Use of Machine Learning with Temporal Photoluminescence Signals from CdTe Quantum Dots for Temperature Measurement in Microfluidic Devices

Lewis, Charles; Erikson, James; Sanchez, Derek; McClure, C. Emma; Nordin, Gregory P.; Munro, Troy; Colton, John

doi:10.1021/acsanm.0c00065

Cited by 32 publications

(40 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The above mechanisms together with random functional group spatial distributions increase the problem complexity drastically. One possibility to tackle the elevated complexity is to train neural networks with entire stress–strain curves as the input vectors, as inspired by ref . We will consider neural networks in our future work that involves more complex systems and physics.…”

Section: Results and Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of Graphene Oxide Functionalization Using Gradient Boosting: Implications for Material Chemical Composition Identification

Zheng

2021

ACS Appl. Nano Mater.

View full text Add to dashboard Cite

Graphene oxides have exhibited alluring potential for state-of-the-art applications such as biomedical devices and functional nanocomposites. The types and concentrations of oxygen-containing functional groups are fingerprints of graphene oxides, dictating the properties and usage of the nanomaterial. Compared to pure graphene, the properties of graphene oxides are more challenging to model from a theoretical perspective, mainly because of the profound but implicit influences of the functional groups within. Machine learning is a potent method to uncover the hidden structure–property relations and to accelerate material discovery. Here, we develop a machine learning-based strategy to determine the functionalization properties of monolayer graphene oxides, labeled by the oxygen-to-carbon ratio and relative concentrations of functional groups. Trained by mechanical responses upon uniaxial tension computed by reactive molecular dynamics simulations, our proposed gradient boosting machine learning model can accurately identify the chemical composition of graphene oxides in the reserved data set. The difference in prediction accuracies between oxygen coverage and functional group composition is rationalized by graphene oxide molecular mechanisms. The proposed data-driven strategy can contribute to the predictive modeling of functionalized two-dimensional materials of a broad variety.

show abstract

Section: Results and Discussionmentioning

confidence: 99%

“…One possibility to tackle the elevated complexity is to train neural networks with entire stress−strain curves as the input vectors, as inspired by ref. 45 We will consider neural networks in our future work that involves more complex systems and physics.…”

Section: Resultsmentioning

confidence: 99%

Prediction of Graphene Oxide Functionalization Using Gradient Boosting: Implications for Material Chemical Composition Identification

Zheng

2021

ACS Appl. Nano Mater.

View full text Add to dashboard Cite

show abstract

“…We will compare our current results to this previous approach in Section 3.2. The most recent paper in this area [44] used a deep neural network (DNN) trained on both the spectral intensity of fluorescence and the lifetime of the excited molecule to achieve an accuracy of ±0.4 K over a 100-300 K range. However, each measurement took several minutes because hundreds of thousands of laser pulses were needed to create an entire lifetime curve as each pulse measured a single photon event.…”

Section: Demonstration Of Neural Network To Reconstruct Temperatures ...mentioning

confidence: 99%

Demonstration of Neural Networks to Reconstruct Temperatures from Simulated Fluorescent Data towards use in Bio-Microfluidics

Kullberg

Colton

Gregory

et al. 2022

Preprint

View full text Add to dashboard Cite

Biological systems often have a narrow temperature range of operation, which require highly accurate spatially resolved temperature measurements, often near ±0.1K. However, many temperature sensors cannot meet both accuracy and spatial distribution requirements, often because their accuracy is limited by data fitting and temperature reconstruction models. Machine learning algorithms have the potential to meet this need, but their usage in generating spatial distributions of temperature is severely lacking in the literature. This work presents the first instance of using neural networks to process fluorescent images to map the spatial distribution of temperature. Three standard network architectures were investigated using non-spatially resolved fluorescent thermometry (simply-connected feed-forward network) or during image or pixel identification (U-net and convolutional neural network, CNN). Simulated fluorescent images based on experimental data were generated based on known temperature distributions where Gaussian white noise with a standard deviation of ±0.1 K was added. The poor results from these standard networks motivated the creation of what is termed a moving CNN, with an RMSE error of ±0.23 K, where the elements of the matrix represent the neighboring pixels. Finally, the performance of this MCNN is investigated when trained and applied to three distinctive temperature distributions characteristic within microfluidic devices, where the fluorescent image is simulated at either three or five different wavelengths. The results demonstrate that having a minimum of 103.5 data points per temperature and the broadest range of temperatures during training provides temperature predictions nearest to the true temperatures of the images, with a minimum RMSE of ±0.15 K. When compared to traditional curve fitting techniques, this work demonstrates that greater accuracy when spatially mapping temperature from fluorescent images can be achieved when using convolutional neural networks.

show abstract

“…Another class of conundrums, which pertain to the conceptual level and to data post-processing, have been only pragmatically and less systematically dealt with. As a result, the field of luminescence thermometry is repleted by the day with new luminescent probes but few are the improvements at the methodological level; that is, with the exception of a handful of recent works involving multiparametric readouts and machine learning-based regression models 15 – 17 . By exploring the use of dimensionality reduction (DR), our work fits in this frame of underexplored data processing methods, catering to researchers whose goal is to maximize the performance of a luminescence thermometry approach.…”

Section: Introductionmentioning

confidence: 99%

“…Because the calibration of a luminescent thermometer generates large datasets (e.g., intensity vs. several wavelengths at different temperatures), extending DR approaches to luminescence thermometry is a natural step to identify the numerical quantities that better correlate with temperature. However, only few examples of DR methods applied to luminescence thermometry have been reported 15 – 17 , Lewis et al, for instance, obtained a thermal readout through a long short-term memory neural network trained with a combination of raw spectral and time-resolved luminescence data obtained from quantum dots 15 . Šević et al, on the other hand, used principal component analysis to infer the temperature from the luminescence of Sr 2 CeO 4 :Eu 3+ nanophosphors 17 .…”

Section: Introductionmentioning

confidence: 99%

Less is more: dimensionality reduction as a general strategy for more precise luminescence thermometry

et al. 2022

View full text Add to dashboard Cite

Thermal resolution (also referred to as temperature uncertainty) establishes the minimum discernible temperature change sensed by luminescent thermometers and is a key figure of merit to rank them. Much has been done to minimize its value via probe optimization and correction of readout artifacts, but little effort was put into a better exploitation of calibration datasets. In this context, this work aims at providing a new perspective on the definition of luminescence-based thermometric parameters using dimensionality reduction techniques that emerged in the last years. The application of linear (Principal Component Analysis) and non-linear (t-distributed Stochastic Neighbor Embedding) transformations to the calibration datasets obtained from rare-earth nanoparticles and semiconductor nanocrystals resulted in an improvement in thermal resolution compared to the more classical intensity-based and ratiometric approaches. This, in turn, enabled precise monitoring of temperature changes smaller than 0.1 °C. The methods here presented allow choosing superior thermometric parameters compared to the more classical ones, pushing the performance of luminescent thermometers close to the experimentally achievable limits.

show abstract

Use of Machine Learning with Temporal Photoluminescence Signals from CdTe Quantum Dots for Temperature Measurement in Microfluidic Devices

Cited by 32 publications

References 44 publications

Prediction of Graphene Oxide Functionalization Using Gradient Boosting: Implications for Material Chemical Composition Identification

Prediction of Graphene Oxide Functionalization Using Gradient Boosting: Implications for Material Chemical Composition Identification

Demonstration of Neural Networks to Reconstruct Temperatures from Simulated Fluorescent Data towards use in Bio-Microfluidics

Less is more: dimensionality reduction as a general strategy for more precise luminescence thermometry

Contact Info

Product

Resources

About