https://chemrxiv.org/engage/chemrxiv/article-details/6769dc3a81d2151a02b75ef6

Skip to main content Accessibility information
We use cookies to distinguish you from other users and to provide you
with a better experience on our websites. Close this message to
accept cookies or find out how to manage your cookie settings. Learn
more about our Privacy Notice... [opens in a new tab]
Cambridge Open Engage home
What is Cambridge Open Engage?
ChemRxiv Home
How to Submit
Browse
About
 
News [opens in a new tab]
Log in
Back to Nanoscience
Search within Nanoscience[                    ]
 
RSS feed for Nanoscience

Inverse Design of Complex Nanoparticle Heterostructures via Deep
Learning on Heterogeneous Graphs

26 December 2024, Version 1
Working Paper

Authors

  * Eric Sivonxay ,
  * Lucas Attia ,
  * Evan Walter Clark Spotte-Smith Author ORCID: We display the ORCID
    iD icon alongside authors names on our website to acknowledge
    that the ORCiD has been authenticated when entered by the user.
    To view the users ORCiD record click the icon. [opens in a new
    tab] ,
  * Benjamin Lengeling ,
  * Xiaojing Xia ,
  * Daniel Barter ,
  * Emory Chan ,
  * Samuel Blau

Show author details
[NonPeerRev] This content is a preprint and has not undergone peer
review at the time of posting.
Download
Cite
Comment

Abstract

Applications of deep learning (DL) to design nanomaterials are
hampered by a lack of suitable data representations and training
data. We report efforts to overcome these limitations and leverage DL
to optimize the nonlinear optical properties of core-shell
upconverting nanoparticles (UCNPs). UCNPs, which have applications in
e.g., biosensing, super-resolution microscopy, and 3D printing, can
emit visible and ultraviolet light from near-infrared excitations. We
report the first large-scale dataset of UCNP emission spectra based
on accurate but expensive kinetic Monte Carlo simulations (N > 6,000)
and use this data to train a heterogeneous graph neural network (GNN)
using a novel representation of UCNP nanostructure. Applying
gradient-based optimization on the trained GNN, we identify
structures with 6.5 times higher predicted emission under 800nm
illumination than any UCNP in our training set. Our work reveals new
design principles for UCNPs and presents a roadmap for DL-based
inverse design of nanomaterials.

Keywords

upconverting
nanoparticle
emission
machine learning
ml
graph neural network
heterogeneous
deep learning
dataset
core-shell
doping
optimization
representation

Supplementary materials

Title
Description
Actions
Title
[pdfIcon]
Supporting Information
Description
Additional information about data pre-processing; consideration of
the effect of NaYF4 structure on simulation results; details for
tabular, image, and homogeneous graph representations for UCNPs;
dataset biases and attempts to mitigate bias; optimal model
hyperparameters; additional details regarding model validation and
optimization.
Actions
Download (1 MB)

Comments

You are signed in as . Your name will appear with any comment you
post.
Comments are not moderated before they are posted, but they can be
removed by the site moderators if they are found to be in
contravention of our Commenting Policy [opens in a new tab] - please
read this policy before you post. Comments should be used for
scholarly discussion of the content in question. You can find more
information about how to use the commenting feature here [opens in a
new tab] .
 
[                    ]
[                    ]
[                    ]
300 words allowed
You can enter up to 300 words. Post comment
Log in or register with ORCID to comment
Comments are not moderated before they are posted, but they can be
removed by the site moderators if they are found to be in
contravention of our Commenting Policy [opens in a new tab] - please
read this policy before you post. Comments should be used for
scholarly discussion of the content in question. You can find more
information about how to use the commenting feature here [opens in a
new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy 
[opens in a new tab] and Terms of Service [opens in a new tab] apply.

Version History

Dec 26, 2024 Version 1

Metrics

0
0
0
Views
Downloads
Citations

License

CC logo
CC
BY logo
BY
The content is available under CC BY 4.0 [opens in a new tab]

DOI

 
10.26434/chemrxiv-2024-1dw4q
D O I: 10.26434/chemrxiv-2024-1dw4q [opens in a new tab]

Funding

Charter Hill Laboratory Directed Research and Development, Lawrence
Berkeley National Laboratory
DE-AC02-05CH11231
DOE Computational Science Graduate Fellowship
DE-SC0022158
Carnegie Bosch Institute

Author's competing interest statement

The author(s) have declared they have no conflict of interest with
regard to this content

Ethics

The author(s) have declared ethics committee/IRB approval is not
relevant to this content

Share

How to submit
Browse
About us
News [opens in a new tab]
Technology Provided By
Cambridge University Press home
ChemRxiv home ISSN 2573-2293 (Online) ChemRxiv Archive
Legal Notices
Privacy Policy
Accessibility
Contact and Help
Public API
RSS feed for ChemRxiv
1.0.2740