Effect of Amino Acids on the Synthesis of NiFe₂O₄/Au Hybrid Nanoparticles

Abstract

Hybrid nanoparticles, composed of magnetic oxides and gold, have garnered significant interest due to their potential applications in various fields, including catalysis, diagnostics, and nanomedicine. In this study, the effect of reaction parameters on the reduction of HAuCl₄ by different non-sulfur amino acids (glycine, L-serine, L-tryptophan, and L-tyrosine) was determined using the design of experiment (DOE) method. The results of the analysis of the regression equations were used to select the conditions and develop a methodology for the preparation of the hybrid magnetic NiFe₂O₄/Au nanoparticles (NPs) by direct reduction of gold on the magnetic surface using the aforementioned amino acids as the reducing and stabilizing agents simultaneously. The materials were characterized using XRD, TEM, XPS, and Vis/FTIR spectroscopy. The results indicate the successful synthesis of magnetic NiFe₂O₄/Au nanoparticles with all amino acids used, but the size of the gold crystals, their surface density, and the details of the NP structure (inlaid or a core–shell structure) depend on the amino acid used. The mechanism of the gold deposition on the magnetic core surface and the difference in the effect of various amino acids are discussed. The developed synthesis strategy can be extended to other metal ferrites and iron oxides.

1. Introduction

In recent years, research on the preparation, study, and application of hybrid nanoparticles has developed rapidly. Hybrid materials are obtained by the interaction of chemically different components that form a specific spatial structure with improved characteristics compared to its individual components [1].

Diverse hybrid nanostructures such as core–shell [2,3,4]; Janus-like [5,6,7], combining two (or more) components close in size; and structures with nanocomponents incorporated or deposited on the surface (inlaid particles) [8,9] have been mentioned.

The interest in these materials is determined by the new useful properties that a hybrid nanomaterial achieves as a result of combining the individual components. As a rule, the application of hybrid materials is determined by the properties of the main component. The add-on components either enhance the functional properties in a specific application or increase the overall efficiency of the system. Thus, nanostructures with magnetic nanoparticles that can be manipulated by an external magnetic field are used as contrast agents in magnetic resonance imaging or for targeted drug delivery [10,11]. It is convenient to use nanoparticles of magnetically soft ferrimagnetics with low, but not zero, values of coercivity and residual magnetization in targeted delivery because these materials facilitate their magnetic control with insignificant particle agglomeration in the absence of a magnetic field [12,13]. One such material is nickel ferrite, which was chosen for this study. The addition of gold to magnetic NPs imparts the inherent properties of gold such as low cytotoxicity, high electrical and thermal conductivity, chemical inertness, and stability in biological media [14].

Additionally, the catalytic activity of such nanoparticles has been demonstrated, in particular, in the reduction of nitroaromatic compounds to amines [15] or in the photodegradation of organic compounds [16], which allows the application of these materials for the removal of environmentally hazardous substances [17].

There are two main approaches for producing magnetic gold hybrid nanoparticles. In the first approach, gold seeds are deposited on the surface of magnetic nanoparticles. This is usually realized with a polymer, such as polyethyleneimine [18,19], after or during the polymer shell formation on the surface of the magnetic nanoparticles. In this case, gold seeds are obtained separately, for example via borohydride or sodium citrate reduction Au(III) species [20,21]. The attached Au⁰ seeds can be grown into a continuous gold shell by adding a gold precursor and a reducing agent.

Another approach is to deposit gold directly on the surface of the magnetic core without the use of Au seeds and a polymer. A significant mismatch in the crystal structure of the magnetic material and gold makes it difficult to fix Au⁰ seeds on the surface. This leads to “parasitic” reduction of gold in the solution with the formation of separated large (50–80 nm) particles. Careful selection of the most appropriate stabilizer and reducing agent is the main way to avoid this problem.

Amino acids can be suitable reducing agents because their molecules contain several functional groups (e.g., -COOH, -NH₂, =S, -OH) that can also stabilize and anchor gold nanoclusters to magnetic particles. It is well known that amino acids have reducing properties due to electron donor atoms (S and N). Sulfur-containing amino acids (methionine, cysteine, cystine, taurine, glycylmethionine), whose interaction with gold salts has been well studied, are the most easily oxidized [22,23,24,25,26,27,28,29,30,31]. The interaction of the [AuCl₄]⁻ ions with non-sulfur amino acids has been much less studied. In the paper [32], the reducing activity towards Au(III) species of 18 amino acids is shown, among which the maximum yield of Au⁰ was observed for aspartic acid (C₄H₇NO₄), arginine (C₆H₁₄N₄O₂), threonine (C₄H₉NO₃), tryptophan (C₁₁H₁₂N₂O₂), and valine (C₅H₁₁NO₂). However, the authors used a 400 W xenon lamp to initiate and accelerate the process. Photoirradiation of the reaction mixture of HAuCl₄ and glycine was also used in [33]. In [34,35], gold nanoparticles with a size of 5–30 nm were obtained by prolonged boiling of HAuCl₄ with glutamic acid C₅H₉NO₄, phenylalanine C₉H₁₁NO₂ and tryptophan C₁₁H₁₂N₂O₂. In [36], the kinetics of the reduction of Au(III) to Au(I) by glycine in acetate buffer solution at pH ~4.5 and a large excess of amino acid was studied. Thus, in our opinion, there are not enough systematic studies of gold(III) species reduction by non-sulfur amino acids in the literature. In addition, some research has been carried out under rather harsh conditions such as UV photoirradiation or prolonged boiling.

In this paper, the effect of the reaction parameters on the reduction of HAuCl₄ by amino acids was determined by the design of experiment (DOE) method. Different types of amino acids were studied: the aliphatic nonpolar and polar glycine (Gly) and L-serine (Ser) and the aromatic nonpolar and polar L-tryptophan (Tryp) and L-tyrosine (Tyr). The data obtained were used to select the conditions and develop a methodology for the preparation of the NiFe₂O₄/Au nanoparticles. To the best of our knowledge, our study is the first to successfully obtain magnetic hybrid NPs by depositing gold with only non-sulfur amino acids without any stabilizer on the surface of nickel ferrite NPs under ambient conditions.

2. Materials and Methods

2.1. Chemicals

Tetrachloroauric (III) acid (HAuCl₄∙6H₂O), L-methionine (C₅H₁₁NO₂S), glycine (C₂H₅NO₂), L-serine (C₃H₇NO₃), L-tryptophan (C₁₁H₁₂N₂O₂), and L-tyrosine (C₉H₁₁NO₃) were of analytical grade and were purchased from Sigma-Aldrich; Ni(NO₃)₂·6H₂O, Fe(NO₃)₃∙9H₂O, NaOH, and other chemicals were of chemically pure grade and were purchased from Chemreaktivsnab (Ufa, Russia). All chemicals were used as received. The strong-base anion-exchange resin AV-17-8 was produced by “LLC Production company TOKEM” (Kemerovo, Russia) in the chloride form with a bead size of 0.4–0.6 mm (Russian GOST 20301-74 [37]). This resin has a gel matrix, based on polystyrene cross-linked with divinylbenzene and the functional group quaternary ammonium (type I). It is widely used in processes involving separation, purification, and decontamination.

2.2. Reduction of Tetrachloroaurate(III) Ions with Amino Acids

The interaction of Au(III) with amino acids was investigated according to the following procedure: Amino acids (volume and concentration were provided in

Table 1. The natural values of the control factors at the high (x_max) and low (x_min) levels.

Variation Levels	X1		X2	X3	X4	X5	X6
	Model 1	Model 2
high	L-tyrosine	glycine	0.03 M	25 °C	30 min	9	3
low	L-tryptophan	L-serine	0.1 M	37 °C	60 min	11	5

) were added to 2 mL of 0.0002 M chloroauric acid, pH was adjusted by a 0.2 M solution of NaOH to 9 or 11, and the mixture was diluted with distilled water to 15 mL. The mixture was stirred on a magnetic stirrer at 25 or 37 °C for 30 or 60 min. A more complete description of the study’s conditions can be found in Table 1. It was observed that the color of the solution changed to purple due to the formation of gold nanoparticles. The hydrosols thus obtained were transferred into a 1 cm cuvette for subsequent investigation by using a TUV6U spectrophotometer (SILab, Hangzhou, China).

2.3. Synthesis of Nickel Ferrite Nanoparticles

In this study, nickel ferrite nanoparticles were synthesized by anion-exchange resin precipitation technique [38,39,40] based on the ion exchange between the anions of the aqueous solution and a solid substance (an anion-exchange resin, containing the OH groups). The process of obtaining the precursor of nickel ferrite can be described by the following equation:

Fe(NO₃)₃ + Ni(NO₃)₂ + 5R-OH → Fe(OH)₃ + Ni(OH)₂ + 5R-Cl,

(1)

where R is the anion-exchange resin AV-17-8. The precipitation of metal hydroxides occurs at the resin–solution interface, namely on the resin beads. When the thickness of the surface deposit reaches 1–1.5 μm, it is exfoliated, and an individual product phase forms.

The strong-base anion-exchange resin AV-17-8 in Cl-form was treated with 1 M NaOH for 1 h and then 5–6 times with 2 M NaOH for 1 h. The last portion was kept for 24 h. After that, the anion-exchange resin was washed with water to pH = 6–7 and dried at 60 °C. The total exchange capacity (TEP) of AV-17-8 was determined by 0.1 M HCl (TEP = 1.4 meq g⁻¹).

Excess (150%) AV-17-8(OH) (34.3 g) was added to a solution containing nickel and ferric salts (V = 10 mL, C = 0.4 M and 0.8 M, respectively). The quantity of reagents was chosen according to the stoichiometry of the reaction. The mixture was stirred (120 rpm) at a temperature of 25 °C, using a magnetic stirrer, for 1.5 h. To remove the anion-exchange resin beads from the reaction products, a sieve with round holes (0.16 mm in diameter) was used; the precipitate was centrifuged, washed with distilled water, dried in air at 80 °C, and then annealed in a muffle furnace at 650 °C for 3 h [41].

2.4. Synthesis of Hybrid NiFe₂O₄/Au Nanoparticles

Nickel ferrite nanoparticles (0.025 g) were dispersed in 20 mL amino acid solution (0.1 mol/L, L-methionine, glycine, L-serine, L-tryptophan) by ultrasonic agitation (ultrasonic bath “Sapphire”, 35 kHz, 100 W, Sapphire, Moscow, Russia) for 30 min. Chloroauric acid (0.3 mL, 0.1 mol/L) was added to the mixture; the pH was adjusted by a 0.2 M solution of NaOH to 11 (most of the acidic and basic sites of the amino acids were deprotonated at this pH). The deposition process was conducted at 37 °C for 4 h under stirring (800 rpm). The resulting nanoparticles were separated by magnetic separation and thoroughly washed with distilled water and ethanol.

2.5. Nanoparticle Characterization

X-ray powder diffraction (XRD) data were collected using a Shimadzu XDR-600 diffractometer (Shimadzu Corporation, Kyoto, Japan) with CuKα radiation. Scanning parameters included a range from 5 to 70° on the 2θ scale with a step size of 0.013° and a time interval of 50 s/step, conducted in air at room temperature. The size of the coherent scattering region of the obtained particles was calculated using the Debye–Scherrer formula:

d = (0.94∙λ)/(β∙cosθ),

(2)

where d is the nanoparticles’ crystalline size, λ is the wavelength of the X-ray beam used, β is the full width at half maximum (FWHM) of the peak, and θ is the Bragg angle.

Transmission electron microscopy (TEM) analysis was conducted using a Hitachi 7700 M microscope (Hitachi Corporation, Tokyo, Japan) at an accelerating voltage of 100 kV. The STATISTICA package was used for statistical data processing. The particle size distribution histograms were obtained from more than 300 particles [4].

Fourier-transform infrared (FTIR) spectra were recorded using the multiple internal reflection (MIR) attachment of the Fourier-transform infrared spectrometer Bruker Tensor 27 (Bruker, Bremen, Germany). Initially, the background spectrum was obtained from a crystal made of ZnSe. The spectral range was 4000–400 cm⁻¹, with a resolution of 4 cm⁻¹ and 32 scans. The resulting spectrum was obtained by subtracting the background spectrum from the sample spectrum. Spectra were processed using the OPUS 7.5 software package.

UV–Vis absorption spectra were obtained using a TUV6U UV–Vis spectrophotometer (SILab, Hangzhou, China) with a glass cell having an optical path length of 1 cm, covering the range from 400 to 1000 nm.

Measurements of the hydrodynamic diameter of particles were determined with a Malvern Zetasizer Nano ZS instrument (Malvern Panalytical, Worcestershire, UK) with a laser wavelength of 632.8 nm and a scattering angle of 173°. Nickel ferrite powders (0.1 g) were dispersed in 20 mL of distilled water under ultrasonic treatment for 10 min. One milliliter of the sol was transferred to a plastic cuvette (l = 1 cm).

X-ray photoelectron spectroscopy (XPS) studies were performed using a hydrosol that had been dried with highly oriented pyrolytic graphite (HOPG) and gently rinsed with water. The resulting spectra were acquired using a SPECS spectrometer (SPECS Gmbh, Berlin, Germany) equipped with a PHOIBOS 150-MCD-9 hemispherical electron analyzer (SPECS GmbH, Berlin, Germany). Spectra were recorded upon excitation with a monochromatic radiation of AlKα (E = 1486.6 eV) [4,41]. The analyzer pass energy was set to 10 electronvolts (eV) for high-resolution scans and 20 eV for survey spectra. To ensure the uniformity of the electrostatic charge on the samples, an electron flood gun was employed. The C 1s peak at 284.45 eV from HOPG was utilized as a reference for the alignment of the spectra. The high-resolution spectra were then fitted after subtraction of Shirley-type background with Gaussian–Lorentzian peak profiles using CasaXPS software (version 2.3.16, Casa Software, Teignmouth, UK) [4,41].

3. Results and Discussion

3.1. Interaction of Tetrachloroaurate(III) Ions with Amino Acids

In this study, amino acids, namely glycine, L-serine, L-tryptophan, and L-tyrosine (E⁰_Gly = −1.44 V, E⁰_Ser = −1.08 V, E⁰_Tryp = −1.53 V, E⁰_Tyr = −1.39 V), that possess the capacity to reduce Au³⁺ to Au⁰ (E⁰ = +1.498 B) were used [42,43].

Initially, the focus was on the influence of reaction parameters of gold reduction with amino acids on the yield of the Au nanoparticle (NP) in solutions. We used the design of the experiment (DOE) method—the fractional (1/16) two-level, seven-factor factorial design (FFD 2⁷⁻⁴), which minimizes the number of runs without sacrificing quality [44,45]. This approach is advantageous for the purpose of planning, conducting, analyzing, and interpreting the results of experiments to estimate the factors controlling the value of a parameter or group of parameters. The following reaction parameters were used as the control factors (X_i):

X₁—amino acid (discrete factor: L-tyrosine, L-tryptophan (model 1); glycine, L-serine (model 2));

X₂—concentration of the amino acid;

X₃—temperature;

X₄—synthesis time;

X₅—pH;

X₆—volume ratio of gold and amino acid solutions;

X₇ = X₁ × X₂ × X₃.

The interval (factor’s domain) of variation (I) of the controlling factors was determined by the following formula:

$$\mathrm{I}={\displaystyle \frac{{\mathrm{x}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{x}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}{2}},$$

(3)

where x_max is the maximum value of the factor and x_min is the minimum value of the factor.

To simplify the design of the experiment in the DOE method, instead of the natural values of the factors $\widetilde{x_j}$, the coded values of the factors obtained by normalizing procedure were used:

$${\mathrm{x}}_{\mathrm{j}}={\displaystyle \frac{{\tilde{\mathrm{x}}}_{\mathrm{j}}-{\mathrm{x}}_{\mathrm{j}0}}{{\mathrm{I}}_{\mathrm{j}}}},$$

(4)

where $\widetilde{{\mathrm{x}}_{\mathrm{j}}}$ is the natural value of the factor, I_j is the interval of variation, x_j0 is the zero level of the factor x_j0 = (x_max + x_min)/2, and x_j is the coded value.

As a result, x_j = +1 (high level) or −1 (low level). The values of reaction parameters at high and low levels are presented in Table 1. We used the first-degree model of the FFD 2⁷⁻⁴ design, and the result of the experiment in this case is a regression equation (linear model), describing the behavior of the object:

Y = b₀ + ∑ b_j × x_j

(5)

where Y is the experimental result (response function) and b₀ and b_j are the coefficients of the polynomial (regression coefficients).

As each factor has only two levels in this design, two regression models (models 1 and 2) were obtained to demonstrate the effect of the four amino acids on gold reduction.

The experimentally determined absorbance (A_i) of the obtained NPs hydrosols at a wavelength of 540 nm (representative of the surface plasmon resonance (SPR) peak of gold nanostructures) was selected as the response (Y) in both models, as it is directly related to the amount of reduced gold. It is well established that the SPR peak of nanoparticles of different metals is characteristic, with the intensity of the peak reflecting the amount of metallic gold and its position depending on the size and morphology of the NPs. The optical properties of spherical particles’ dispersion are predicted by the Mie–Drude theory [46]. Equation (5) estimates the size of nanoparticles based on the dielectric properties of gold and solvent and the value of the extinction coefficient:

$${\mathrm{C}}_{\mathrm{e}\mathrm{x}\mathrm{t}}=24\mathsf{\pi }\mathrm{R}{\mathsf{\epsilon }}_{\mathrm{M}}^{3/2}/\mathsf{\lambda }{\mathsf{\epsilon }}^{″},$$

(6)

where R is the radius of the nanoparticles, ε_M is the dielectric constant of the media surrounding the nanoparticle, ${\mathsf{\epsilon }}^{″}$ is the imaginary part of the dielectric constant of gold, λ is the wavelength, and C_ext is the extinction coefficient.

The plasmon peak of spherical gold nanoparticles with a size of 2–5 nm is located around 520 nm (in water). With an increase in the size of gold NPs, the absorption band shifts towards longer wavelengths (bathochromic shift).

The standard FFD 2⁷⁻⁴ design matrix delineating the experimental conditions (

Table 2. FFD 2⁷⁻⁴ matrix.

Run No.	X0	X1	X2	X3	X4	X5	X6	X7
1	+1	−1	−1	−1	+1	+1	+1	−1
2	+1	+1	−1	−1	−1	−1	+1	+1
3	+1	−1	+1	−1	−1	+1	−1	+1
4	+1	+1	+1	−1	+1	−1	−1	−1
5	+1	−1	−1	+1	+1	−1	−1	+1
6	+1	+1	−1	+1	−1	+1	−1	−1
7	+1	−1	+1	+1	−1	−1	+1	−1
8	+1	+1	+1	+1	+1	+1	+1	+1

) was used. Two sets of eight runs each were carried out for each model. The experimental results are shown in

Table 3. The measured response (A_i(exp.)), the response calculated from the postulated model (A_i(pred.)), and their difference.

Run No.	Model 1			Model 2
	Ai(exp.)	Ai(pred.)	|Ai(pred.) − Ai(exp.)|	Ai(exp.)	Ai(pred.)	|Ai(pred.) − Ai(exp.)|
1	0.407 ± 0.004	0.408	0.001	0.113 ± 0.003	0.111	0.002
2	0.079 ± 0.006	0.078	0.001	0.087 ± 0.003	0.089	0.002
3	0.413 ± 0.009	0.414	0.001	0.101 ± 0.006	0.101	0.000
4	0.084 ± 0.015	0.084	0.000	0.084 ± 0.010	0.079	0.005
5	0.421 ± 0.009	0.418	0.003	0.121 ± 0.006	0.123	0.002
6	0.083 ± 0.002	0.088	0.005	0.080 ± 0.006	0.077	0.003
7	0.383 ± 0.024	0.388	0.005	0.110 ± 0.007	0.113	0.003
8	0.062 ± 0.004	0.058	0.004	0.062 ± 0.003	0.067	0.005

.

The statistical processing of experimental data was performed in STATISTICA (DOE) software (Ver. 12.5.192.7). The regression coefficients b_i and their standard deviation values $\mathsf{\Delta }\mathrm{b}$ were calculated using Equations (6)–(8) with a confidence level of 95%.

$${\mathrm{b}}_{\mathrm{i}}={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{X}}_{\mathrm{i}\mathrm{j}}\cdot {\overline{\mathrm{Y}}}_{\mathrm{i}}}{\mathrm{N}}}$$

(7)

where X_ij is the coded value of the factor from the matrix, $\overline{{\mathrm{Y}}_{\mathrm{i}}}$ is the mean of the measured response, and N is the number of runs;

$${\mathrm{S}}_{\mathrm{b}}^2={\displaystyle \frac{{\mathrm{S}}_{\mathrm{Y}}^2}{\mathrm{N}}},$$

(8)

where ${\mathrm{S}}_{\mathrm{Y}}^2$ is the mean squared error of the measured response and ${\mathrm{S}}_{\mathrm{b}}^2$ is the mean squared error of the regression coefficient;

$$\mathsf{\Delta }\mathrm{b}={\displaystyle \frac{\pm {\mathrm{t}}_{\left(0.95\right)}\cdot {\mathrm{S}}_{\mathrm{b}}}{\sqrt{\mathrm{n}}}},$$

(9)

where $\pm t_{\left(0.95\right)}$ represents the critical values of t, n is the number of replications.

Table 4. Values of the coefficients of the regression equations.

Model	Δb	b0	b1	b2	b3	b4	b5	b6	b7
1	0.004	0.242	−0.165	−0.006	−0.004	0.002	−0.001	−0.009	0.002
2	0.003	0.095	−0.017	−0.005	−0.002	0.000	−0.006	−0.002	−0.002

shows the results of the statistical analysis of the data from the experiment.

Figure 1 shows the optical absorption spectra of the solutions from Run 5 (models 1 and 2). The spectra exhibit an SPR peak at 530–540 nm, indicating the formation of gold nanoparticles by the reduction of HAuCl₄ with different amino acids. According to literature data, this position of the SPR peak is consistent with the particle size of 25–30 nm [32].

Consequently, regression Equations (9) and (10) were obtained, which describe the effect of the reaction parameters on the absorbance at 540 nm. The regression coefficients are indicative of the degree to which a given factor influences the response function. If the regression coefficient is less than or equal to Δb, the factor is considered to be insignificant and is thus excluded from the regression equation. The direction of influence of this factor is indicated by the sign in front of the coefficient.

Model 1: A_i = 0.242 − 0.165 X₁ − 0.006 X₂ − 0.009 X₆;

(10)

Model 2: A_i = 0.095 − 0.017 X₁ − 0.005 X₂ − 0.006 X₅.

(11)

The lack-of-fit test was employed to ascertain the adequacy of the obtained models in describing the observed data. The goodness of the models was examined by the determination coefficients and the multiple correlation coefficients (R). The values of R were 0.962 and 0.981 for models 1 and 2, respectively, indicating a strong correlation between the observed and predicted values. As presented in Table 3, the differences between the experimental and predicted A_i were found to be minimal, nearly approaching zero, thereby validating the models.

The used amino acid, its concentration, the reagent volume ratio, and the pH were determined as critical factors affecting gold reduction. It should be noted that the degree of influence of these factors differs slightly between models 1 and 2. pH does not affect the reduction of gold by L-tyrosine and L-tryptophan (model 1). The volume ratio of gold and amino acid solutions is insignificant for model 2.

In these conditions, L-serine and L-tryptophan are the most effective reducing agents for gold. Glycine and L-tyrosine have weaker reducing properties, but a distinct SPR maximum is observed in the spectrum of the sample obtained by glycine in contrast to tyrosine (Figure 1b). So, L-serine, L-tryptophan, and glycine were selected for further investigation. L-methionine, used to synthesize NiFe₂O₄/Au and CoFe₂O₄/Au hybrid nanoparticles earlier [4,41], was also investigated for comparison.

The models that were obtained demonstrate that neither the synthesis time nor the temperature has a significant effect on the synthesis of Au⁰ NPs, likely due to the limited variation in these factors. In order to obtain more precise results, a new, improved test was carried out. Specifically, 2 mL of 0.002 M HAuCl₄ was added to 10 mL of 0.1 M amino acid solution, with the pH adjusted to 11 and the volume diluted to 15 mL. The mixture was heated at 50 °C for 1 h or 37 °C for 4 h. The analysis of the absorption spectra of the gold hydrosols obtained with L-serine, L-tryptophan, glycine, and L-methionine (Figure 2) indicates that at 50 °C, there is a formation of separated gold crystals of 40–60 nm (SPR peak is at 540–560 nm) for all amino acids. In addition, the appearance of a “gold mirror” on the walls of the reaction vessel is observed. Such “parasitic” gold recovery is undesirable because it leads to a significant non-productive waste of gold. Consequently, further experiments were carried out at 37 °C for 4 h.

3.2. Synthesis of NiFe₂O₄/Au Hybrid Nanoparticles

In this study, NiFe₂O₄ nanoparticles were synthesized by the anion-exchange resin precipitation approach from Ni(NO₃)₂ and Fe(NO₃)₃ solutions using anion-exchange resin AV-17-8 [38,39,40]. The representative XRD pattern of the NPs is shown in Figure 3c. The result aligns with the standard cubic spinel structure of bulk NiFe₂O₄ (JCPDS, Card No. 10-0325); no other secondary phase is observed. The reported lattice parameter of NiFe₂O₄ is a = 8.343 Å, which is close to previously documented values [4]. The average crystallite size determined from the peak broadening for the four most intense XRD peaks using the Scherrer equation (Equation (1)) was 20.4 ± 1.0 nm [4]. This result is in good agreement with transmission electron microscopy characterization (Figure 3a,b), which shows that the obtained nanoparticles are well crystalized and have a mean size of 22.7 ± 1.0 nm.

To obtain hybrid nanoparticles based on nickel ferrite and gold, we used L-serine, L-tryptophan, glycine, and L-methionine [4] as reducer agents of gold(III) species. In a typical experiment, 20 mL of 0.1 M amino acid solution was added to 25 mg of nickel ferrite and treated with ultrasound for 30 min, and then 0.3 mL of HAuCl₄ solution with a concentration of 0.1 M was added, adjusted to pH ≈ 11, and stirred on a mechanical stirrer for 4 h at 37 °C. The obtained hybrid particles were separated by magnetic separation, washed thoroughly with distilled water, and dispersed in ethanol or in water for characterization.

The solutions remaining after the separation of the NiFe₂O₄/Au particles were studied by a spectrophotometer in the wavelength range of 400–700 nm. In the case of L-serine, L-tryptophan, and L-methionine (Figure 4a,b,d), only tiny amounts of gold were found in the solution (the shoulder at 540 nm for L-tryptophan and 580 nm for L-methionine are observed in the spectra). This indicates that the reduction of gold does not take place in the volume of the solution, but on the surface of the nickel ferrite. On the other hand, a well-defined SPR peak at 550 nm is observed in the spectrum of the solution after the separation of the nanoparticles obtained with glycine (Figure 4c). Therefore, it can be assumed that glycine reduces the Au(III) both in the solution and on the surface of the nickel ferrite.

The surface composition and chemical state of the elements in the NiFe₂O₄/Au samples were investigated by X-ray photoelectron spectroscopy. The wide XPS spectra of all samples contain peaks of Au 4f, O 1s, N 1s, C 1s, Fe 2p, and Ni 2p (Figure 5). The surface composition of the hybrid nanoparticles obtained with L-serine, L-tryptophan, and glycine is summarized in

Table 5. Surface composition of NiFe₂O₄/Au NPs after gold deposition by the amino acids derived from XPS.

Element	Relative Concentrations (At. %)
	Amino Acid
	L-Serine	L-Tryptophan	Glycine	L-Methionine
Au	0.9	0.7	<0.1	2.9
Ni	1.9	0.7	0.9	0.9
Fe	4.5	1.5	2.5	1.9
O	44.4	27.5	15.0	30.5
N	2.2	2.1	-	0.4
C	46.1	67.5	81.5	61.8
S	-	-	-	1.6
Au/Ni	0.5	1	<0.1	3.2
Ni/Fe	0.4	0.5	0.4	0.5

. The results obtained earlier for L-methionine are also given there for comparison. The high carbon concentration in the samples is due to the sample substrate used (pyrolytic graphite), adsorbed ligands, and adventitious carbon-containing impurities. In addition, in the spectra of all samples, excluding the sample obtained with glycine, the peak of the N 1s line was observed. This indicates that L-serine, L-tryptophan, and L-methionine were adsorbed on the surface of nickel ferrite nanoparticles. In addition, the spectrum of the sample with L-methionine contains a S 2p peak, and the nitrogen concentration was reduced. As shown previously [4,46], not only L-methionine but also its products of oxidation by Au(III) that do not contain nitrogen atoms were fixed on the sample surface.

It can be seen that the gold content on the surface of nickel ferrite depends on the amino acid used. Its amount on the surface of the sample obtained by L-methionine was maximal: the Au/Ni atomic ratio was 3. We believe that this is due to the effective fixation of gold on the surface of the adsorbed amino acid through the stable S-Au dative bond [4]. Glycine was not efficiently adsorbed on the surface of nickel ferrite, so the surface concentration of gold in this sample was low (Au/Ni < 0.1).

In order to gain a better insight into the chemical state of the gold at the surface of NiFe₂O₄/Au NPs depending on the amino acid used, the high-resolution XP spectra for Au 4f were recorded (Figure 6). The Au 4f_7/2,5/2 spectra of the nanoparticles synthesized by glycine or L-serine gold reduction are the same as those of elemental gold (

Table 6. Relative concentrations of Au species (%) and binding energies of the Au 4f_7/2 peaks derived from XPS.

Amino Acid	Au0		Au(I)		Ausupp.
	BE (eV)	%	BE (eV)	%	BE (eV)	%
Glycine	83.93	100	-	-	-	-
L-Tryptophan	84.16	57.79	85.08	24.43	83.29	17.78
L-Serine	84.10	100.00	-	-	-	-
L-Methionine	84.15	71.25	85.27	24.33	83.10	4.42

, for the sake of simplicity, the decomposition of solely the Au 4f7/2 peak is presented in Table 6, as it is generally accepted in the relevant literature). The spectra of NiFe₂O₄/Au nanoparticles obtained with tryptophan and L-methionine are similar and can be decomposed into three components, with the binding energies given in Table 6. The major doublet with the binding energy of the Au 4f_7/2 peak of 84.16 and 84.15 eV corresponds to Au⁰; the component at 85.08 and 85.27 eV is due to Au(I) species. The presence of Au (I) spectral lines can be explained by the adsorption of [AuCl₂]⁻ on the surface of the nanoparticles [4]. The third doublet exhibits the Au 4f_7/2 binding energy of 83.10 eV and 83.29 eV (Au_supp.) for L-methionine- and tryptophan-derived NPs, respectively. Such a significant negative shift may be surprising at first, but it is likely associated with electron transfer from the support to the Au particles [47,48]. This phenomenon has been previously observed in our earlier study [4]. Such an interpretation of XPS data agrees with conclusions from the papers [49,50]: an excess of oxygen atoms is observed on the surface of NiFe₂O₄ due to the presence of oxygen vacancies in the lattice. The gold atoms are adsorbed on these sites, and the charge is transferred from electron-rich O atoms to Au atoms. Such electronic transitions are of particular importance in photocatalytic applications of the NiFe₂O₄/Au hybrid nanoparticles, as they increase the lifetime of photogenerated charges: electrons and holes [51].

The oxygen vacancies in the lattice of nickel ferrite nanoparticles and electron density transferring from the support to the Au particles are confirmed by the high-resolution XPS spectra O 1s (Figure 7). The shape of O 1s lines in all spectra is complex and consists of several peaks between 530 and 534 eV (from two to four components depending on the amino acid used). The binding energies of the components are given in

Table 7. Relative concentrations of O species (%) and binding energies of the O 1s peaks derived from XPS.

Amino Acid	O I		O II		O III		O IV
	BE (eV)	%	BE (eV)	%	BE (eV)	%	BE (eV)	%
Glycine	530.17	65.05	532.08	23.01	-	-	533.63	11.93
L-Tryptophan	531.20	76.22	-	-	532.81	23.78	-	-
L-Serine	530.26	38.04	532.30	61.96	-	-	-	-
L-Methionine	530.84	13.52	532.02	33.04	532.96	34.32	533.92	19.13

. A peak corresponding to the lattice oxygen in the Ni/Fe-O framework was observed at 530.4 eV [50], which is close to the binding energy of the OI component in the O 1s spectra of NPs obtained with glycine and L-serine. However, in the case of L-methionine and L-tryptophan, a positive shift of about 0.44 eV and 0.80 eV in the binding energy for the O I peak, respectively, due to a more positive charge for O atoms, indicates a charge transfer from O atoms to Au atoms. In [52,53,54], it was noted that the higher binding energy of the O 1s peak (around 531.3 eV) is attributed to oxygen defects in the matrix of metal oxides, related to oxygen vacancies. Anionic vacancies alter the net electronic charge density. This non-lattice oxygen peak was attributed to surface O-ions with lower electron density. Thus, the reduction of gold by L-tryptophan and L-methionine leads to a strong coupling via the -O band between NiFe₂O₄ and gold in the hybrid nanoparticles. Such a phenomenon is not observed in NiFe₂O₄/Au nanoparticles obtained by glycine and L-serine.

The peaks at 532.0–532.2 (O II) and 532.8–533. 0 eV (O III) can be attributed to C-OH and C-O- bonds in non-deprotonated -COOH and deprotonated -COO groups of amino acids [55]. The highest binding energy components at BE = 533.6–533.9 eV (O IV) are those due to adsorbed water (these components are observed in the spectra of NPs prepared using L-methionine and glycine) [56,57]. Although additional information is necessary to assign the peaks clearly, we speculate that these findings could indicate binding interactions between NiFe₂O₄ and amino acids via the O atom of the COOH group.

This claim is supported by the analysis of the FTIR data. Figure 8 shows the FTIR spectra of nickel ferrite nanoparticles obtained with glycine (curve 1), L-serine (curve 2), and L-methionine (curve 3). The spectra of all samples are similar and contain absorption bands at 596, 598, and 600 cm⁻¹, which are characteristic of nickel ferrite and correspond to the vibrations of the M-O bond. The broad peaks at 3400 cm⁻¹ are due to vibrations of OH groups of water molecules adsorbed on the surface of NiFe₂O₄ nanoparticles. In the range of 3000–2800 cm⁻¹, small maxima corresponding to CH₂ groups (symmetrical and asymmetrical C-H stretching) are observed. In the range of 1650–1000 cm⁻¹, some weak lines corresponding to NH₂ scissoring and/or H₂O scissoring vibrations (1618–1622 cm⁻¹), asymmetric and symmetric stretching vibrations of COO- groups (1550–1581 and 1385 cm−1), NH₂ twist + CH₂ twist (1342–1355 cm⁻¹), C-OH rocking (1119–1138 cm⁻¹) and C-N asymmetric stretching (1038–1049 cm⁻¹) of amino acids were observed [58,59,60,61,62]. The band positions of νs(COO-)and νas(CN) are shifted by 10–20 cm⁻¹ due to the interaction of COO- and NH₂ groups of adsorbed amino acids with the surface of nanoparticles. [24,63,64]. As predicted by DFT calculations [65,66], gold nanoparticles attach to amino acid molecules on the surface of NiFe₂O₄ NPs through the most energetically stable N-Au anchoring dative bond.

Figure 9 shows the electron diffraction patterns and TEM images of the hybrid materials, and also the particle size distribution diagrams for Au⁰ deposited on the NiFe₂O₄ surface. Diffraction rings of both nickel ferrite and gold are observed in the electron microdiffraction patterns of all samples.

It can be seen that the size of the gold nanoparticles and their distribution on the surface of the material is dependent on the amino acid that was used. In the case of glycine, the formation of separated gold particles with a median size of 44 nm can be clearly seen. Au crystals have a higher contrast in the TEM image compared to the nickel ferrite NPs. According to the scanning transmission electron microscopy (STEM) data (Figure 10), the gold atoms in this sample are located at different surface areas than the Fe and Ni atoms. In the product obtained with L-tryptophan (Figure 9h), the formation of large gold particles with a median size of 204 nm was also observed. However, the corresponding elemental mapping images (Figure 11) show that the Au can be detected on the entire nanoparticle, with the Au signals located in the same spatial region as the Fe and Ni signals. This indicates the growth of a continuous gold shell on the surface of the numerous agglomerated nickel ferrite crystals. Previously [4], we observed the formation of the gold shell on the surface of NiFe₂O₄ NP aggregates in a three-step L-methionine-assisted gold deposition.

The reduction of gold with L-serine (Figure 9e) leads to the formation of both small gold seeds with an average size of 5.7 ± 0.4 nm and some larger gold particles with a size of 53.6 ± 4.9 nm. This behavior results in the formation of a bimodal size distribution (Figure 9f has two PSD diagrams (the second one is in the inset) for fine and large gold particles). Large gold particles are probably formed as a result of the agglomeration of small particles, which may be due to insufficient stabilization by adsorbed amino acid or its oxidation products.

Under these reaction conditions, L-methionine (Figure 9k) forms only tiny gold seeds, much smaller than those produced by L-serine (1.46 ± 0.07 nm). In addition, the density of Au⁰ seeds on the surface of this sample is much higher than that of the sample prepared with L-serine: 26 versus 3 gold NPs per 400 nm².

Thus, by using different amino acids and varying the number of gold reduction steps, hybrid NiFe₂O₄/Au nanoparticles can be obtained with a specific gold grain size and surface density (inlaid or core–shell structure) for a particular application. Gold is firmly attached to the NP surface and does not come off during washing, redispersion in solvents, or ultrasonic treatment.

The chemical mechanism of amino acid oxidation by [AuCl₄]⁻ ions depends on the amino acid used and has not been studied enough. Most papers describe the oxidation of sulfur-containing amino acids [27,28,29,30,31]. In particular, in the case of L-methionine, the intermediate complex [AuCl₂(Met)]⁺ is formed, which reacts with the second L-methionine molecule to generate methionine sulfoxide CH₃S(O)CH₂CH₂CH(NH₂)CO₂H and [AuCl₂]⁻. [AuCl₂]⁻ undergoes disproportionation, forming [AuCl₄]⁻ and Au⁰ [30].

In [36], the following mechanism of Au(III) reduction by glycine (at pH = 4.5) is proposed: Stage 1 (slow) involves the nucleophilic attack of -COO⁻ to Au(III) species with the formation of an iminic cation, CO₂, and [AuCl₂]⁻:

[AuCl₄]⁻ + ⁺H₃NCH₂COO⁻ → H₂C = NH₂⁺ + CO₂ + [AuCl₂]⁻ + H⁺ + 2Cl⁻.

(12)

In Stage 2 (fast), the iminic cation undergoes hydrolysis to produce formaldehyde and an ammonium ion.

H₂C = NH₂⁺ → HCHO + NH₄⁺.

(13)

In [67], it was shown by NMR spectroscopy that at pH = 2.4, the oxidation of glycine leads to glyoxylic acid, formic acid, and CO₂. Further, glyoxylic acid is oxidized to HCOOH by excess gold(III). The authors of [36,67] did not observe a complete reduction of gold to the metallic state under their experimental conditions. The disproportionation reaction

3[AuCl₂]⁻ = 2Au⁰ + [AuCl₄]⁻ +2Cl⁻

(14)

is facilitated by an alkaline condition [68].

Increasing the number of carbon atoms on the hydrocarbon chain of aliphatic amino acids, as well as the use of aromatic amino acids, generally does not significantly change the mechanism of their oxidation [34]. For example, in the case of the oxidation of L-a-amino-n-butyric acid by permanganate, the same iminic cation (CH₃CH₂CH=NH₂) was formed, which underwent fast hydrolysis to give CH₃CH₂CHO as the only organic product [69], and the oxidation of DL-alanine, b-phenylalanine, and DL-leucine produces NH₄⁺, CO₂, and the corresponding aldehydes [70,71,72].

Tight binding between Au⁰ seeds and the surface of NiFe₂O₄ in the case of L-methionine, confirmed by our XPS data, can be explained by the fact that the reduced gold atoms still remain in the coordination sphere of L-methionine sulfur attached to the surface of ferrite NPs and stabilize gold NPs [24]. Glycine-assisted Au(III) reduction produces only low-molecular-weight products that are not fixed to the nickel ferrite surface and do not interfere with the binding of NiFe₂O₄ and Au NPs, resulting in a low gold concentration and separate gold crystal formation. As the length of the carbon chain of the amino acids and the number of functional groups in its molecules increase, the products of its oxidation are more efficiently adsorbed on the surface of nickel ferrite, as well as on the formed Au NPs. This has been demonstrated for the polar amino acid L-serine, which has an OH group in the side carbon chain, and especially for L-tryptophan, which contains an aromatic indole nucleus. The favorable interactions of the indole group with other molecular partners are well documented. In addition, SERS data show that L-tryptophan is tightly bound to the surface of gold NPs [73].

4. Conclusions

In this paper, the effect of the reaction parameters on the reduction of [AuCl₄]⁻ ions by glycine, L-serine, L-tryptophan, and L-tyrosine on the yield of Au nanoparticles in solutions was determined. It was determined that L-serine and L-tryptophan are the most effective reducing agents for [AuCl₄]⁻ ions under the investigated conditions, while glycine and L-tyrosine have weaker reducing properties. The found optimal conditions of Au (III) reduction have been used to prepare NiFe₂O₄/Au hybrid material in one-pot gold deposition by non-sulfur amino acids.

The XP spectra of gold-coated nickel ferrite nanoparticles indicate that L-serine and L-tryptophan were efficiently adsorbed on the surface of nickel ferrite nanoparticles while glycine was not adsorbed. The gold concentration on the NiFe₂O₄ surface depends on the amino acid used and increases in the order Gly << Ser ≈ Tryp < Met (taken for comparison). The Au 4f and O 1s spectra demonstrate the charge transfer from electron-rich O atoms of the NiFe₂O₄ surface to Au atoms in the case of tryptophan, indicating a tight binding between Au⁰ seeds and the magnetic core.

The TEM data show the formation of only separated gold particles with a median size of 44 nm in the case of the glycine-assisted sample. In Au⁰ deposition with serine, both small gold seeds with an average size of 5.7 ± 0.4 nm and some larger gold particles with a size of 53.6 ± 4.9 nm are formed. In the product obtained with L-tryptophan, the formation of a continuous gold shell on the surface of the numerous agglomerated nickel ferrite crystals is observed. According to the FTIR spectroscopy data, the interaction between the surface of NiFe₂O₄ NPs and the adsorbed amino acids occurs through the O atom of the COOH group. The gold nanoparticles attach to the adsorbed amino acid molecules via N-Au anchoring and/or O-Au (in the case of side-chain OH groups) dative bonds.

Thus, by using different amino acids and varying the number of gold reduction steps, hybrid NiFe₂O₄/Au nanoparticles with a specific gold grain size and surface density (inlaid or core–shell structure) can be obtained for a particular application.