Deep Learning Methods for Colloidal Silver Nanoparticle Concentration and Size Distribution Determination from UV-Vis Extinction Spectra
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911907524378624 |
|---|---|
| author | Klinavičius, Tomas Khinevich, Nadzeya Tamulevičienė, Asta Vidal, Loic Tamulevičius, Sigitas Tamulevičius, Tomas |
| author_facet | Klinavičius, Tomas Khinevich, Nadzeya Tamulevičienė, Asta Vidal, Loic Tamulevičius, Sigitas Tamulevičius, Tomas |
| contents | Electron microscopy, while reliable, is an expensive, slow, and inefficient technique for thorough size distribution characterization of both mono- and polydisperse colloidal nanoparticles. If rapid in-situ characterization of colloid samples is to be achieved, a different approach, based on fast, widely accessible, and inexpensive optical measurements such as UV-Vis spectroscopy in combination with spectra interpretation related to Mie scattering theory, is needed. In this article, we present a tandem deep neural network (DNN) for the size distribution and concentration prediction of close to spherical silver colloidal nanoparticle batches synthesized via the seeded growth method. The first DNN identified the dipole component of the localized surface plasmon resonance and the second one determined the size distribution from the isolated spectral component. The training data was engineered to be bias-free and generated numerically. High prediction accuracy with root mean square percentage error of mean size down to 1.2% was achieved, spanning the entire prediction range from 1 nm up to 150 nm in radius, suggesting the possible extension limits of the effective medium theory used for simulating the spectra. The DNN-predicted nanoparticle concentrations also were very close to the ones expected based on synthesis precursor contents as well as measured by atomic absorption spectroscopy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_10891 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Deep Learning Methods for Colloidal Silver Nanoparticle Concentration and Size Distribution Determination from UV-Vis Extinction Spectra Klinavičius, Tomas Khinevich, Nadzeya Tamulevičienė, Asta Vidal, Loic Tamulevičius, Sigitas Tamulevičius, Tomas Computational Physics Disordered Systems and Neural Networks Electron microscopy, while reliable, is an expensive, slow, and inefficient technique for thorough size distribution characterization of both mono- and polydisperse colloidal nanoparticles. If rapid in-situ characterization of colloid samples is to be achieved, a different approach, based on fast, widely accessible, and inexpensive optical measurements such as UV-Vis spectroscopy in combination with spectra interpretation related to Mie scattering theory, is needed. In this article, we present a tandem deep neural network (DNN) for the size distribution and concentration prediction of close to spherical silver colloidal nanoparticle batches synthesized via the seeded growth method. The first DNN identified the dipole component of the localized surface plasmon resonance and the second one determined the size distribution from the isolated spectral component. The training data was engineered to be bias-free and generated numerically. High prediction accuracy with root mean square percentage error of mean size down to 1.2% was achieved, spanning the entire prediction range from 1 nm up to 150 nm in radius, suggesting the possible extension limits of the effective medium theory used for simulating the spectra. The DNN-predicted nanoparticle concentrations also were very close to the ones expected based on synthesis precursor contents as well as measured by atomic absorption spectroscopy. |
| title | Deep Learning Methods for Colloidal Silver Nanoparticle Concentration and Size Distribution Determination from UV-Vis Extinction Spectra |
| topic | Computational Physics Disordered Systems and Neural Networks |
| url | https://arxiv.org/abs/2404.10891 |