The Influence of Temperature on the Spatial Distribution of AuNPs on a Ceramic Substrate for Biosensing Applications

Abstract

In this study, we investigated the spatial distribution and homogeneity of gold nanoparticles (AuNPs) on an alumina (Al₂O₃; AAO) substrate for potential application as surface-enhanced Raman scattering (SERS) sensors. The AuNPs were synthesized through thermal treatment at 450 °C at varying times (5, 15, 30, and 60 min), and their distribution was characterized using field-emission scanning electron microscopy (FE-SEM) and scanning transmission electron microscopy (STEM). The FE-SEM and STEM analyses revealed that the size and interparticle distance of the AuNPs were significantly influenced by the duration of thermal treatment, with shorter times promoting smaller and more closely spaced nanoparticles, and longer times resulting in larger and more dispersed particles. Raman spectroscopy, using Rhodamine 6G (R6G) as a probe molecule, was employed to evaluate the SERS enhancement provided by the AuNPs on the AAO substrate. Raman mapping (5 µm × 5 µm) was conducted on five sections of each sample, demonstrating improved homogeneity in the SERS effect across the substrate. The topological features of the AuNPs before and after R6G incubation were analyzed using atomic force microscopy (AFM), confirming the correlation between a decrease in surface roughness and an increase in R6G adsorption. The reproducibility of the SERS effect was quantified using the maximum intensity deviation (D), which was found to be below 20% for all samples, indicating good reproducibility. Among the tested conditions, the sample synthesized for 15 min exhibited the most favorable characteristics, with the smallest average nanoparticle size and interparticle distance, as well as the most consistent SERS enhancement. These findings suggest that AuNPs on AAO substrates, particularly those synthesized under the optimized condition of 15 min at 450 °C, are promising candidates for use in SERS-based sensors for detecting cancer biomarkers. This could be attributed to temperature propagation promoted at the time of synthesis. The results also provide insights into the influence of thermal treatment on the spatial distribution of AuNPs and their subsequent impact on SERS performance.

1. Introduction

Cancer, a leading cause of death worldwide, accounted for approximately 9.7 million deaths in 2022, with an estimated 20 million new cases reported globally each year [1]. Thus, the monitoring and early detection of diseases such as cancer are of great importance in improving patient outcomes and survival rates [2]. Early diagnosis significantly enhances treatment efficacy and survival rates, highlighting the critical need for advanced diagnostic tools [3]. Traditional diagnostic methods, while effective, often involve invasive procedures that can be uncomfortable and risky for patients. Consequently, there is a growing demand for non-invasive sensors that can detect disease biomarkers with a high sensitivity and specificity. Non-invasive sensors offer numerous advantages, including reduced patient discomfort, lower risk of complication, and the ability to perform frequent monitoring [4].

Among the emerging technologies, surface-enhanced Raman spectroscopy (SERS) stands out for its potential to revolutionize disease diagnostics. SERS, a powerful technique in Raman spectroscopy, leverages the plasmonic properties of metallic nanoparticles to amplify Raman signals from molecules adsorbed on their surface [5]. This amplification allows for the detection of biomolecules at extremely low concentrations, making SERS an ideal candidate for non-invasive diagnostic applications. The recent advancements in plasmonic sensors, particularly those based on SERS, have opened new avenues in health monitoring [6]. These sensors typically utilize noble metal nanoparticles, such as gold nanoparticles (AuNPs), to enhance the Raman scattering of target molecules. The unique optical properties of AuNPs, including their strong surface plasmon resonance, make them highly effective in amplifying Raman signals [7,8]. This capability has led to the development of a new generation of SERS-based plasmonic sensors that can detect cancer biomarkers, pathogens, and other disease indicators with remarkable precision. In addition, gold nanoparticles have become a focal point in SERS research due to their excellent biocompatibility, stability, and tunable optical properties [9]. The controlled synthesis and functionalization of AuNPs are crucial for optimizing their performance in SERS applications. Research groups are particularly interested in understanding the growth mechanisms of AuNPs, as this knowledge is essential for developing reliable and reproducible SERS substrates [10]. The formation process of AuNPs significantly influences their size, shape, and surface properties, which in turn affect their plasmonic behavior and SERS performance. The development of SERS supports based on metallic nanoparticles supported on ceramic substrates is a promising trend in this field [5,11]. Ceramic materials, such as alumina (Al₂O₃), provide a stable and inert platform for anchoring AuNPs, enhancing their thermal and chemical stability. These hybrid substrates combine the advantageous properties of ceramics and noble metals, resulting in robust and efficient SERS platforms [12]. The integration of AuNPs with ceramic support has shown great potential in various applications, including environmental monitoring [13], food safety [14], and medical diagnostics [15]. Understanding the implications of the growing mechanism in the thermal-based fabrication of SERS substrates is vital for advancing this technology. The temperature at which AuNPs are synthesized can significantly impact their nucleation and growth processes, leading to variations in particle size, distribution, and morphology [16]. These parameters are critical in determining the plasmonic properties and SERS performance of the substrates. By comprehensively studying the thermal effects on AuNP growth, researchers can optimize fabrication protocols to achieve consistent and high-quality SERS substrates. Despite these promising advancements, several challenges remain in the implementation of SERS-based sensors. One of the primary hurdles is the reproducibility of AuNP formation processes. Variation in synthesis conditions can lead to significant differences in the optical properties and SERS activity of the nanoparticles [16,17,18,19]. Achieving uniform and reproducible AuNPs is essential for the widespread adoption of SERS technology in clinical and industrial settings. Additionally, the stability and robustness of SERS substrates under real-world conditions are critical factors that need to be addressed to ensure reliable performance over extended periods [20]. Recent literature highlights a growing interest in the development of SERS substrates, with various architectures being explored, such as heterostructures [21], nanoparticles on nanowire arrays [22], three-dimensional nanoporous structures [23], mesoscopic star-shaped particles [24], and nano-garland arrays [25]. All of the aforementioned SERS substrate architectures are gold-based, and the literature focuses on the following two main aspects: (1) increasing the Raman enhancement factor and (2) achieving a greater homogeneity in the Raman enhancement across the entire sample. Recent studies have proposed the use of Rhodamine-functionalized nanoparticles for identifying carcinogenic cells and biomarkers [26,27,28]. Consequently, achieving the detection of Rhodamine via SERS substrates could pave the way for the development of biosensing technology [29].

In this work, the spatial distribution of AuNPs on a ceramic substrate was controlled by varying the length of time of thermal treatment on a gold film. The time was varied as a parameter to control the temperature propagation on the sample during thermal treatment. The SERS enhancements resulting from the different distributions of AuNPs on the ceramic were studied to explore the viability of employing the proposed material as a SERS substrate for breast cancer biomarkers.

2. Materials and Methods

2.1. Chemicals and Materials

For the development of this research, the following commercially available materials were obtained: sodium hydroxide (NaOH, SKU: 306576-100G), Rhodamine 6G (R6G, C₂₈H₃₁N₂O₃Cl, SKU: 56226-25MG), aluminum foil (Al, SKU: 326860-3.6G), oxalic acid (C₂H₂O₄, SKU: 247537-500G), perchloric acid (HClO₄, SKU: 311421-250ML), phosphoric acid (H₃PO₄, SKU: 695017-500ML), and deionized water (H₂O, SKU: W4502-1L, 18 MΩ). All chemical reagents were commercially acquired from SIGMA-ALDRICH and used as received without any additional purification.

2.2. The Synthesis of AuNPs on the AAO Substrate

The AAO substrate was synthesized using the two-step anodization method [5]. Briefly, an aluminum sheet was cut into 1 cm × 1 cm squares and mechanically polished to a mirror-like finish. The aluminum was then electrochemically polished in a solution of perchloric acid and ethanol mixed at a 4:1 volume ratio using a potentiostat under a constant voltage of 20 V (with respect to the reference electrode) for 1 min at room temperature. The samples were cleaned by subjecting them to alternating ultrasonic baths of deionized water and methanol for four cycles, each lasting 10 min. After cleaning, the samples were dried using a nitrogen stream. Next, the polished substrate underwent anodization to form alumina on the aluminum surface. This process was carried out using a potentiostat in a 3 M oxalic acid solution, applying a voltage of 20 V (with respect to the reference electrode) for 2 h at 4 °C.

The initial alumina layer was removed by immersing the samples in a 0.5 M phosphoric acid solution for 20 min. A second anodization was then performed under the same conditions as the first, but extended over a period of 6 h. After completing this process, the samples were cleaned using six alternating ultrasonic baths of deionized water and methanol, each lasting 15 min, and were subsequently dried using a nitrogen stream. After obtaining the AAO substrate, a gold layer was deposited as a precursor for nanoparticle formation. The gold layer was deposited using sputtering coater equipment, employing an ionic current density of 16 mA/cm² and a vacuum of 5 Pa. Following gold deposition, an isothermal treatment was carried out in a horizontal furnace with a nitrogen flow (30 L/min) at 450 °C. Four duration times were explored: (a) 5 min, (b) 15 min, (c) 30 min, and (d) 60 min. Based on the equipment calibration, the Au layer deposited had a thickness of 90 nm. Following deposition, the substrates underwent a cleaning process involving four ultrasonic bath cycles, alternating between methanol for 10 min and deionized water for 15 min. The samples were then dried using a nitrogen stream. Figure 1 summarizes the methodology followed to obtain the AuNPs on the AAO.

2.3. Physical Characterization of the Samples

The samples were characterized by field emission scanning electron microscopy (FE-SEM) using an AURIGA microscope equipped with a secondary electron (SE) detector (Zeiss, Jena, Germany). Images were acquired at an accelerating voltage of 2 keV and a working distance of 8 mm. For a detailed study of the gold nanostructures on the alumina substrate surface, images were obtained using scanning transmission electron microscopy (STEM) in a JEOL ARM200F (JEOL, Tokyo, Japan). Sixty images covering a total area of 10 µm × 10 µm were taken for each sample to obtain the statistical analysis for diameter size and interparticle distance. Prior to imaging, samples were prepared using a focused ion beam (FIB) system (JIB-4500 MultiBeam, JEOL, Tokyo, Japan) with a gallium liquid metal ion source (Ga). Atomic force microscopy (AFM) images were recorded with a microscope from SOL instruments (NT-MDT model) (NT-MDT spectrum instruments, Moscow, Russia). The samples were measured by AFM at 4 °C with an acquisition time of 10 min to record a 5 µm × 5 µm area for each sample. To obtain the particle size and interparticle distance, DigitalMicrograph software (version number 1.0) was used. A sample of 5000 particles were randomly analyzed through the acquisition of electron microscopy images. The measurements were then computed using image processing techniques included in the software, as has been conducted in other research work [19].

2.4. Spatial Distribution Analysis for Biosensing Applications

The spatial distribution of AuNPs under the four different conditions was analyzed using Raman spectroscopy (LabRam HR800, Horiba Jovin Yvon spectrometer with a 633 nm excitation source) (Horiba, Kyoto, Japan). Rhodamine 6G (R6G) was used as the probe molecule due to its well-documented Raman response in SERS substrates [5]. In addition to analyzing the distribution of AuNPs, Raman mapping was performed on the substrate to assess the uniformity of signal enhancement across the proposed SERS substrate (AuNPs on AAO).

3. Results and Discussion

The anodized aluminum oxide (AAO) achieved was cylinder-like and porous, with a diameter of 102 nm ± 11 nm and an average interpore distance of 145 nm ± 15 nm, as shown in the frontal view (Figure 2a) and lateral view (Figure 2b) of the AAO images recorded by FE-SEM. Once the AuNPs on the AAO substrate were obtained, top-view images were acquired using FE-SEM at the same magnification (100,000×) to compare the distribution of nanoparticles on the alumina substrate (see Figure 3). At 450 °C for 5 min, the resulting AuNPs exhibited a spherical shape with relatively homogeneous distribution across the AAO substrate (see Figure 3a). When the exposure time was increased to 15 min, the nanoparticles decreased in diameter and the distance between them also reduced (see Figure 3b). With a further increase in time to 30 min at 450 °C, the AuNPs dramatically increased in their average diameter and the distance between them also increased (see Figure 3c). Finally, when the exposure time exceeded 60 min, the AuNPs lost their regular shape and formed larger gold particles, seemingly due to the accumulation of AuNPs (see Figure 3d). The distance between these larger gold particles increased, reflecting the loss of their nanoscale size. Note that the darker lines in Figure 3b are the result of charge effects within the scanning electron microscope, which is common in alumina-based sample analysis, as it is an electrical isolator material [30]. It is possible that AAO allows for the special dispersion of AAO due to the temperature propagation during thermal treatment and the surface energy relationship between the gold and the aluminum oxide.

Combined with the FE-SEM analysis, a STEM analysis was conducted to study the spatial distribution of AuNPs on the alumina substrate. Figure 3a shows the AuNPs attached to the alumina pre-processed by FIB and obtained under a synthesis time of 15 min, where it can be seen that chemical bonding allows for the attachment of the AuNPs to the support [31]. A detailed image of the highlighted section in Figure 4a is shown in Figure 4b, where a clear difference in contrast indicates the phase difference between the nanoparticles and the substrate, attributed to gold and aluminum oxide, respectively [32,33]. Additionally, as metrics for the distribution of nanoparticles on the AAO substrate, two magnitudes were considered: (1) the diameter of the AuNPs, and (2) the interparticle distance between the AuNPs. Thus, Figure 4c compares the average diameter size measured for each condition, showing values of 50 nm ± 5.9 nm, 27 nm ± 2.6 nm, 150 nm ± 60.1 nm, and 250 nm ± 80.6 nm for the samples synthesized under conditions 1 (5 min), 2 (15 min), 3 (30 min), and 4 (60 min), respectively. On the other hand, the average interparticle distance was experimentally determined (see Figure 4d) to be 21 nm ± 2.3 nm, 9.0 nm ± 0.8 nm, 60.1 nm ± 6.4 nm, and 95.3 nm ± 11.2 nm for synthesis conditions 1, 2, 3, and 4, respectively. These results suggest that temperature has a significant effect on the spatial distribution of AuNPs on a ceramic AAO substrate. A 15-min exposure promotes the smallest nanoparticle size diameter and the smallest interparticle distance, while a 60-min exposure results in the largest size for both the average diameter and the interparticle distance. This behavior can be explained according to classical nucleation theory and the growth phase in crystals, promoted by temperature, as reported in other systems [16,34].

To determine the homogeneity offered by the substrates as potential sensors assisted by Raman spectroscopy, a 1 × 10⁻¹⁰ M Rhodamine 6G (R6G) solution was applied to interact with the synthesized substrates (AuNPs on alumina). A comparison of the lateral view of the interface between the alumina and AuNPs, imaged by recording backscattered electrons (BSE; Figure 5a), and secondary electrons (SE; Figure 5b), in the FE-SEM of samples incubated with the R6G solution, suggests that the spaces generated by the AuNPs and the porosity in the alumina play a key role in R6G adsorption on the surface (as depicted in the scheme shown in Figure 5c). This is supported by the fact that the AuNP accumulation appears thinner in the BSE image compared to the SE image, as organic materials do not provide sufficient signals in the BSE image due to their lower density and atomic number compared to the AuNPs [35,36]. Consequently, a SERS mapping was constructed from the R6G measured on all synthesized substrates. Raman spectroscopy was performed on the sample surface with 10 accumulative measurements and a recording time of 10 s each. The measurements were then repeated at 100 points on the sample to build a matrix of 100 pixels, where the intensity of the peak at 1361 cm⁻¹ was used as the parameter to assign a color representing the magnitude. Figure 5e shows the resulting image of this adaptation, corresponding to the Raman mapping of the AuNPs on the AAO substrate, obtained at 450 °C for 15 min. The R6G can be adsorbed by the AuNPs and the AAO, as shown in Figure 5a,b. Thus, the variation observed in Figure 5d,e could be influenced by the spatial distribution of AuNPs on the AAO.

Figure 5d shows a Raman mapping of AuNPs on an alumina support loaded with R6G; this image allows us to see the difference between the maximum intensity and the minimum intensity given by the SERS substrate developed in this work. Figure 5f compares the Raman spectra between the points with the highest and lowest intensities in the Raman mapping. The primary difference between these spectra is the noise level in the signal, while the peaks observed are consistent with those typically exhibited by R6G [5,37], as shown in

Table 1. Vibrational moods attributed to Rhodamine 6G (in Figure 5f).

Peak Position in Figure 5f	Vibrational Mode
769 cm−1	C–H out-of-plane bending
1184 cm−1	C–H in-of-plane bending
1311 cm−1	C–O–C stretching in COOC2H5 group
1358 cm−1	C–C stretching in xanthene ring
1574 cm−1	C–C stretching in phenyl ring
1645 cm−1	C–C stretching in xanthene ring

. A common issue with substrates used in SERS-based sensors is the homogeneity of the enhancement effect over a large area of the sample. To evaluate the improvement in homogeneity of the Raman scattering, a large Raman mapping (5 µm × 5 µm area) was conducted on the synthesized samples with R6G in five sections of the substrate, with each section separated by at least 4 mm. In addition, the enhancement factor was computed for each sample. Thus, the known methodology for computing the EF was employed [5]. The minimum concentration detected for each sample was employed to compute the EF, 1 × 10⁻¹⁰ M, 1 × 10⁻⁹ M, 1 × 10⁻⁷ M, and 1 × 10⁻⁶ M, in the samples treated for 5 min, 15 min, 30 min, and 60 min, respectively. A 1 µL drop was applied to the AAO/AuNPs and allowed to dry prior to measurement. To assess the intensity from the R6G alone, 1 × 10⁻¹¹ g of R6G was analyzed in its solid state. Based on the probe diameter (10 µm), the estimated number of irradiated R6G molecules was 1.257 × 10¹¹ in the lyophilized sample, 1.3257 × 10⁵ for the 1 × 10⁻¹⁰ M, 1.1134 × 10⁶ in the 1 × 10⁻⁹ M, dilution, 2.5344 × 10⁷ for the 1 × 10⁻⁸ M, 3.4615 × 10⁸ for the 1 × 10⁻⁷ M, and 2.6522 × 10⁹ in the 1 × 10⁻⁶ M dilution. Raman spectra were subsequently recorded for all samples. The peak intensity corresponding to the vibrational mode at 1184 cm⁻¹ was evaluated, yielding intensity values. The solid R6G sample recorded an intensity value of 184.35. Then, the relationship between Raman intensity due to the SERS effect and the R6G sole was computed to determine the EF. The computation was repeated randomly at ten different points on the sample, recording both the minimum and maximum Raman intensity values for each sample, resulting in a range of minimum and maximum values of 0.1 × 10⁵–0.5 × 10⁵, 0.9 × 10⁶–1.1 × 10⁶, 5.2 × 10⁴–6.8 × 10⁴, and 2.3 × 10²–4.5 × 10³ for the samples, synthesized at 5 min, 15 min, 30 min, and 60 min, respectively.

Figure 6 shows two images obtained by AFM which show a 5 micrometer by 5 micrometer area of the AuNPs on the alumina substrate, particularly those synthesized at 15 min of thermal treatment. The topological features of the AuNPs on a single alumina substrate are displayed in Figure 6a, where the AuNPs on the alumina are distinguishable. The size of AuNPs is in good agreement with the FE-SEM and STEM results. After incubating the sample with the R6G solution for 60 min, the roughness of the sample decreases from 0.16 to 0.08, as seen in Figure 6b. Note that the mentioned difference can be better appreciated by comparing the roughness profile of the selected area shown in the inserts of Figure 6a,b. Note that sample becomes less rough because the R6G gets trapped on the surface. By comparing the Raman mapping with the topological characteristics of the sample, both before and after R6G incubation, the intensity variation of each point constructing the image exhibits only a slight average change compared to Figure 6a,b. This could be attributed to the RG6 solution concentration, which was 1 × 10⁻¹⁰ M, the electrostatic forces in the sample that promote the physisorption of R6G, and the solvent used in the solution (deionized water).

On the other hand, the recorded Raman mapping of the same sample exhibited in Figure 6b is shown in Figure 7. Note that images shown in Figure 6 and Figure 7 correspond to the same area in the sample. It can be observed that the roughness profile shown (insets in Figure 7) indicate that the sample has a homogeneous SERS effect across the entire area. The maximum intensity deviation (D) is a parameter that provides information on the reproducibility of the SERS effect in active SERS substrates. For the samples, this parameter was calculated using Equation (1), where the difference between the maximum peak intensity and the average peak intensity (∆I) of a certain number of samples is divided by the average peak intensity ($\overline{I}$) and multiplied by 100 to obtain the percentage value of D.

$$D={\displaystyle \frac{∆I}{\overline{I}}}·100\%$$

(1)

The intensity of fifteen points at different places on the same sample were measured to calculate the statistics (

Table 2. A comparison between the characteristic values for computing the maximum intensity deviation of the synthesized substrates.

Sample	$\overline{\mathit{I}}$	∆I	D (%)
1	140.18	24.59	17.54
2	408.2	74.96	18.36
3	191.46	30.39	15.87
4	57.88	11.31	19.54

). According to M. J. Natan [38], the values of maximum intensity deviation should be below 20%. Based on the calculated values, all samples meet these criteria, indicating that even though all the samples are suitable for application as SERS substrates, the sample synthesized in condition 2 is the best candidate for this application.

To evaluate the influence of a ceramic substrate on the performance of SERS substrates for biosensing applications, the process of SERS substrate formation was repeated using a (111) silicon wafer as a reference. Figure 8 presents FE-SEM images of samples treated for 5, 15, 30, and 60 min. It is evident that the formation of AuNPs is not homogeneous under any condition, with a broad size distribution observed in all cases. This indicates that the ceramic substrate plays a crucial role in the spatial distribution of AuNPs, particularly during the thermal treatment process. These findings suggest that ceramic substrates significantly enhance the homogeneity of SERS substrates over large areas, which is essential for obtaining reliable Raman spectroscopy measurements in biosensing applications.

The novelty of this work addresses one of the main challenges in applying SERS substrates to biosensing technology: achieving homogeneity in the SERS effect across the entire substrate. Compared to other studies, the proposed substrate offers a significant contribution by achieving large-area substrates with high reproducibility in SERS amplification and acceptable homogeneity [39,40,41]. Despite these advancements, SERS substrates must combine homogeneity, accessibility, and enhancement to be commercially viable in the medical sector.

4. Conclusions

This study successfully demonstrates the potential of gold nanoparticles (AuNPs) on alumina (AAO) substrates as effective surface-enhanced Raman scattering (SERS) biosensors due to an increase in the SERS effect’s homogeneity by thermal treatment. Through meticulous control of thermal treatment duration, we were able to significantly influence the spatial distribution and size of AuNPs, where shorter treatments (15 min at 450 °C) yielded smaller and more closely spaced nanoparticles and longer treatments resulted in larger and more dispersed particles. The combined analyses using field-emission scanning electron microscopy (FE-SEM), scanning transmission electron microscopy (STEM), and atomic force microscopy (AFM) confirmed the morphology and topological features of the AuNPs on the alumina substrate. AAO can influence the propagation of temperature and allow for a high special dispersion of AuNPs. The Raman mapping, conducted with Rhodamine 6G (R6G) as a probe molecule, revealed a homogeneous SERS enhancement across the substrate, with the sample synthesized for 15 min exhibiting the most favorable characteristics. This condition provided the smallest average nanoparticle size and interparticle distance, contributing to a more uniform and reproducible SERS signal. The reproducibility of the SERS effect was further validated by the maximum intensity deviation (D), which remained below 20% in all samples, confirming their reliability for sensor applications. These findings underscore the significance of having precise control over synthesis parameters when optimizing the performance of SERS substrates. The optimized AuNPs on the alumina substrate synthesized under conditions of 15 min at 450 °C show great promise as SERS-based sensors. These optimized features were achieved due to temperature distribution during the thermal treatment and were controlled by the duration time of the process.