Robust optimization of silver nanoparticles synthesis to improve quantitative performances of surface enhanced Raman scattering analysis

Nowadays, surface-enhanced Raman scattering (SERS) is a widely used vibrational technique to perform rapid qualitative analyses with a high specificity and sensitivity. This technique is also recognized for its ability to quantify molecules in complex matrices particularly in pharmaceutical and biomedical fields. However, the literature on routine quantitative SERS analyses remains underdeveloped. This can be explained by various potential sources of signal variability including equipment, substrate-analyte interactions and the variability of the substrates themselves. Generally, SERS substrates are either gold nanoparticles (AuNps) or silver nanoparticles (AgNps) in suspension which are homemade prepared leading to inconsistencies. Many researchers work to optimize synthesis protocols to produce homogeneous and repeatable nanoparticles but achieving substrates with uniform shape and size continues to be a significant challenge. This thesis focused on the improvement of the repeatability of SERS analyses.
Initially, a characterization method using single particle inductively coupled plasma mass spectrometry (spICP-MS) was successfully developed to provide information on size, size distribution, nanoparticles concentration and dissolved metal content. The established method enables the characterization of AgNps and AuNps in suspension which are the two main substrates used for SERS analyses.
The next focus was the optimization of AgNps in suspension as they tend to exhibit more variability compared to AuNps. To address this, an innovative approach based on the Quality by Design (QbD) strategy was used to robustly optimize the synthesis of AgNps targeting homogeneous shape and uniform size. The commonly used Lee-Meisel protocol was considered as the starting point for its speed, ease of implementation and low cost. This process involves to reduce silver nitrate by trisodium citrate under boiling conditions. A screening design was initially conducted to identify parameters that have a critical influence on the formation of AgNps. To reduce the variability sources during the AgNps synthesis, microwave irradiation was incorporated into the chemical reduction process. This technique offers better control over key parameters such as temperature and pressure and enables the reaction to occur in a sealed vial under a homogeneous heating. Applying the QbD strategy led to a robust and optimized method to produce AgNps with consistent shape and size. This optimized reaction takes only 3.36 minutes at 130 °C with a stirring at 600 rpm. In this framework, AgNps achieved a maximal SERS variability of 15% for intra-batch and 20% for intra and inter batches.
Finally, the quantitative performances of the optimized AgNps by microwave-assisted synthesis were evaluated by the determination of an impurity, 4-aminophenol, in a complex matrix from a paracetamol-based medicine. For this study and for the first time, a transmission detection mode was used in SERS analyses (SETRS) demonstrating the feasibility of this approach. This detection mode is known to consider a larger sample volume which may help to reduce signal variability. The standard deviation of the repeatability across the quantification range was between 6.7% at a concentration level of 5 μg/mL and 1.2% at 20 μg/mL.
In conclusion, this thesis presents an innovative approach that combines the robust optimization of SERS substrates with the use of transmission detection mode thereby contributing to the reduction of signal variability which addresses one of the main limitations for the routine application of SERS.