Intrinsic Visible Plasmonic Properties of Colloidal PtIn2 Intermetallic Nanoparticles

Abstract Materials that intrinsically exhibit localized surface plasmon resonance (LSPR) in the visible region have been predominantly researched on nanoparticles (NPs) composed of coinage metals, namely Au, Ag, and Cu. Here, as a coinage metal‐free intermetallic NPs, colloidal PtIn2 NPs with a C1 (CaF2‐type) crystal structure are synthesized by the liquid phase method, which evidently exhibit LSPR at wavelengths similar to face‐centered cubic (fcc)‐Au NPs. Computational simulations pointed out differences in the electronic structure and photo‐excited electron dynamics between C1‐PtIn2 and fcc‐Au NPs; reduces interband transition and stronger screening with smaller number of bound d‐electrons compare with fcc‐Au are unique origins of the visible plasmonic nature of C1‐PtIn2 NPs. These results strongly indicate that the intermetallic NPs are expected to address the development of alternative plasmonic materials by tuning their crystal structure and composition.

The size and morphology of each NP were characterized by TEM (HITACHI HT7820) operating at 200 kV.NPs were purified and a drop of solution was drop-casted on carbon-coated Cu grids for TEM.The atomic resolution analytical TEM (JEOL JEM-ARM200CF) observation was operated at an acceleration voltage of 120 kV.A Wiener filter was applied to original TEM-EDX images for noise reduction [41] using DigitalMicrograph Software (Gatan).
The crystal structure was confirmed at multiple locations with and without the filter.
HAADF images and LSPR properties of single NP were obtained by EELS combined with monochromated STEM (JEOL JEM-ARM200F).STEM-EELS was conducted at an acceleration voltage of 200 kV.We heated the SiN membrane at 400°C in the STEM experiments to prevent contamination from organic molecules.The crystal structure of PtIn2 at 400°C was confirmed by high-temperature XRD (Figure S10).
The crystalline structure of the NPs at room temperature was characterized by XRD (PANalytical AERIS).More-detailed XRD patterns were obtained by synchrotron X-ray radiation experiments performed at the SPring-8 BL02B2 (0.5272 Å).In situ high-temperature XRD was measured with a furnace (PANalytical X'Pert PRO MPD).
The elemental ratios of the synthesized NPs were determined by EDX (AMETEK Apollo XF) combined with SEM (HITACHI S-4800).
The optical response of LSPR was determined by UV-vis spectroscopy (HITACHI UH5700) with quartz cuvettes.SERS spectra were obtained with a Laser Raman Microspectroscopy System Nanofinder 30 (Tokyo Instruments Inc.) equipped with a 532-nm laser diode-pumped solidstate laser (SOC J050GS-1H) at an output power of 5.132 mW through a 50× microscope objective (Olympus MPlanFLN) over the range of 590-1819 cm −1 .The exposure time was 1 s and not accumulated (only 1×) to avoid heavy damage to rhodamine 6G (R6G).A chargecoupled device camera (1024 px × 255 px) operating at −70°C was used as the detector.To verify that the spectra were reproducible, more than 10 different areas were measured for each sample.

Synthesis of Pt seed NPs
To obtain PtIn2 NPs, Pt seed NPs were synthesized first.Oleic acid (31.5 mmol, 10 mL) and oleylamine (97.2 mmol, 32 mL) were mixed in a 100-mL, three-necked flask; followed by stirring at 220°C in an N2 atmosphere.Pt(acac)2 (0.5 mmol, 196.68 mg) was added to oleylamine (24.3 mmol, 8 mL) in a 50-mL, three-necked flask; followed by stirring and heating in an N2 atmosphere.After the Pt(acac)2 was dissolved in the oleylamine, the solution turned bright yellow at ca. 120°C.The mixture was quickly injected into the preheated oleic acid/oleylamine solution.Immediately after injection, the color of the solution changed to blackish brown after 1 min and black after 5 min.The solution was maintained at 220°C for 30 min and cooled at room temperature.Four-fold scale-up was carried out through this method.
Pt seed NPs were well dispersed in chloroform.To remove excess oleic acid and oleylamine, 15 mL of ethanol was added as a poor solvent to 25 mL of crude solution collected with chloroform, followed by centrifugation for 10 min at 9390×g, and then stored in chloroform.

Synthesis of PtIn2 NPs
C1-PtIn2 NPs were synthesized by reacting Pt seed NPs with In amide complexes in oleylamine under strong basic conditions. [26]The expected reaction is as follows: a strong Brønsted base deprotonates a long-chain primary amine, which reacts with InCl3 to form an In oleylamide. [26]he in-situ formed amide is then reduced or thermally decomposed to react with the Pt seed NPs to form the target product.In this method, the size of the PtIn2 NPs is larger than that of the Pt seed NPs, and the formation of indium oxide during the reaction process is also suppressed because of the absence of oxygen atom in precursors and solvents.
Pt seed NPs were additionally purified for the next step.Pt NPs (0.5 mmol of Pt atoms) were collected in a 100-mL three-necked flask with chloroform; added to distilled oleylamine (30.4 mmol, 10 mL); and sonicated for a few minutes.Then, chloroform was removed by evaporation, additionally dried under vacuum (45 min at 120°C), and heated to 180°C under N2.LDA-oleylamine solution was injected, after 30 s, followed by injection of InCl3oleylamine solution.The reaction solution was maintained at 180°C for 2 h and then cooled to room temperature.Although the oleylamine used as a solvent provided the dispersibility of the NPs as a ligand, the dispersibility of the synthesized NPs was improved by injecting 2 mL of oleic acid when the temperature reached 60°C during cooling before opening to the air.Largersized PtIn2 NPs were obtained by setting the heating temperature to 240°C to induce the fusion of NPs (Figure 2d).
As is clear from the phase diagram (Figure S3), the region where PtIn2 is stable is narrow.
When the reaction conditions differed, Pt2In3 and Pt3In7 phases were readily evident.
The sizes of the NPs were statistically determined from the TEM images.Size distributions of NPs are shown in Figure S2.

Purification of C1-PtIn2 NPs
There were two steps of purification: (1) removal of oleylamine and (2) size separation.In step (1), 20 mL of ethanol was added to 15 mL of a chloroform-diluted sample solution and centrifuged at 9390×g for 10 min.In step (2), a small quantity (~5 mL) of ethanol was added to the chloroform dispersion, and the mixture was centrifuged.Centrifugal acceleration was modified depending on the NPs size to be separated.These steps were repeated several times and the target size of the PtIn2 NPs was collected.To prevent a decrease in ligand concentration, oleylamine and oleic acid were added during the purification.PtIn2 NPs were well dispersed in chloroform and retains absorption at least two years even when stored in dispersed solution after purification.

Synthesis and purification of Au NPs
Oleylamine-capped Au NPs were synthesized as reported. [42]Briefly, oleylamine (8.8 mmol, 2.9 mL) was added to toluene (48 mL) in a 200-mL, three-necked flask; followed by stirring and boiling at ca. 120°C.Then, a solution of HAuCl4⋅4H2O (53.7 mg, 0.13 mmol) dissolved in oleylamine (1.2 mL) and toluene (2.0 mL) were quickly injected.The reaction mixture changed to bright magenta in 5 min and gradually deepened over the course of 25 min.Heating was stopped after 2 h and 17-nm-sized Au NPs were precipitated by adding 100 mL of methanol.
The precipitates were centrifuged at 9390×g for 5 min with methanol as a poor solvent.
The Au NPs were purified again by redispersion in toluene.Small-sized Au NPs were removed by centrifugation by using only toluene.

In situ high-temperature XRD of C-supported PtIn2 NPs
To prevent interparticle fusion as much as possible, a PtIn2 NPs chloroform dispersion was added into carbon (CABOT Vulcane XC-72) chloroform dispersion and left to stand.After washing with ethanol, the C-supported sample was dried thoroughly and packed in an Al2O3 holder.The sample was covered with a readily burnable polymeric film, Prolene Thin-Film (Chemplex), to prevent the sample from blowing off during evacuation.
In high-temperature XRD measurements, a vacuum was first drawn to ca. 7 Pa.Then, the XRD pattern was measured at room temperature, following which the temperature was increased at a constant heating rate of 10°C min −1 , and the in situ XRD pattern was measured every 100°C from 100°C to 400°C.The measurement was repeated 3 times (ca. 2 h) at 400°C to fully confirm the phase stability at 400°C, the same temperature used in EELS measurements.
Finally, XRD measurements were carried out again after the temperature decreased to 40°C.
XRD patterns of C-support, Al2O3 holder, and prolene are shown by reference measurements under atmospheric conditions at room temperature.
The quality factors (Q-factors) were evaluated from the measured extinction spectra of the NP dispersion.Q = w / G, where w is the center of the plasmon resonance with a full width at half maximum G.

XRD pattern simulation
The XRD pattern of PtIn2 was simulated using CrystalMaker X and CrystalDiffract 6.9 (CrystalMaker Software Ltd.) (Figure S14).The occupancy of Pt at the Pt sites in the C1 structure was taken as an independent variable, and the occupancy of In sites was taken as a dependent variable so that the overall molar ratio of Pt : In = 1 : 2 was maintained (Table S1).
As shown in Figure S14, the superlattice peaks of C1 structure are 111, 200, and 311 from lower angle.For example, the intensity of the 111 peak is smaller for lower Pt occupancy of the Pt sites.

Theory and computational details
The time-dependent Kohn-Sham equation solved in the Scalable Ab-initio Light-Matter simulator for Optics and Nanoscience (SALMON) is as follows (in atomic units): Where  !"# (, ),  $ (, ),  %& (, ), and  '() (, ) are the nuclear attraction potential, Hartree potential, exchange-correlation potential, and external potential of an applied laser field, respectively.In this study, we adopted the laser pulse, the functional form of which is where  * () is the laser field given by where w is the laser frequency, and t determines the laser pulse duration.The electron dynamics under the laser pulse were analyzed by using the Fourier-transformed formula where  +' is the normalized factor given by We used only the imaginary part because the sin-type wave was used [Eq.( 3)].Then, the imaginary part of Δρ exhibits the spatial configuration of the resonance response of the photoexcited electrons. [4]The Fourier-transformed electric field was also analyzed to evaluate the photogenerated electric field around the NPs.
We investigated the photoexcited electron dynamics of large clusters of ca.600 atoms, Au561 and Pt249In432, with SALMON. [43]The geometrical structures were determined from bulk crystal data.All the electrons other than the Au 5d 10 6s 1 electrons were replaced with effective core potentials obtained by the Troullier-Martins scheme implemented in the fhi98PP program. [44,45]For the exchange-correlation potential, the local density approximation functional was used.The computation box and grid spacing were set to 46 Å × 46 Å × 46 Å and Δx = Δy = Δz = 0.25 Å, respectively.The propagation time was 12.8 fs for the oscillator strength and 20.0 fs for the laser-induced electron dynamics.The applied laser pulse intensity was set to 10 9 W cm −2 .room temperature, the peak of the Prolene film used for the cover appeared strongly, but this film burned out at 100°C.We observed the peak that corresponds to PtIn2 at high intensity even at 400°C, although it was slightly shifted towards lower angles due to the thermal expansion of the lattice.At 400°C, we identified a small quantity of Pt2In3 in the diffraction pattern as the binary phase with a broader FWHM than that of PtIn2 and did not increase over time.Thus, we propose that nearly unreacted Pt seed NPs contained as minor components reacted with the PtIn2 phase to give the Pt2In3 phase.The aforementioned examination strongly supports our proposal that an isolated PtIn2 NP can retain its C1 structure even during EELS measurements at 400°C.S1).220 peak intensity was taken as 100%.220 peak intensity was set as 100%.Pt occupancy at Pt sites vs. f) 111/220 ratio, g) 200/220, and h) 311/220 of XRD simulation results.

Figure S1 .
Figure S1.Rietveld refinement of 44-nm PtIn2 NPs.We measured the synchrotron XRD pattern of 44-nm PtIn2 NPs (red, solid line) with λ = 0.527 Å.We conducted Rietveld analysis by using the standard pattern of Pt and two different patterns of PtIn2 with different sizes obtained from Ref. [32].Total simulated result (blue dotted line) is the sum of the simulation results for two different sized C1-PtIn2 (violet and light blue dotted lines) and Pt (brown dotted line).The reference patterns for PtIn2 (violet solid line, PDF #01-071-5016) and Pt (brown solid line, PDF #00-004-0802) calculated at λ = 0.527 Å and the difference between the experimentally observed and simulated patterns (gray solid line) are shown.Rwp was 6.87% and the goodnessof-fit (GOF) was 3.45.The NPs comprised C1-PtIn2 phase (92.6 wt%) with smaller fcc-Ptbased crystallites (7.4 wt%).

Figure S5 .Figure S6 .
Figure S5.Dependence of λmax of PtIn2 NPs on the surrounding refractive index.We plotted the λmax values of C1-PtIn2 NPs dissolved in cyclohexane (n = 1.4268, red), chloroform (n = 1.4467, green), and toluene (n = 1.4978, blue) against the refractive indices of the solvents.PtIn2 NPs (17 nm) were used because of their high solubility in these solvents.The λmax is proportional to the refractive index of the solvent, which also supports the LSPR feature of C1-PtIn2 NPs.The sensitivity of LSPR λmax to the changes in the surrounding RIU is defined as λmax/RIU = 42.4.

Figure S7 .
Figure S7.The line-scan EDS curves for Pt (blue), In (red), and O (yellow).a) Two line-scan positions (i, ii) and b) the line-scan curves.As seen in line-scan curves in (i), Pt and In exist alternately.On the other hand, the longitudinal line-scan curves in (ii) clearly show the formation of a double-shell structure at the surface.

Figure S8 .
Figure S8.STEM-EELS maps of a single PtIn2 NP.EEL spectra of 44-nm PtIn2 NP exhibit higher-energy losses at a) 4.7-5.1 eV, b) 8.2-8.6 eV, and c) 14.0-14.4eV.(a) and (b) show the large intensity at the surface of the NP.This corresponds to LSPR at the high energy reportedhere (i.e., higher-order LSPR), which we consider to be dipole and quadrupole modes.In contrast, interband transitions and/or bulk plasmons (c) corresponded to a large intensity at the center.

Figure S9 .
Figure S9.Crystal structure analysis of a single PtIn2 NP by high-resolution TEM.a, b) Highresolution TEM images of a PtIn2 NP measured at 400°C after EELS measurements.c) FFT image of the red square region of (b).The spots corresponding to the 200, 220, and 400 reflections of the C1 structure are indicated in yellow.d) Noise-filtered inverse FFT result of (c).The inset in (d) is the crystal lattice of C1-PtIn2, where lighter and darker dots correspond to Pt and In atoms, respectively.The atomic arrangement in (d) is similar to that in the inset.

Figure S10 .
Figure S10.High-temperature XRD measurements of C1-PtIn2 NPs supported on carbon.The stability of PtIn2 at high temperatures was investigated every 100°C from room temperature.At

Figure S11 .
Figure S11.TEM images of Au NPs and optical properties of SERS samples.a) TEM images of Au NPs used for SERS.Nearly spherical NPs are 17 ± 4 nm in diameter.b) Normalised extinction spectra of 24-nm C1-PtIn2 NP dispersion in chloroform (violet), Au NP dispersion (pink), and R6G solution (black); with λmax at 534, 527, and 530 nm, respectively.The quality factor of this resonance can be described by Q =w / G. where w is the centre of the plasmon resonance with a FWHM of G. QPtIn2 = 2.80 was smaller than QAu = 7.98.

Figure S12 .
Figure S12.SERS measurements.SERS spectra of R6G (black), R6G with 24-nm PtIn2 NPs (violet), and R6G with 17-nm Au NPs (pink) on a Si wafer with a 532-nm laser as the excitation light source.Although the R6G sample exhibited almost no clear peaks, both NPs drastically enhanced the SERS signals caused by the electric field enhancement effect of LSPR.As expected from the Q values, the SERS enhancement by the Au NPs was stronger than that by the PtIn2 NPs.

Figure S13 .
Figure S13.SEM image of the substrate used for surface-enhanced Raman measurements: a) Rhodamine 6G (R6G) with Au NPs, b) with C1-PtIn2 NPs and c) only R6G on Si wafer.