Substituent‐Dependent SERS Kinetics of Citrate‐Stabilized Gold Nanoparticles: Balancing Hotspot Formation and Colloidal Flocculation

ABSTRACT

Surface‐enhanced Raman scattering (SERS) is among the most sensitive vibrational techniques, but reliable quantification in liquid‐phase colloidal systems remains limited by the coupling between molecular adsorption, hotspot formation, and nanoparticle stability. Here, using citrate‐stabilized gold nanoparticles functionalized with para‐substituted benzenethiols, we show that analyte physicochemical properties determine not only the SERS intensity, but also the accessible measurement window and colloidal regime. Concentration‐dependent and time‐resolved SERS measurements, supported by UV–vis absorption spectroscopy and ζ‐potential analysis, reveal three response regimes. Halogenated thiols preserve colloidal stability and display delayed formation of SERS‐efficient configurations. Partially destabilizing thiols generate faster and stronger responses, but undergo earlier signal attenuation. Strongly destabilizing substituents produce intense but short‐lived signals governed by rapid aggregation and flocculation. The temporal response is described by a semi‐mechanistic model combining SERS‐active‐state formation with its loss through colloidal evolution. These results show that reliable quantitative SERS requires reporter selection based not only on Raman response and surface affinity, but also on the balance between hotspot generation and colloidal persistence.

Reporter chemistry controls the operational window of liquid‐phase SERS. Para‐substituted benzenethiols induce distinct AuNPs colloidal regimes, ranging from delayed but stable hotspot formation to rapid flocculation‐dominated signal decay. The balance between SERS‐active‐state formation and colloidal loss defines whether a reporter provides reliable quantitative performance or only intense but short‐lived signals.

1. Introduction

Surface‐enhanced Raman scattering (SERS) spectroscopy is one of the most sensitive vibrational spectroscopic techniques, enabling detection down to the single‐molecule level [1, 2]. The observed enhancement arises from two main contributions: chemical and electromagnetic enhancement [3], with the latter generally providing the dominant amplification in plasmonic SERS substrates [1, 2, 4]. Electromagnetic enhancement originates from localized surface plasmon resonance (LSPR), where excitation of metallic nanostructures generates highly amplified local electric fields [5, 6]. Because this enhancement decays rapidly with distance from the metal surface, efficient SERS detection requires the analyte to be adsorbed onto, or located in proximity to, the plasmonic interface [7]. Molecules combining strong affinity for noble‐metal surfaces with high Raman scattering cross‐sections are therefore especially suitable as SERS reporters [8, 9]. Benzenethiol and its derivatives are widely used in this context because they bind strongly to gold through Au─S bond formation and exhibit intense aromatic Raman bands [10, 11, 12]. However, despite their extensive use, the relationships between substituent chemistry, adsorption behavior, colloidal stability, and time‐dependent SERS response remain insufficiently understood.

SERS sensing frequently relies on the formation of electromagnetic hot spots, particularly in nanogaps generated by interparticle coupling [13, 14, 15, 16, 17]. Controlled aggregation and the fabrication of three‐dimensional plasmonic architectures are therefore common strategies to increase SERS sensitivity [18, 19, 20, 21, 22, 23]. In liquid‐phase measurements, however, the same interparticle interactions that generate intense hot spots can also compromise quantitative performance. Although colloidal SERS is attractive for quantitative analysis because of its simple handling, homogeneous sampling, and compatibility with solution‐phase assays [24, 25], its response depends on the coupled evolution of molecular adsorption, molecular orientation [26], nanoparticle concentration, interparticle interactions, and colloidal stability. In many cases, the addition of analytes or reporters to citrate‐stabilized nanoparticles partially displaces or screens the stabilizing ligand shell, inducing uncontrolled aggregation, precipitation, or time‐dependent changes in the number of optically active hot spots [27]. Protective strategies, including porous silica shells [28, 29] and metal‐organic frameworks (MOFs) [30, 31, 32], have been developed to improve colloidal stability, but these architectures can also restrict molecular accessibility and modify target–reporter interactions. Thus, the physicochemical properties of the analyte itself, particularly the nature of its functional groups, can be decisive in defining both SERS efficiency and the useful experimental time window.

Although thiophenol adsorption and substituent effects on nanoparticle surface charge and aggregation have been investigated previously, a systematic connection between substituent chemistry, colloidal evolution, temporal SERS response, and quantitative usability in liquid‐phase systems is still lacking. Here, a series of para‐substituted benzenethiols bearing functional groups with distinct electronic and interfacial properties was investigated using citrate‐stabilized gold nanoparticle (AuNPs) suspensions. The SERS response was evaluated as a function of AuNPs concentration and time, while UV–vis absorption spectroscopy and ζ‐potential measurements were used to monitor the accompanying colloidal evolution. The results show that substituent groups control not only the SERS intensity, but also the activation kinetics, colloidal stability, and operational measurement window. The time‐dependent response can be rationalized as the competition between the progressive formation of a SERS‐efficient state and its subsequent loss through aggregation and flocculation. These findings establish reporter selection as a key design parameter for reliable quantitative SERS measurements in liquid‐phase colloidal systems.

2. Methods

2.1. Materials

Hydrogen tetrachloroaurate(III) trihydrate (99.99%, HAuCl₄·3H₂O), 4‐methoxybenzenethiol (98%, BT─OMe), 4‐chlorothiophenol (98%, BT─Cl), 4‐hydroxybenzenethiol (98%, BT─OH), and p‐thiocresol (98%, BT─Me) were purchased from Thermo Scientific. Biphenyl‐4‐thiol (97%, BT─Ø) was purchased from Sigma‐Aldrich. Sodium citrate tribasic dihydrate (≥ 98%, CA) was obtained from Acros Organics. Thiophenol (≥ 98%, BT) was purchased from Fluka, and 4‐fluorothiophenol (97%, BT─F) was obtained from Alfa Aesar. All reagents were used as received, without further purification. Milli‐Q water (18 MΩ·cm) was used for the preparation of all aqueous solutions. Thiol stock solutions were prepared in ethanol. All glassware was cleaned with freshly prepared aqua regia, thoroughly rinsed with deionized water, and finally washed with ethanol before use.

2.2. Synthesis of Citrate‐Stabilized Spherical Gold Nanoparticles

Citrate‐stabilized spherical gold nanoparticles (AuNPs) with an average diameter of approximately 25 nm were synthesized following a modified literature procedure [33]. Briefly, 100 mL of HAuCl₄ solution $(1.4\times {10}^{-3}$ M) was heated to boiling under continuous stirring. Once boiling was reached, 10 mL of freshly prepared CA solution (1% w/v) was added under vigorous stirring. After 15 min, the solution changed from pale yellow/colorless to deep red, indicating nanoparticle formation. Heating was then stopped, and the colloidal suspension was allowed to cool to room temperature. This synthesis yielded spherical AuNPs with an average diameter of 23.43 ± 2.51 nm and a final Au concentration of 1.4 mM. The colloidal suspension was stored in the dark until further use.

2.3. Nanoparticle Characterization

Transmission electron microscopy (TEM, JEOL JEM 1011), operating at an accelerating voltage of 80 kV, was used to determine the size and morphology of the AuNPs. The optical properties of the nanoparticles were characterized by UV–vis absorption spectroscopy using a Thermo Scientific Evolution 201 UV–vis spectrophotometer.

2.4. Concentration‐Dependent Study

A series of 12 AuNPs suspensions was prepared from the stock colloid, covering the following nominal Au concentrations: $1.4\times {10}^{-3}$, $1.2\times {10}^{-3}$, $9.33\times {10}^{-4}$, $7.00\times {10}^{-4}$, $6.00\times {10}^{-4}$, $4.60\times {10}^{-4}$, $3.50\times {10}^{-4}$, $2.33\times {10}^{-4}$, $1.75\times {10}^{-4}$, $8.75\times {10}^{-5}$, $5.80\times {10}^{-5}$, and $3.80\times {10}^{-5}$ M. These suspensions were used for the concentration‐dependent SERS and UV–vis measurements.

2.5. SERS Measurements for the Concentration‐Dependent Study

SERS measurements were carried out using a Renishaw inVia Reflex Raman microscope equipped with a 2D CCD detector, a Leica confocal microscope, and a 785 nm NIR diode laser as the excitation source. Samples were prepared by adding 30 µL of each thiol stock solution $(1\times {10}^{-3}$ M) to 3.00 mL of each AuNPs suspension, yielding a final thiol concentration of $1\times {10}^{-5}$ M. After 1 min of incubation, spectra were recorded directly in liquid using quartz cuvettes and a macro‐sampling lens. The instrumental parameters were 785 nm excitation, 300 mW laser power, 1200 L/mm grating, 10 s exposure time, and 1 accumulation.

2.6. UV–vis Absorption Measurements for the Concentration‐Dependent Study

UV–vis absorption measurements were acquired using a Thermo Scientific Evolution 201 UV–vis spectrophotometer. Samples were prepared by adding 30 µL of the corresponding thiol stock solution $(1\times {10}^{-3}$ M) to 3.00 mL of each AuNPs suspension, yielding a final thiol concentration of $1\times {10}^{-5}$ M. After 1 min of incubation, aliquots were diluted to a final Au concentration of $3.8\times {10}^{-5}$ M before measurement to avoid detector saturation.

2.7. Kinetic Study

For the kinetic experiments, one working Au concentration was selected for each thiol based on the concentration‐dependent SERS and UV–vis results. The selected concentrations corresponded to conditions that provided a strong measurable SERS response while avoiding, as far as possible, immediate signal loss caused by excessive aggregation or precipitation. Higher Au concentrations were therefore used for the more stable halogenated thiols, whereas lower Au concentrations were selected for the strongly destabilizing thiols. The selected Au concentrations were 1.4 mM for BT─F and BT─Cl, 0.6 mM for BT and BT─OMe, 0.46 mM for BT─Ø, and 0.17 mM for BT─Me and BT─OH.

2.8. Time‐Resolved SERS Measurements

Time‐resolved SERS measurements were performed using the Renishaw inVia Reflex system described above. All measurements were conducted directly in liquid using quartz cuvettes and a macro‐sampling lens. The instrumental parameters were 785 nm excitation, 300 mW laser power, 1200 L/mm grating, 10 s exposure time, and 1 accumulation. For each thiol, a spectrum was first acquired at time 0 before thiol addition. Then, 30 µL of each thiol stock solution $(1\times {10}^{-3}$ M) was added to 3.00 mL of the selected AuNPs suspension, yielding a final thiol concentration of $1\times {10}^{-5}$ M. Once the kinetic experiment started, the cuvette remained undisturbed in the macro‐sampling position for 6 h. Spectra were recorded continuously at 1 min intervals.

2.9. Time‐Dependent UV–vis Absorption Measurements

Time‐dependent UV–vis absorption measurements were carried out using a Thermo Scientific Evolution 201 UV–vis spectrophotometer. Samples were prepared by adding 30 µL of each thiol stock solution $(1\times {10}^{-3}$ M) to 3.00 mL of the selected AuNPs suspension in a glass vial, yielding a final thiol concentration of $1\times {10}^{-5}$ M. The vial was left undisturbed for 6 h. For all thiols, a spectrum at time 0 was recorded before thiol addition. The measurement time points were selected based on the corresponding SERS kinetic profiles, focusing on the most representative stages of the time‐dependent response: 3 min, 30 min, 2 h 30 min, and 6 h for BT─F; 3 min, 30 min, 3 h 30 min, and 6 h for BT─Cl; 2 min, 4 min, 2 h 30 min, and 6 h for BT; 1 min, 20 min, 2 h 30 min, and 6 h for BT─Ø; 30 s, 1 h, 3 h, and 6 h for BT─Me; 1 min, 2 h 30 min, and 6 h for BT─OMe; and 30 s, 1 h, and 6 h for BT─OH. When required, aliquots were diluted with Milli‐Q water before acquisition to a final Au concentration of $1.75\times {10}^{-4}$ M to avoid detector saturation.

2.10. Zeta Potential Measurements

Zeta potential $(\zeta )$ measurements were performed using a Malvern Zetasizer Pro. Samples were prepared by adding 30 µL of each thiol stock solution $(1\times {10}^{-3}$ M) to 3.00 mL of the selected AuNPs suspension in a glass vial, yielding a final thiol concentration of $1\times {10}^{-5}$ M. The vial was left undisturbed for 6 h. The measurement time points were selected based on the corresponding SERS behavior, focusing on the most representative stages of the time‐dependent response: 3 min, 30 min, 2 h 30 min, and 6 h for BT─F; 3 min, 30 min, 3 h 30 min, and 6 h for BT─Cl; 7 min, 2 h 30 min, and 6 h for BT; 1 min, 20 min, 2 h 30 min, and 6 h for BT─Ø; 30 s, 1 h, 3 h, and 6 h for BT─Me; 1 min, 2 h 30 min, and 6 h for BT─OMe; and 30 s, 1 h, and 6 h for BT─OH. For all thiols, a measurement at time 0 was performed before thiol addition. For each measurement, an aliquot was withdrawn from the vial and transferred to the zeta‐potential cell.

2.11. Optical Images

Optical images were acquired using a smartphone camera. For each condition, 30 µL of thiol stock solution $(1\times {10}^{-3}$ M) was added to 3.00 mL of the selected AuNPs suspension in a glass vial, yielding a final thiol concentration of $1\times {10}^{-5}$ M. The vial was left undisturbed for 6 h. Images were recorded at time 0 before thiol addition and at 1 min, 5 min, 15 min, 30 min, and 1 h, 2 h, 3 h, 4 h, 5 h, and 6 h after thiol addition.

2.12. Theoretical Basis

The experimental SERS intensity can be treated as an effective response resulting from the interplay between the progressive formation of a SERS‐active state and the time‐dependent evolution of the colloidal environment. This approximation is justified because SERS intensity is not governed solely by analyte amount, but also by hotspot formation, analyte accessibility to high‐field regions, and the structural stability of the plasmonic configuration [2]. Under this assumption, the fitted intensity can be written as

$$I_{\mathrm{fit}}\left(t\right)=S_0+A\alpha \left(t\right){\mathrm{\Phi }}_{\mathrm{floc}}\left(t\right)$$ (1)

where $S_0$ is the residual baseline contribution, $A$ is the effective signal amplitude, $\alpha (t)$ describes the formation of the SERS‐efficient state, and ${\mathrm{\Phi }}_{\mathrm{floc}}(t)$ accounts for the attenuation or modulation introduced by colloidal aggregation and flocculation.

To describe the progressive appearance of the SERS‐active state, a Prout–Tompkins‐type kinetic law is adopted. In its general form

$$\frac{d\alpha }{dt}=k_{\mathrm{PT}}{\alpha }^m{(1-\alpha )}^n$$ (2)

This expression is commonly used to describe sigmoidal processes involving an induction period, an accelerated growth regime, and a final saturation regime [34]. For the minimal case $m=1$ and $n=1$, Equation (2) becomes

$$\frac{d\alpha }{dt}=k_{\mathrm{PT}}\alpha \left(1-\alpha \right),$$ (3)

Separation of variables gives

$$\frac{d\alpha }{\alpha \left(1-\alpha \right)}=k_{\mathrm{PT}}dt,$$ (4)

and integration yields

$$\ln \left(\frac{\alpha }{1-\alpha }\right)=k_{\text{PT}}(t-t_0).$$ (5)

After redefining the integration constant as $-k_{\mathrm{PT}}{t}_0$, this expression becomes

$$\ln \left(\frac{\alpha }{1-\alpha }\right)=k_{\mathrm{PT}}\left(t-t_0\right).$$ (6)

Solving for $\alpha (t)$ leads to the logistic expression

$$\alpha \left(t\right)=\frac{1}{1+e^{-k_{\mathrm{PT}}\left(t-t_0\right)}}.$$ (7)

This term describes a slow initial stage, followed by rapid activation and, finally, an approach to saturation. In the present context, it represents the effective buildup of a SERS‐efficient state, which may arise from progressive aggregation to reach hotspot generation.

The second component of the model accounts for the temporal evolution of the colloidal system. For diffusion‐controlled binary coagulation, the classical Smoluchowski expression can be written as

$$K_{\mathrm{B}}=4\pi \left(R_i+R_j\right)\left(D_i+D_j\right),$$ (8)

where $R_i$ and $R_j$ are the hydrodynamic radii of the interacting particles, and $D_i$ and $D_j$ are their diffusion coefficients. For identical particles, $R_i=R_j=R_h$ and $D_i=D_j=D$, giving

$$K_{\mathrm{B}}=16\pi R_{h}D.$$ (9)

The diffusion coefficient is related to the hydrodynamic radius through the Stokes–Einstein relation

$$D=\frac{k_{\mathrm{B}}T}{6\pi \eta R_h}.$$ (10)

Substitution of Equation (10) into Equation (9) gives

$$K_{\mathrm{B}}=\frac{8k_{\mathrm{B}}T}{3\eta },$$ (11)

which corresponds to the diffusion‐limited coagulation rate constant for identical spherical particles. In real colloidal systems, deviations from the ideal fast‐coagulation limit are expected because of electrostatic repulsion, ligand‐shell effects, solvation, interparticle potential barriers, and non‐ideal aggregate structures. Therefore, an effective coagulation constant $K_{\mathrm{coag}}$ is introduced instead of the ideal Brownian value [35].

If $N(t)$ denotes the number concentration of free or effectively accessible primary colloidal entities, binary coagulation can be written as

$$\frac{dN}{dt}=-K_{\mathrm{coag}}{N}^2,$$ (12)

which gives

$$\frac{dN}{N^2}=-K_{\mathrm{coag}}dt.$$ (13)

Integration yields

$$-\frac{1}{N}=-K_{\mathrm{coag}}t+C^{\prime }.$$ (14)

Using $N(0)=N_0$, the time‐dependent concentration becomes

$$\frac{1}{N}=\frac{1}{N_0}+K_{\mathrm{coag}}t\Rightarrow N\left(t\right)=\frac{N_0}{1+K_{\mathrm{coag}}{N}_{0}t}.$$ (15)

Defining $k_f=K_{\mathrm{coag}}{N}_0$, the normalized number of accessible colloidal entities is

$$\frac{N\left(t\right)}{N_0}=\frac{1}{1+k_{f}t}.$$ (16)

This hyperbolic decay is the expected signature of binary coagulation. However, the SERS signal does not necessarily scale linearly with the number of dispersed primary particles. Instead, it depends nonlinearly on local electromagnetic enhancement, hotspot accessibility, and the optical activity of evolving aggregates. For this reason, the colloidal contribution is introduced through a phenomenological power‐law mapping

$${\mathrm{\Phi }}_{\mathrm{floc}}\left(t\right)={\left(\frac{N\left(t\right)}{N_0}\right)}^p={(1+k_{f}t)}^{-p}.$$ (17)

Here, $k_f$ sets the characteristic time scale of the colloidal evolution, whereas $p$ describes how strongly this evolution is translated into the measured SERS response. The exponent $p$ is, therefore, not a fundamental coagulation parameter, but an effective phenomenological descriptor that accounts for the nonlinear relationship between colloidal evolution and SERS intensity [2].

Substitution of the sigmoidal activation term and the flocculation term into Equation (1) gives

$$I_{\mathrm{fit}}\left(t\right)=S_0+A\left[\frac{1}{1+e^{-k_{\mathrm{PT}}\left(t-t_0\right)}}\right]{(1+k_{f}t)}^{-p}.$$ (18)

This expression combines a Prout–Tompkins‐type activation term with a Brownian‐coagulation‐inspired attenuation term. It should be interpreted as a semi‐mechanistic model: the logistic component describes the progressive emergence of a SERS‐efficient state, whereas the hyperbolic component describes the gradual modulation or loss of this state as the colloidal system evolves towards flocculation.

For limited experimental time windows, or when $k_{f}t\ll 1$, the hyperbolic term can be approximated using

$$\ln \left(1+k_{f}t\right)\approx k_{f}t.$$ (19)

Thus,

$${(1+k_{f}t)}^{-p}=e^{-p\ln (1+k_{f}t)}\approx e^{-p{k}_{f}t}.$$ (20)

Defining $k_{\mathrm{eff}}=p{k}_f$, the working fitting expression becomes

$$I_{\mathrm{fit}}\left(t\right)=S_0+A\left[\frac{1}{1+e^{-k_{\mathrm{PT}}\left(t-t_0\right)}}\right]e^{-k_{\mathrm{eff}}t}.$$ (21)

This logistic‐times‐exponential form preserves the physical meaning of a competition between hotspot formation and flocculation attenuation, while reducing the number of adjustable parameters. In this formulation, $k_{\mathrm{eff}}$ should be interpreted as an effective attenuation constant that includes both the time scale of colloidal evolution and the nonlinear sensitivity of the SERS response to this evolution. It is, therefore, not a direct Smoluchowski coagulation constant, but an operational descriptor of signal loss associated with flocculation [35]. In this formulation, $I_{\mathrm{fit}}(t)$ is the model‐predicted SERS intensity at time $t$; $S_0$, the residual baseline intensity; $A$, the effective amplitude of the SERS‐active contribution; $k_{\mathrm{PT}}$, the effective kinetic constant governing the sigmoidal activation; $t_0$, the characteristic inflection time of the logistic component; $K_{\mathrm{coag}}$, the effective binary coagulation constant introduced in the Smoluchowski‐type treatment; $k_f=K_{\mathrm{coag}}{N}_0$, the effective colloidal‐evolution constant derived from the coagulation‐inspired term; $p$, describes how strongly colloidal evolution is translated into the measured SERS response; and, $k_{\mathrm{eff}}=p{k}_f$ is the effective attenuation constant used in the reduced logistic‐times‐exponential fitting model.

3. Results and Discussion

Here, citrate‐stabilized AuNPs were used as a model colloidal SERS platform to evaluate how thiol substituents influence both signal generation and colloidal evolution. These nanoparticles were selected because of their high affinity toward thiols, chemical stability, and widespread use in SERS substrates [36]. The AuNPs were synthesized by citrate reduction following a modified literature procedure [29], yielding spherical particles with an average diameter of 23.43 ± 2.51 nm, as determined by TEM analysis (Figure 1A). The corresponding UV–vis absorption spectrum displayed a single dipolar localized surface plasmon resonance (LSPR) band centered at ca. 525 nm (Figure 1B), consistent with well‐dispersed spherical AuNPs in this size range [37]. The initial ζ‐potential of approximately −29 mV further confirmed that the citrate‐capped AuNPs suspension provided a sufficiently stable colloidal starting point for monitoring analyte‐induced changes [34].

FIGURE 1.

(A) Representative TEM image of the synthesized AuNPs. The inset shows the particle size distribution, with an average diameter of 23.43 ± 2.51 nm. (B) UV–vis absorption spectrum of the AuNPs suspension, showing a single dipolar localized surface plasmon resonance (LSPR) band centered at ca. 525 nm, characteristic of well‐dispersed spherical AuNPs in this size range.

A series of para‐substituted benzenethiols was then selected to probe the influence of substituent chemistry on the SERS response. Benzenethiol (BT) was used as the reference molecule, while 4‐fluorothiophenol (BT─F), 4‐chlorothiophenol (BT─Cl), biphenyl‐4‐thiol (BT─Ø), 4‐methylbenzenethiol (BT─Me), 4‐methoxybenzenethiol (BT─OMe), and 4‐hydroxybenzenethiol (BT─OH) were chosen to introduce functional groups with different electronic and interfacial properties. These molecules are particularly suitable for this study because the thiol group promotes adsorption onto Au through Au─S bond formation, whereas the aromatic ring provides intense and structurally informative Raman features [12, 38]. Representative SERS spectra of the investigated thiols adsorbed on AuNPs are shown in Figure 2. Each molecule exhibited a characteristic spectral fingerprint, confirming successful interaction with the plasmonic surface. Despite substituent‐dependent spectral differences, all spectra displayed common aromatic thiolate features, including bands in the ca. 1060–1110 cm⁻¹ region, assigned mainly to ring‐breathing/ring‐stretching modes coupled to ν(C─S), and bands around 1570–1585 cm⁻¹, associated with aromatic ν(C═C) stretching vibrations [12, 39, 40, 41, 42]. These shared features enabled direct comparison of SERS intensity trends among the different thiols while preserving molecular specificity. With the AuNPs platform and reporter fingerprints established, the effect of AuNPs concentration on the substituent‐dependent SERS response was then examined.

FIGURE 2.

Representative SERS spectra of benzenethiol (BT), 4‐fluorothiophenol (BT─F), 4‐chlorothiophenol (BT─Cl), biphenyl‐4‐thiol (BT─Ø), 4‐methylbenzenethiol (BT─Me), 4‐methoxybenzenethiol (BT─OMe), and 4‐hydroxybenzenethiol (BT─OH), together with their corresponding molecular structures. The highlighted spectral regions indicate common aromatic thiolate bands: ca. 1060–1110 cm⁻¹, assigned mainly to ring‐breathing/ring‐stretching modes coupled to ν(C─S), and ca. 1570–1585 cm⁻¹, assigned to aromatic ν(C═C) stretching vibrations.

To assess how nanoparticle availability modulates the substituent‐dependent response, the SERS intensity was first evaluated as a function of AuNPs concentration at a fixed thiol concentration of 10⁻⁵ M. The intensity of the ca. 1070 cm⁻¹ aromatic thiolate band was used as a common descriptor of the SERS response, enabling direct comparison among the different substituted benzenethiols. As shown in Figure 3, variation of the AuNPs concentration revealed three distinct response regimes, indicating that the substituent not only affects the spectral intensity but also determines the colloidal state in which the SERS signal is generated. The first regime comprises the halogenated thiols BT─F and BT─Cl. These molecules showed weak SERS signals at low AuNPs concentration, followed by a marked increase at higher AuNPs concentration and an eventual tendency toward saturation (Figures 3A and S1). In parallel, the UV–vis spectra retained the main LSPR band without pronounced broadening or red‐shifting (Figures 3B and S1), indicating that the colloidal suspension remained largely stable under these conditions. Therefore, the increase in SERS intensity cannot be primarily assigned to irreversible aggregation, but rather to a higher probability of forming transient or dynamic SERS‐efficient interparticle configurations at elevated nanoparticle concentration. This behavior is consistent with a regime in which the colloidal stability is mostly preserved while hotspot generation remains concentration‐dependent. The second regime includes BT, BT─Ø, and BT─OMe. For these molecules, the SERS intensity increased sharply at lower AuNPs concentrations than for BT─F and BT─Cl, reached a maximum, and then decreased at higher AuNPs concentrations (Figures 3A and S1). The corresponding UV–vis spectra displayed clear LSPR broadening and red‐shifting (Figures 3C and S1), evidencing stronger interparticle coupling and aggregation. In this regime, partial aggregation promotes the formation of efficient hot spots and enhances the SERS response. However, when aggregation becomes excessive, the number of optically active suspended structures decreases, leading to signal attenuation. Thus, these thiols define a productive but less stable colloidal regime, where SERS enhancement and colloidal destabilization are closely coupled. The third regime is represented by BT─Me and BT─OH. These thiols produced the most abrupt concentration‐dependent response, with the SERS maximum and subsequent signal decay occurring at lower AuNPs concentration than for the other molecules (Figures 3A and S1). At the same time, the UV–vis spectra showed pronounced broadening and red‐shifting across the concentration range (Figures 3D and S1), confirming strong colloidal destabilization. The high maximum intensities obtained in this regime suggest the formation of highly enhancing aggregates, but the rapid decrease in signal indicates that these structures are not colloidally persistent. BT─Me and BT─OH, therefore, behave as strongly destabilizing analytes, for which intense SERS signals are generated only within a narrow concentration window before aggregation and flocculation dominate the response. Thus, the concentration‐dependent experiments show that the SERS response is not controlled solely by the amount of plasmonic material available. Instead, the effective signal results from a balance between nanoparticle concentration, hotspot formation, and substituent‐dependent colloidal stability. This first classification defines three operational regimes: a stable but delayed‐enhancement regime for BT─F and BT─Cl, a productive aggregation regime for BT, BT─Ø, and BT─OMe, and a strongly destabilized regime for BT─Me and BT─OH. Since aggregation and flocculation are inherently time‐dependent processes, kinetic SERS measurements were then performed to determine how these concentration‐defined regimes evolve over the experimental time window.

FIGURE 3.

A schematic summary of the Au concentration‐dependent SERS response of monosubstituted benzenethiols on AuNPs suspensions. Analytes can be classified into three groups based on their SERS response in relation to colloidal stability. (A) Representative SERS intensity profiles as a function of Au concentration for the three groups. Illustrations of the corresponding colloidal stability behavior are shown above each profile. (B–D) A representative absorption UV–vis spectra recorded at increasing Au concentration for (B) group I, (C) group II, and (D) group III.

To determine whether the concentration‐defined regimes persist over time, time‐resolved SERS measurements were performed over 6 h at a selected AuNPs concentration for each thiol. These concentrations were chosen from the concentration‐dependent study (Figure S1) as the conditions providing a strong and measurable SERS response while avoiding, as far as possible, immediate signal loss caused by excessive aggregation or precipitation. Thus, higher AuNPs concentrations were selected for BT─F and BT─Cl, because these thiols preserved colloidal stability and required a larger nanoparticle density to generate efficient interparticle configurations. Intermediate AuNPs concentrations were used for BT, BT─Ø, and BT─OMe, where partial aggregation already produced strong SERS enhancement, but excessive AuNPs concentration accelerated signal attenuation. For BT─Me and BT─OH, lower AuNPs concentrations were selected because these analytes strongly destabilized the colloid even under diluted conditions. In this way, the selected concentrations provided the most informative experimental window for comparing the time‐dependent formation and loss of SERS‐efficient states. The kinetic traces revealed that the same three groups identified in Figure 3 also display distinct temporal behaviors, confirming that substituent chemistry controls not only the intensity of the SERS response but also the time window over which the colloidal system remains analytically useful. For BT─F and BT─Cl, the kinetic measurements were performed at 1.4 mM Au. Both halogenated thiols displayed a delayed SERS response, characterized by a weak initial signal followed by a gradual increase and a maximum after several hours (Figures 4A and S2). The UV–vis spectra showed moderate broadening and red‐shifting of the LSPR band over time (Figures 4B and S2), indicating the development of interparticle coupling. However, the ζ‐potential values remained highly negative throughout the experiment, close to or below −30 mV, showing that the overall colloidal stability was largely preserved [43]. Optical images further confirmed that precipitation was limited compared with the other thiols (Figure S3). Therefore, BT─F and BT─Cl define a delayed‐development regime in which SERS‐efficient configurations form slowly, but the colloidal suspension remains sufficiently stable to provide a relatively broad measurement window. BT, BT─Ø, and BT─OMe exhibited a faster and stronger temporal response. At their selected AuNPs concentrations, the SERS signal increased sharply within the first hour, reached a maximum earlier than BT─F and BT─Cl, and then progressively decreased (Figures 4A and S2). This behavior was accompanied by pronounced LSPR broadening and red‐shifting, together with ζ‐potential values closer to neutrality, indicating stronger colloidal destabilization (Figures 4B and S2). The optical images showed an initial darkening of the suspension, consistent with aggregation, followed by progressive clarification due to flocculation and sedimentation (Figure S3). These results indicate that partial aggregation initially promotes efficient hotspot formation, but the same process later reduces the number of optically active structures remaining in suspension. This group, therefore, corresponds to a productive aggregation regime, where high SERS intensities are achieved at the expense of a shorter operational time window. The most rapidly evolving behavior was observed for BT─Me and BT─OH. At 0.17 mM Au, both thiols produced intense SERS signals at the earliest measurable times, followed by a rapid decay over the 6 h experiment (Figures 4A and S2). In contrast to the previous groups, the signal‐growth stage was not experimentally resolved, indicating that colloidal destabilization and hotspot formation occurred faster than the acquisition time scale. The UV–vis spectra showed severe LSPR broadening, red‐shifting, and intensity loss, while ζ‐potential values close to neutrality confirmed strong destabilization of the citrate‐stabilized AuNPs (Figures 4B and S2). Optical images showed rapid darkening followed by visible precipitation, confirming extensive aggregation and flocculation (Figure S3). BT─Me and BT─OH, therefore, define an early‐collapse regime, where strong but short‐lived SERS signals are dominated by rapid colloidal destabilization rather than by a stable SERS‐active state. The kinetic experiments demonstrate that the concentration‐dependent classification in Figure 3 also translates into distinct temporal regimes. Halogenated thiols provide slower, but longer‐lived SERS responses, BT, BT─Ø, and BT─OMe generate stronger signals over an intermediate time window, and BT─Me and BT─OH induce rapid colloidal collapse. These results show that the highest instantaneous SERS intensity is not necessarily the most useful condition for quantitative measurements. Instead, reliable liquid‐phase SERS requires balancing signal intensity, activation time, and colloidal persistence. The differences among the three regimes can be rationalized by considering how the para substituent modifies the interfacial environment after Au─S bond formation. The halogenated thiols perturb the citrate‐stabilized interface only weakly, as supported by their persistently negative ζ‐potential values and limited precipitation. Their SERS response, therefore, develops as nanoparticle concentration and time increase the probability of forming SERS‐efficient interparticle configurations without inducing rapid colloidal collapse. In contrast, BT, BT─Ø, and BT─OMe promote stronger interparticle association through the combined effects of hydrophobicity, aromatic or lateral adsorbate interactions, substituent polarizability, and partial screening or displacement of the citrate shell. This interfacial perturbation favours productive aggregation and hotspot formation, but also reduces the operational measurement window. Finally, BT─Me and BT─OH induce rapid colloidal collapse, indicating a stronger disruption of the stabilizing interface. For BT─Me, this behavior is consistent with enhanced hydrophobic adsorbate–adsorbate interactions, whereas for BT─OH, it is consistent with additional hydrogen‐bonding or acid–base interactions involving citrate and the aqueous environment.

FIGURE 4.

(A) SERS intensity as a function of time over 6 h at fixed thiol concentration (10⁻⁵ M) and at the AuNPs concentration selected for each thiol from the concentration‐dependent study. The kinetic traces reveal three temporal regimes: delayed SERS‐state development, productive aggregation followed by attenuation, and rapid flocculation‐dominated decay. (B) UV–vis absorption spectra recorded at selected times after thiol addition, showing the time‐dependent evolution of the LSPR band. Broadening, red‐shifting, and loss of absorbance indicate increasing interparticle coupling, aggregation, and flocculation. The spectrum of the initial AuNPs suspension is included for comparison.

Taken together, the concentration‐dependent and time‐dependent results show that the SERS intensity is governed by two coupled processes: the formation of SERS‐efficient interparticle configurations (hotspots) and their subsequent loss or attenuation through colloidal flocculation. This distinction is important because, in colloidal SERS, the measured signal does not depend only on analyte adsorption, but also on hotspot generation, hotspot accessibility, and the persistence of optically active nanoparticle assemblies [1]. Accordingly, the kinetic profiles were fitted using the semi‐mechanistic expression described above (Equation (21)), in which colloidal evolution is described with a logistic activation term that accounts for the progressive emergence of SERS due to the aggregation that forms efficient stable hotspots, and an exponential attenuation term that accounts for signal loss associated with colloidal growth of the aggregates, which leads to flocculation.

The fitted parameters are summarized in Table 1, and the normalized fitted profiles are compared in Figure 5A. The model reproduced the main kinetic features observed experimentally, including the delayed signal development for BT─F and BT─Cl, the faster rise and subsequent decay for BT, BT─Ø, and BT─OMe, and the rapid attenuation observed for BT─Me and BT─OH. Importantly, the fitted curves should not be interpreted as direct molecular adsorption kinetics. Instead, they represent effective kinetic envelopes that integrate molecular adsorption, interparticle coupling, hotspot formation, and colloidal destabilization into a single observable SERS response. For BT─F and BT─Cl, the fitted profiles show relatively delayed activation, with lower $k_{\mathrm{PT}}$ values and larger $t_0$ values than the intermediate group. This agrees with the experimental observation that these thiols preserve colloidal stability more effectively and require longer times to develop highly efficient SERS‐active configurations (Figures 4 and S2). Their moderate $k_{\mathrm{eff}}$ values indicate that signal attenuation occurs, but not through immediate colloidal collapse. Therefore, these thiols define a delayed‐development regime, characterized by slower hotspot maturation but a broader operational time window. BT, BT─Ø, and BT─OMe displayed larger $k_{\mathrm{PT}}$ values and earlier $t_0$ values, consistent with faster formation of SERS‐efficient configurations. This behavior matches the concentration‐dependent and UV–vis results, where partial aggregation promotes stronger interparticle coupling and higher SERS intensity (Figures 3 and 4). Their lower‐to‐moderate $k_{\mathrm{eff}}$ values indicate that the active structures persist long enough to generate intense signals before flocculation progressively attenuates the response. These thiols therefore correspond to a productive aggregation regime, in which partial destabilization is beneficial for SERS enhancement but reduces the available measurement window. BT─Me and BT─OH were dominated by rapid signal decay from the earliest measurable times. For these molecules, the activation stage was not experimentally resolved, and therefore, $k_{\mathrm{PT}}$ and $t_0$ should not be treated as meaningful activation descriptors. Instead, the dominant fitted parameter is $k_{\mathrm{eff}}$, which is markedly higher than for the other thiols. This agrees with the UV–vis, ζ‐potential, and optical‐image data, all of which show rapid destabilization, aggregation, and precipitation (Figures 4, S2, and S3). BT─Me and BT─OH are therefore better described as an early‐collapse regime, where the SERS response is controlled primarily by rapid colloidal destabilization rather than by the gradual development of a stable SERS‐active state. The $k_{\mathrm{PT}}$ versus $k_{\mathrm{eff}}$ representation in Figure 5B summarizes this classification. $k_{\mathrm{PT}}$ describes how rapidly the SERS‐efficient state develops when the activation regime is experimentally resolved, whereas $k_{\mathrm{eff}}$ describes the effective attenuation associated with colloidal evolution. In this parameter space, the investigated thiols separate into three regimes: delayed development for BT─F and BT─Cl, productive aggregation for BT, BT─Ø, and BT─OMe, and early collapse for BT─Me and BT─OH. This representation reinforces the central conclusion that the most useful SERS reporter is not necessarily the one that gives the highest instantaneous signal, but the one that provides the best compromise between signal intensity, response time, and colloidal persistence.

TABLE 1.

Fitted kinetic parameters obtained from the logistic‐times‐exponential model (Equation (21)) for the time‐dependent SERS response of para‐substituted benzenethiols.

PROBE	S 0 a	A b	k PT (h−1) c	t 0 (h) d	k eff (h−1) e	R 2
BT─F	0.30	4.63	1.62	2.20	0.61	0.93
BT─Cl	0.40	4.92	1.39	2.98	0.51	0.82
BT	0.23	1.27	3.50	1.50	0.23	0.96
BT─Ø	0.39	0.97	3.68	1.53	0.22	0.95
BT─OMe	0.28	0.98	3.92	1.47	0.19	0.95
BT─Me	0.71	0.23	n.r. f	n.r. f	6.67	0.94
BT─OH	0.68	0.26	n.r. f	n.r. f	6.31	0.95

^a Residual baseline intensity.

^b Effective amplitude of the SERS‐active contribution.

^c Effective activation constant.

^d Inflection time of the logistic component.

^e Effective attenuation constant associated with colloidal evolution.

^f The parameters $k_{\mathrm{PT}}$, and $t_0$ are not reported for BT─Me and BT─OH because the activation stage was not experimentally resolved and the response was dominated by rapid flocculation.

FIGURE 5.

(A) Normalized fitted SERS kinetic profiles obtained using the logistic‐times‐exponential model (Equation (21)). The profiles compare the effective temporal evolution of the SERS response for all investigated thiols and reveal three regimes: delayed SERS‐state development for BT─F and BT─Cl, productive aggregation for BT, BT─Ø, and BT─OMe, and rapid flocculation‐dominated decay for BT─Me and BT─OH. (B) $k_{\mathrm{PT}}$ versus $k_{\mathrm{eff}}$ representation of the fitted parameters and classification of their SERS behavior.

4. Conclusion

In summary, here we demonstrate that the performance of liquid‐phase colloidal SERS is governed not only by the intrinsic SERS response or surface affinity of the analyte, but also by its effect on nanoparticle stability and colloidal evolution. By studying a series of para‐substituted benzenethiols on citrate‐stabilized AuNPs, we show that substituent chemistry controls the balance between SERS‐efficient hotspot formation and signal loss through aggregation and flocculation. The combined concentration‐dependent and time‐resolved SERS measurements, supported by UV–vis absorption spectroscopy and ζ‐potential analysis, enabled the classification of the studied thiols into three distinct regimes. BT─F and BT─Cl preserved colloidal stability more effectively, leading to delayed but longer‐lived SERS responses. BT, BT─Ø, and BT─OMe promoted partial aggregation, producing stronger signals over an intermediate time window before attenuation. In contrast, BT─Me and BT─OH induced rapid colloidal destabilization, resulting in intense but short‐lived responses dominated by flocculation. The temporal behavior was rationalized using a semi‐mechanistic kinetic model that combines the progressive formation of a SERS‐efficient state with its subsequent attenuation through colloidal evolution. This approach highlights that the highest instantaneous SERS intensity does not necessarily correspond to the most reliable analytical condition. Instead, quantitative liquid‐phase SERS requires optimization of the compromise between signal intensity, response time, and colloidal persistence. Overall, these findings establish substituent‐dependent colloidal stability as a key design parameter in reporter selection. Careful choice of the analyte or reporter molecule, together with optimization of nanoparticle concentration and measurement time, is therefore essential for improving the reproducibility and reliability of quantitative SERS measurements in colloidal suspensions.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: chem71432‐sup‐0001‐SuppMat.docx

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.