Wonjin Yun | Stanford University (original) (raw)

Papers by Wonjin Yun

Energy & Fuels, 2020

We investigated the Young’s modulus and Poisson’s ratio of Green River oil shale at elevated temp... more We investigated the Young’s modulus and Poisson’s ratio of Green River oil shale at elevated temperatures under triaxial conditions using an X-ray computed tomography (CT) scanner for in situ imagi...

This report describes classification methods that recognize the genres of music using both superv... more This report describes classification methods that recognize the genres of music using both supervised and unsupervised learning techniques. The five genres, classical(C), EDM(E), hip-hop(H), jazz(J) and rock(R), were examined and classified. As a feature selection method, discrete Fourier transform (DFT) converted the raw wave signals of each song into the signal amplitude ordered by their frequencies. Based on the analysis of the characteristics of data set, final feature set were collected by averaging the amplitudes of corresponding two different frequency division (XL and XM ). For supervised learning, a training set (mtrain = 50/genre) was used to train the CART (Classification and Regression Tree), and the performance of the genre prediction by CART classifier was evaluated using a test set (mtest = 10/genre). A recognition rate of 86.7% for three genre classification (C, H, and R) was observed, and 60.7% for five genre classification (C, E, H, J, and R) was obtained. For unsu...

Understanding wettability of porous media, namely the relative affinity of the fluids for the sol... more Understanding wettability of porous media, namely the relative affinity of the fluids for the solid, and its influence on the efficiency of wetting-phase displacement of non-wetting phase is a key factor in estimating storage CO and recovery efficiencies of hydrocarbons. Hence, we want to achieve a systematic way to study the complex interplay between wettability and multi-phase trapping. It could be done by developing a time-efficient method for in-situ detection of pore-level wettability from massive number of high resolution micro-pore scale images that covers wide range of field of view (FOV) in the rock. In this report, we addresses an application of the deep-learning technique to classify automatically pore-level wettability. To achieve the objectives, we trained and optimized two deep-learning algorithms, fully connected neural network (FCN) and convolutional neural network (ConvNet), on the micro-scale FOV image data set collected using inhouse fabricated micro-fluidic devi...

In this paper, FCN and CNN were trained on the image data set that were prepared from the in-hous... more In this paper, FCN and CNN were trained on the image data set that were prepared from the in-house fabricated micro-fluidic device and fluid injection experiment. Trained FCN and CNN model provide the best validation accuracy of 0.81 and 0.83, respectively, for the surface wettability classification of sandstone and carbonate pore structure.

Advances in Water Resources

Scientific Reports, Jan 21, 2020

Enhanced oil recovery (EOR) plays a significant role in improving oil production. Tertiary EOR, i... more Enhanced oil recovery (EOR) plays a significant role in improving oil production. Tertiary EOR, including surfactant flooding, can potentially mobilize residual oil after water flooding. Prior to the field deployment, the surfactant performance must be evaluated using site-specific crude oil at reservoir conditions. Core flood experiments are common practice to evaluate surfactants for oil displacement efficiency using core samples. Core flood experiments, however, are expensive and time-consuming and do not allow for pore scale observations of fluid-fluid interactions. This work introduces the framework to evaluate the performance of EOR surfactants via a Reservoir-on-a-Chip approach, which uses microfluidic devices to mimic the oil reservoir. A unique feature of this study is the use of chemically modified micromodels such that the pore surfaces are representative of carbonate reservoir rock. To represent calcium carbonate reservoir pores, the inner channels of glass microfluidic devices were coated with thin layers of calcium carbonate nanocrystals and the surface was modified to exhibit oil-wet conditions through a crude oil aging process. During surfactant screening, oil and water phases were imaged by fluorescence microscopy to reveal the micro to macro scale mechanisms controlling surfactant-assisted oil recovery. The role of the interfacial tension (IFT) and wettability in the microfluidic device was simulated using a phase-field model and compared to laboratory results. We demonstrated the effect of low IFT at the oil-water interface and wettability alteration on surfactant-enhanced oil displacement efficiency; thus providing a time-efficient and low-cost strategy for quantitative and qualitative assessment. In addition, this framework is an effective method for pre-screening EOR surfactants for use in carbonate reservoirs prior to further core and field scale testing.

The Journal of Physical Chemistry C

Dual-mode surface-enhanced Raman scattering (SERS)−surface-enhanced fluorescence (SEF) composite ... more Dual-mode surface-enhanced Raman scattering (SERS)−surface-enhanced fluorescence (SEF) composite nanoparticles have been developed for possible use as oil reservoir tracers. These composite nanoparticles are composed of metal Ag nanostructured cores, specific dye molecules, and a SiO 2 shell coating. Herein, we show that the embedded dye molecules are detectable by both Raman and fluorescence spectroscopies and yield dramatically enhanced detectability due to strong SERS−SEF phenomena with limits of detection (LOD) as low as 1 ppb by fluorescence spectroscopy and 10 ppb by Raman spectroscopy. To determine the optimal structures for signal enhancement for both SERS and SEF, we show how these phenomena are significantly affected by morphologies of the composite nanoparticles. The aggregation status of metal dots and the distance between the metal and dye probe molecules are the crucial factors for enhancement of SERS and SEF signals. Through well-controlled one-pot reactions in microemulsion media, composite nanoparticles with designed morphologies, Ag@SiO 2 core−shell structures, or Ag@SiO 2 /Ag satellite structures have been synthesized, and various dyes have been encoded into these composite nanoparticles. We have demonstrated that the Ag@SiO 2 /Ag satellite nanoparticles exhibit the highest dye molecule signal enhancement through both SERS and SEF phenomena. Imaging studies on the detection and mobility of these specifically designed nanoparticles in microchannels show their detection within micron-sized pores and at low concentrations. The multifunctional composite nanoparticles presented herein contain different dyes which exhibit different fluorescence emission wavelengths and fingerprinted Raman signals. Thus, these strategically designed nanoparticles provide a possible pathway for future use as barcoded smart reservoir tracers.

Lab on a chip, Jan 11, 2017

Silicon-based microfluidic devices, so-called micromodels in this application, are particularly u... more Silicon-based microfluidic devices, so-called micromodels in this application, are particularly useful laboratory tools for the direct visualization of fluid flow revealing pore-scale mechanisms controlling flow and transport phenomena in natural porous media. Current microfluidic devices with uniform etched depths, however, are limited when representing complex geometries such as the multiple-scale pore sizes common in carbonate rocks. In this study, we successfully developed optimized sequential photolithography to etch micropores (1.5 to 21 μm width) less deeply than the depth of wider macropores (>21 μm width) to improve the structural realism of an existing single-depth micromodel with a carbonate-derived pore structure. Surface profilimetry illustrates the configuration of the dual-depth dual-porosity micromodel and is used to estimate the corresponding pore volume change for the dual-depth micromodel compared to the equivalent uniform- or single-depth model. The flow chara...

ACS Applied Nano Materials

Composite magnetic nanoparticles are designed and synthesized with different morphologies as surf... more Composite magnetic nanoparticles are designed and synthesized with different morphologies as surfaceenhanced Raman scattering (SERS) substrates or SERS-active particles. Through the incorporation of a magnetic functionality, we provide a means to concentrate SERS-active nanoparticles in a low-volume microfluidic channel where the detected entity is now either a flowing analyte (e.g., tracer or chemical) or SERS-active particles contained in a target reservoir fluid. This collection strategy allows for detection using small amounts of material and can be optimized to provide selectivity for tracelevel materials detection at the wellsite. We also demonstrate low-concentration detection of dye molecules used for reservoir tracer materials by optimizing the fluid flow rate and the intensity of the magnetic field. Thus, we developed an efficient magnetic SERS microfluidic detection platform for in situ monitoring trace level of analyte molecules. The integration of SERS with microfluidic systems also can extend the application of Raman detection in biomedical research and microreactor monitoring where low volumes of expensive samples make traditional detection methods ineffective or cost prohibitive.

Energy & Fuels, 2020

We investigated the Young’s modulus and Poisson’s ratio of Green River oil shale at elevated temp... more We investigated the Young’s modulus and Poisson’s ratio of Green River oil shale at elevated temperatures under triaxial conditions using an X-ray computed tomography (CT) scanner for in situ imagi...

This report describes classification methods that recognize the genres of music using both superv... more This report describes classification methods that recognize the genres of music using both supervised and unsupervised learning techniques. The five genres, classical(C), EDM(E), hip-hop(H), jazz(J) and rock(R), were examined and classified. As a feature selection method, discrete Fourier transform (DFT) converted the raw wave signals of each song into the signal amplitude ordered by their frequencies. Based on the analysis of the characteristics of data set, final feature set were collected by averaging the amplitudes of corresponding two different frequency division (XL and XM ). For supervised learning, a training set (mtrain = 50/genre) was used to train the CART (Classification and Regression Tree), and the performance of the genre prediction by CART classifier was evaluated using a test set (mtest = 10/genre). A recognition rate of 86.7% for three genre classification (C, H, and R) was observed, and 60.7% for five genre classification (C, E, H, J, and R) was obtained. For unsu...

Understanding wettability of porous media, namely the relative affinity of the fluids for the sol... more Understanding wettability of porous media, namely the relative affinity of the fluids for the solid, and its influence on the efficiency of wetting-phase displacement of non-wetting phase is a key factor in estimating storage CO and recovery efficiencies of hydrocarbons. Hence, we want to achieve a systematic way to study the complex interplay between wettability and multi-phase trapping. It could be done by developing a time-efficient method for in-situ detection of pore-level wettability from massive number of high resolution micro-pore scale images that covers wide range of field of view (FOV) in the rock. In this report, we addresses an application of the deep-learning technique to classify automatically pore-level wettability. To achieve the objectives, we trained and optimized two deep-learning algorithms, fully connected neural network (FCN) and convolutional neural network (ConvNet), on the micro-scale FOV image data set collected using inhouse fabricated micro-fluidic devi...

In this paper, FCN and CNN were trained on the image data set that were prepared from the in-hous... more In this paper, FCN and CNN were trained on the image data set that were prepared from the in-house fabricated micro-fluidic device and fluid injection experiment. Trained FCN and CNN model provide the best validation accuracy of 0.81 and 0.83, respectively, for the surface wettability classification of sandstone and carbonate pore structure.

Advances in Water Resources

Scientific Reports, Jan 21, 2020

Enhanced oil recovery (EOR) plays a significant role in improving oil production. Tertiary EOR, i... more Enhanced oil recovery (EOR) plays a significant role in improving oil production. Tertiary EOR, including surfactant flooding, can potentially mobilize residual oil after water flooding. Prior to the field deployment, the surfactant performance must be evaluated using site-specific crude oil at reservoir conditions. Core flood experiments are common practice to evaluate surfactants for oil displacement efficiency using core samples. Core flood experiments, however, are expensive and time-consuming and do not allow for pore scale observations of fluid-fluid interactions. This work introduces the framework to evaluate the performance of EOR surfactants via a Reservoir-on-a-Chip approach, which uses microfluidic devices to mimic the oil reservoir. A unique feature of this study is the use of chemically modified micromodels such that the pore surfaces are representative of carbonate reservoir rock. To represent calcium carbonate reservoir pores, the inner channels of glass microfluidic devices were coated with thin layers of calcium carbonate nanocrystals and the surface was modified to exhibit oil-wet conditions through a crude oil aging process. During surfactant screening, oil and water phases were imaged by fluorescence microscopy to reveal the micro to macro scale mechanisms controlling surfactant-assisted oil recovery. The role of the interfacial tension (IFT) and wettability in the microfluidic device was simulated using a phase-field model and compared to laboratory results. We demonstrated the effect of low IFT at the oil-water interface and wettability alteration on surfactant-enhanced oil displacement efficiency; thus providing a time-efficient and low-cost strategy for quantitative and qualitative assessment. In addition, this framework is an effective method for pre-screening EOR surfactants for use in carbonate reservoirs prior to further core and field scale testing.

The Journal of Physical Chemistry C

Dual-mode surface-enhanced Raman scattering (SERS)−surface-enhanced fluorescence (SEF) composite ... more Dual-mode surface-enhanced Raman scattering (SERS)−surface-enhanced fluorescence (SEF) composite nanoparticles have been developed for possible use as oil reservoir tracers. These composite nanoparticles are composed of metal Ag nanostructured cores, specific dye molecules, and a SiO 2 shell coating. Herein, we show that the embedded dye molecules are detectable by both Raman and fluorescence spectroscopies and yield dramatically enhanced detectability due to strong SERS−SEF phenomena with limits of detection (LOD) as low as 1 ppb by fluorescence spectroscopy and 10 ppb by Raman spectroscopy. To determine the optimal structures for signal enhancement for both SERS and SEF, we show how these phenomena are significantly affected by morphologies of the composite nanoparticles. The aggregation status of metal dots and the distance between the metal and dye probe molecules are the crucial factors for enhancement of SERS and SEF signals. Through well-controlled one-pot reactions in microemulsion media, composite nanoparticles with designed morphologies, Ag@SiO 2 core−shell structures, or Ag@SiO 2 /Ag satellite structures have been synthesized, and various dyes have been encoded into these composite nanoparticles. We have demonstrated that the Ag@SiO 2 /Ag satellite nanoparticles exhibit the highest dye molecule signal enhancement through both SERS and SEF phenomena. Imaging studies on the detection and mobility of these specifically designed nanoparticles in microchannels show their detection within micron-sized pores and at low concentrations. The multifunctional composite nanoparticles presented herein contain different dyes which exhibit different fluorescence emission wavelengths and fingerprinted Raman signals. Thus, these strategically designed nanoparticles provide a possible pathway for future use as barcoded smart reservoir tracers.

Lab on a chip, Jan 11, 2017

Silicon-based microfluidic devices, so-called micromodels in this application, are particularly u... more Silicon-based microfluidic devices, so-called micromodels in this application, are particularly useful laboratory tools for the direct visualization of fluid flow revealing pore-scale mechanisms controlling flow and transport phenomena in natural porous media. Current microfluidic devices with uniform etched depths, however, are limited when representing complex geometries such as the multiple-scale pore sizes common in carbonate rocks. In this study, we successfully developed optimized sequential photolithography to etch micropores (1.5 to 21 μm width) less deeply than the depth of wider macropores (>21 μm width) to improve the structural realism of an existing single-depth micromodel with a carbonate-derived pore structure. Surface profilimetry illustrates the configuration of the dual-depth dual-porosity micromodel and is used to estimate the corresponding pore volume change for the dual-depth micromodel compared to the equivalent uniform- or single-depth model. The flow chara...

ACS Applied Nano Materials

Composite magnetic nanoparticles are designed and synthesized with different morphologies as surf... more Composite magnetic nanoparticles are designed and synthesized with different morphologies as surfaceenhanced Raman scattering (SERS) substrates or SERS-active particles. Through the incorporation of a magnetic functionality, we provide a means to concentrate SERS-active nanoparticles in a low-volume microfluidic channel where the detected entity is now either a flowing analyte (e.g., tracer or chemical) or SERS-active particles contained in a target reservoir fluid. This collection strategy allows for detection using small amounts of material and can be optimized to provide selectivity for tracelevel materials detection at the wellsite. We also demonstrate low-concentration detection of dye molecules used for reservoir tracer materials by optimizing the fluid flow rate and the intensity of the magnetic field. Thus, we developed an efficient magnetic SERS microfluidic detection platform for in situ monitoring trace level of analyte molecules. The integration of SERS with microfluidic systems also can extend the application of Raman detection in biomedical research and microreactor monitoring where low volumes of expensive samples make traditional detection methods ineffective or cost prohibitive.