasmonic Cr2N nanocrystals were successfully synthesized using solid-state nitridation of Cr2O3 with Mg3N2. This resulted in the formation of water dispersible Cr2N NPs with an average diameter of 9 ± 5 nm and a positively charged surface. The particles had a varying thickness of oxide and oxynitride layers around the nitride core as determined from XPS and TEM analysis. The calculations predicted these NPs to have LSPR below 200 nm, which aligned with the experimental observations. While a strong LSPR was not observed in the absorbance spectrum limited by the wavelength cutoff, low-loss EELS spectra showed the presence of surface and bulk plasmon resonances below 200 nm. Cr2N can be a promising material for UV-plasmonic applications and future studies will focus on exploring their thermal behavior.
SUPPLEMENTARY MATERIAL
See the supplementary material for TEM images of Cr2O3 and Cr2N NPs, SAED of Cr2N NPs, and calculated extinction cross sections of Au and plasmonic nitrides of varying NP sizes (Figs. S1–S5).
ACKNOWLEDGMENTS
The authors acknowledge funding from the Canada Foundation for Innovation (CFI) and the New Frontiers Research Fund. Dr. Carmen Andrei (Canadian Center for Microscopy) and Dr. Craig Bennett (Acadia University) are thanked for assistance with the TEM, EELS elemental mapping, and SAED experiments. Professor G. Gagnon and H. Daurie are thanked for assistance with DLS and zeta potential measurements. CMC Microsystems is thanked for providing access to COMSOL software. R.A.K. acknowledges funding from Sumner and the Natural Sciences and Engineering Research Council of Canada (NSERC) Graduate Fellowships and Y.E.M. acknowledges Ocean Frontier Institute for a postdoctoral fellowship. The XPS analysis is based on the work supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award No. DE-SC-0020301. (S)TEM and EELS work was performed at the Canadian Centre for Electron Microscopy, a Canada Foundation for Innovation Major Science Initiatives funded facility (also supported by NSERC and other government agencies).
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Reem A. Karaballi: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – original draft (lead). Yashar Esfahani Monfared: Data curation (equal); Formal analysis (equal); Methodology (equal); Writing – review & editing (equal). Isobel C. Bicket: Data curation (equal); Formal analysis (equal); Methodology (equal); Software (equal); Writing – review & editing (equal). Robert H. Coridan: Data curation (equal); Formal analysis (equal); Investigation (equal); Writing – review & editing (equal). Mita Dasog: Conceptualization (equal); Funding acquisition (lead); Project administration (equal); Supervision (lead); Writing – review & editing (equal).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request. Python software used to process EELS data are available at https://zenodo.org/record/807763#.YsuIhnbMLb0.