Enhancing Selectivity, Sensitivity, and Molecular Identification towards High-performance Graphene Gas Sensors

i

Enhancing Selectivity, Sensitivity, and Molecular Identification

towards High-performance Graphene Gas Sensors

Materials Science

Graduate School of Advanced Science and Technology

Mizuta-Manoharan Laboratory

Osazuwa Gabriel Agbonlahor (s1820001)

1. Background

Detecting minute concentrations of gaseous pollutants in the atmosphere and human breath enables

environmental monitoring and non-invasive detection of ailments. Consequently, due to

graphene’s single-molecule sensitivity, it is highly sought after for high sensitivity gas detection.

However, its poor selectivity, huge p-doping in the atmosphere, and inability to identify adsorbed

gases makes graphene less useful for practical applications in environmental and clinical gas

sensors which are typically deployed in atmospheric conditions.

2. Aim

In this work, we aim to develop high-performance CVD graphene-based gas sensors

suitable for environmental monitoring and non-invasive detection of ailments. To this end, we

demonstrate a facile and viable route for the development of an ammonia selective activated carbon

functionalized graphene field-effect transistor (a-CF-GFET) which shows minimal atmospheric

induced doping. The a-CF-GFET gas sensor demonstrated good recoverability, extreme ammonia

sensitivity (500 ppt) in atmospheric air, with ~ 3 seconds response time. Furthermore, retention of

gas adsorption induced charge transfer in the graphene-gas molecule van der Waals complexes is

demonstrated. Hence, a proof-of-concept (the CNPD method) for characterizing adsorbates on the

graphene channel is achieved and new insights on the graphene-gas van der Waals bonding

interaction realized.

3. Research Results

Herein, an activated carbon-functionalized graphene sensor with extreme ammonia

sensitivity and selectivity in atmospheric air is developed (Figure 1a) from the pyrolysis of the of

the electron beam cross-linked Novolac resin on the graphene nanoribbons (Figure 1b). We show

that the activated carbon (a-C) on the graphene channel is oxidized and porous (Figures 1c&d),

consequently, parts per billion concentrations of ammonia (Figure 2) down to 500 parts per trillion

(ppt) were detected in the atmosphere with a response time of 3 seconds making the a-CF-GFET

sensor the most sensitive ammonia selective sensor so far reported. The extreme ammonia

sensitivity makes the a-CF-GFET sensor suitable for environmental monitoring and non-invasive

medical diagnosis of ailments such as ulcers and kidney problems using ammonia as a biomarker.

ii

Figure 1. Morphological characteristics of the a-CF-GFET sensor. (a) Device schematics of the a-CF-GFET sensor showing porous a-C on graphene (left). Inset is the zoomed in image of the pore circled in red. The Raman spectra of the a-CF-GFET sensor showing a broad a-C peak is shown on the right. (b) AFM image showing 38 nm a-C on graphene. (c) TEM image (bright field) showing the fibrous nature of the lithographically cross-linked a-C on the graphene (left), and the more amorphous uncrosslinked a-C on graphene (right). Dark regions are activated carbon, while bright regions are pores leading to graphene. Scale bars in both images are 20 nm. (d) EDX mapping of oxygen around a-C pores on graphene (left) showing minimal oxygen in pores and more oxygen on the surrounding a-C. Top-right is the EDX carbon map of the same region. The bottom Top-right image is the dark field TEM image of the same a-C region on graphene used in EDX mapping. Bright regions are the activated carbon while the dark regions are the pores.

Figure 2. CNP shift of a-CF-GFET sensor in response to ammonia. (a) 500 ppt ammonia in atmospheric air, (b) 1 ppb in atmospheric air (c) 2 ppb ammonia in atmospheric air (d) 5 ppb ammonia in atmospheric air. (e) 200 ppb ammonia in atmospheric air.

iii

Next, we demonstrate that the electric field induced graphene-molecule charge transfer is

retained in the graphene-molecule vdW complex even after the electric field is turned off, (Figure

3a) with the retained charge still unique to the applied electric field magnitude and direction.

Consequently, the vdW bonding memory of adsorbed molecules on graphene was observed (Figure

3a, blue region). This bonding memory enabled the molecular identification of adsorbed gases on

graphene using a proof of concept, the CNPD method (Figures 3b & c), which is a charge transfer

based molecular identification technique that measures the gas adsorption induced charge-transfer

at ∓tuning voltages (∓V

t

).

Figure 3. (a) vdW doping concentration vs time for CO2 adsorption on graphene at Vt = 40 V, 36 V, 0 V, -20 V, -40 V:

white region = tuning voltage applied region, blue region = no tuning voltage applied (vdW bonding memory region). (b) Experimental schematic for the CNPD method (left) and definition of the disparity value for consecutive - Vt and

+ Vt transfer characteristics measurements (right). (c) Dependence of the CNPD value of the a-CF-GFET device to

various gas environments.

The difference between the –V

t

and + V

t

induced charge transfer called the disparity value

(Figure 3b) obtained from the shift of graphene’s charge neutrality point was shown to be

characteristic of various gas environments (Figure 3c), and dependent on the tuning time and

applied tuning voltage.

4. Accomplishment

Publication List

1. Agbonlahor, O.G., Muruganathan, M., Imamura, T., and Mizuta, H. Adsorbed Molecules

as Interchangeable Dopants and Scatterers with a Van der Waals Bonding Memory in

Graphene Sensors. ACS Sensors 5, 2003–2009 (2020).

2. Iwasaki, T., Nakamura, S., Agbonlahor, O. G., Muruganathan, M., Akabori, M., Morita, Y.,

Moriyama, S., Ogawa, S., Wakayama, Y., Mizuta, H., Nakaharai, S. Room-Temperature

Negative Magnetoresistance of Helium-Ion-Irradiated Defective Graphene in the Strong

Anderson Localization Regime. Carbon, 175, 87–92 (2020).

iv

3. Agbonlahor, O.G., Muruganathan, M., Ramaraj, S.G., Wang, Z., Hammam, A.M.M.,

Kareekunnan, A., Maki, H., Hattori, M., Shimomai, K., Mizuta, H. Nanopored

Activated-Carbon Functionalized Graphene Molecular-Sieve for Atmospheric ppt-Level Gas Sensing

and Molecular Identification. Advanced Materials (Under Review)

4. Agbonlahor, O.G., Muruganathan, M., Ramaraj, S.G., Taketomi, K., and Mizuta, H.

Temperature and Pressure Effects on Gas Adsorption-Induced Molecular Dynamics in

Activated Carbon Functionalized Graphene. (In preparation)

Patents

1. 'Sensor devices and their manufacturing method'. Manoharan Muruganathan, Gabriel

Agbonlahor, Amit Banerjee, Hiroshi Mizuta, and 4 other co-inventors. Filing number:

2020-063788 (Filing Date: 31 March 2020)

2. 'Sensor devices, their manufacturing method and gas identification method'

Manoharan Muruganathan, Gabriel Agbonlahor, Amit Banerjee, Hiroshi Mizuta, and 4

other co-inventors. Filing number: 2020-063789 (Filing Date: 31 March 2020)

3. 'Gas identification method and gas identification system'

Manoharan Muruganathan, Gabriel Agbonlahor, and four co-inventors

Filing number: 2019-157961 (Filing Date: 30 August 2019)

Conferences

1. Adsorption Distance Effects on Gas Adsorption Induced Scattering in Graphene Gas

Sensors. Osazuwa Gabriel Agbonlahor, Manoharan Muruganathan, Tomonori Imamura,

Hisashi Maki, Masashi Hattori, Hiroshi Mizuta

2020 International Conference on Solid State Devices and Materials (SSDM2020), 27-30

Sept 2020, Toyama, Japan (Online).

2. Atmosphere-Specific Response in Graphene Field-Effect Transistors

Osazuwa Gabriel Agbonlahor, Manoharan Muruganathan, Hiroshi Mizuta

2020 年 3 月 12‐15 日、第 67 回応用物理学会春季学術講演会、上智大学 四谷キャ

ンパス

.

3. Detection of ammonia via charge transfer-induced shift of graphene charge neutrality

point. Osazuwa Gabriel Agbonlahor, Jothiramalingam Kulothungan, Manoharan

Muruganathan , and Hiroshi Mizuta

Syntheses and Characterizations of Diketopiperazine Monomers and Polyamides

Doctor of Science (Materials Science)

Kaneko Laboratory, 1820004, Hongrong YIN

Part 1: Research Content

1. Background

Chirality and self-assembly are basic attribute of nature and closely related life activity.

Most of biopolymers, such as polypeptides, polynucleotides, and polysaccharides, and

their basic units are chiral. Also, the self-assembly of these biopolymers is important in

life activity, for example, double helix structure of DNA is significant for transcription.

In materials science, chirality and self-assembly of biopolymers also inspire the

development of functional materials, especially in cases of amino acids and polypeptides.

Amino acids are abundant in nature, and also can be produced from biomass. With various

side chains, amino acids can be easily modified, and applied in synthetic polypeptides.

Synthetic polypeptides have been used in in many fields, especially in medical

applications such as nanoparticles, drug delivery system, and tissue engineering.

2,5-Diketopiperazines (DKPs), which are simple cyclic dipeptides, are common in

nature or easily synthesized by the condensation of α-amino acids. Based on two

symmetrical amide groups in a six-membered ring, DKPs have four sites able to form

hydrogen bond, which result in highly molecular arrangement properties. Due to the

strong hydrogen bonding and chiral structure, DKPs have been studied in medicine design,

chiral catalyst, low molecular weight gelators, and annexing agent for polymers. However,

though DKPs have many interesting properties, only a few DKP-based polymers were

reported and the effects of DKP moiety on the corresponding properties of polymers are

not clear. Otherwise, the synthesis of DKP monomer is challenging and important for the

synthesis of polymers.

The aim of this work is to study chirality and self-assembly properties of DKP

monomers and resulting polyamides. This work is divided into three parts: (1) syntheses

and configurational studies of diketopiperazine stereoisomers; (2) syntheses and

solvent-controlled self-assembly of diketopiperazine-based polyamides from aspartame; (3)

syntheses and stereochemistry-property studies of diketopiperazine-based polyamides.

2.

Syntheses and configurational studies of diketopiperazine stereoisomers

To synthesize AB-type DKP monomers, cyclo(aspartyl-4-amino-phenylalanyl)

(ADKP), step-wise protection and deprotection of (

L

/

D

)-aspartic acid and (

L

/

D

)-4-nitro

phenylalanine was performed (scheme 1).

Caused by the solubility problem of precursors

of trans-ADKPs, LD- and DL-ADKP were synthesized in low yield. No racemization

occurred during the synthesis. Structural characterizations and studies of stereoisomers

were performed by

1

H NMR, ROESY, FTIR, and CD. In DMSO solution, configurations

of cis- and trans-ADKP were confirmed, in which DKP ring was in planer structure, and

the folded benzene ring was stabilized by C

β

-H…π interaction in cis-structures and C

α

-H…π interaction in trans-structures. Self-assemble behavior of stereoisomer was studied

by solvent displacement method. LL- and DD-ADKP show similar chrysanthemum-like

morphology, while LD-ADKP shows rose-like morphology. The present study provides

a synthesis method for stereoisomers of DKP, and structural insight of

phenylalanine-aspartic acid-based DKP, which have potential for drug and catalyst design.

Scheme 1 Synthetic route of ADKP stereoisomers form protected aspartic acid and

4-nitro phenylalanine, and structures of ADKP stereoisomers.

3. Syntheses and solvent-controlled self-assembly of diketopiperazine-based

polyamides from aspartame

LL-type ADKP was synthesized from aspartame and subsequently utilized in the

polycondensation of homo-polyamide (PA1) with high molecular weights. By using

various amino acid, dicarboxylic acid, and diamine, random DKP-based copolymers were

also synthesized. The self-assembly properties of ADKP and PA1 were studied via the

solvent displacement method. Notably, PA1 self-assembled into particles with various

morphologies in different solvent systems, such as irregular networks, ellipsoids, and

vesicles (Figure 1). The morphological transformation was also confirmed by dropping

acetone and toluene onto the PA1 particles. Furthermore, infrared spectra and Hansen

solubility parameters of PA1 and different solvents revealed the particle formation

mechanism, which provided more insights into the relationship between the morphology

and strength of the hydrogen bonding of each solvent.

Figure 1 SEM images of particles obtained by the dispersion of PA1 into (a) water, (b)

ethanol, (c) THF, (d) ethyl acetate, (e) acetone, and (f) toluene.

4.

Syntheses and stereochemistry-property studies of diketopiperazine-based

polyamides

Homo-polyamides and co-polyamides from ADKP stereoisomers were synthesized.

Due to same chemical structure, all PAs showed similar molecular weight, thermal

properties and solubility. Determined by CD spectroscopy, all PAs showed optical activity.

Moreover, solvent/third molecule effect on LL-PA was investigated by adding other

solvents into the LL-PA DMSO solution. When water and ethanol were added, LL-PA in

DMSO became optical inactivity. It is suggested that water or ethanol disturbed hydrogen

bond between DKP units in polymer chain, which played an important role in stabilizing

secondary structure of polymer. Self-assembly of PAs were investigated. LL-PA, DD-PA

and LLcoDD-PA self-assembled into vesicles, while LD-PA self-assembled into cubic

structure in toluene and acetone (Figure 2). The present study provides structural insights

of DKP-based polymers with stereochemistry, and reveals their optical and self-assemble

properties.

Figure 2 SEM images of particles obtained by (a) LL-PA in acetone, (b) LL-PA in

toluene (c) DD-PA in acetone, (d) DD-PA in toluene, (e) LLcoDD-PA in acetone, (f)

LLcoDD-PA in toluene, (g) LD-PA in acetone, and (h) LD-PA in toluene.

Part 2: Research Purpose

DKP unit is an important building block for medicines, catalysts, supramolecules, and

has enormous potential for development of functional materials. However, polymers

carrying DKP unit in backbone have not been well explored and are lack of study. In this

study, novel AB-type DKP monomers and relative polyamides with stereochemistry were

successfully synthesized. Chirality and self-assembly properties of obtained monomers

and polyamides were well studied. This study creates a new insight on stereochemistry

and self-assembly of DKP monomers and polymers, which could lead to the development

of medicines and functional materials.

Part 3: Research Accomplishment

Synthesis and solvent-controlled self-assembly of diketopiperazine-based polyamides

from aspartame. Yin, Hongrong; Takada, Kenji;

Kumar, Amit;

Kongprathet, Thawinda

Kaneko, Tatsuo. RSC Advances.

Title: In-situ scanning electron microscopy observation of electrode-electrolyte interfaces in an electrochemical cell

Intended degree: Doctor of Science (Materials Science) Laboratory: Tomitori Laboratory

Student No.: 1820008 Name: He Gada

Part 1: Research Content

Understanding of the dendritic structure growth and stripping mechanism during electrochemical reactions is of great importance to improve the functionality of electrochemical technique devices. Thus, a lot of devoted works had been conducted to uncover the mechanisms qualitatively and quantitatively through combination of typical electrochemical measurements using macro scale electrodes. As an imaging technique during electrochemical reactions, electron microscopies with high resolution meet the desire for both the local structure of electrode and the dendrite formation at nanometer level characterization. For most of the previous works, transmission electron microscopy (TEM) was utilized for in-situ real time observation of metallic crystallization inside the liquid specimen. There have been difficulties like damages due to electron beam irradiation at thigh electric current densities and preparation of liquid specimen with thicknesses of thinner than 100 nm, which is the limitation of transmission of electrons in TEM, and so on. As for the in-situ pseudo real-time scanning electron microscopy (SEM) observation is more practical than TEM observation, because of relatively low current density and no particular requirement for the thickness of the specimen.

We developed a conventional in-situ electrochemical cell (Figure 1) with two electrode terminals for SEM to observe the processes of electro-plating and stripping simultaneously with measuring the cyclic voltammetry.

The cell was sized to fit into the small space of the custom-made specimen holder for Hitachi S-5200 Fe-SEM with

several current feedthroughs, each of which was connected to a coaxial cable. The front plane of the cell-body

possessed a drain on the side of an insulator plate with

Figure 1. (A) Schematic illustration of the exploded view of electrochemical cell. (B) B-1: cell-body, B-2: TEM grid with vapor deposited Au electrode, B-3: cover with O-rings, B-4: form of cell-body and PEEK insulator is installed into SEM holder.

two holes for the cables, as well as a vessel for liquid electrolyte at the center of the insulator plate. The insulator had a circular depressed

seat for a silicon nitrite TEM grid. Two Au wires inserted into Kapton tubes were connected to the two coaxial cables respectively, and inserted through the wire-holes of cell body and insulator; only the Au wires were set to the concentric hole and connected to two vapor-deposited Au electrodes on the one side of the TEM grid. There were two concentric circular seats for O-rings on a cell cover; one was smaller than the silicon nitrate grid and the other one was larger than that. Care was taken so as not for electrolyte to leak to the vacuum environment during SEM experiments. When the applied potential was varied, electrochemical reactions occurred at the end of the vapor-deposited electrode and were observed through a small thin silicon nitrite membrane window of TEM grid with a thickness of 50 nm, through which the SEM electron beam was able to pass.

After inserting the SEM specimen holder assembled with the electrochemical cell into the SEM, the working electrode and the counter electrode of the two Au electrodes were electrically connected to the inputs of a potentiostat/galvanostat instrument (VersaSTAT4, Princeton Applied Research); the inputs for the counter and reference electrodes were connected each together for simple electrochemical measurements. We visualized the electro-plating and stripping

Fig. 2. (a) First shot of time lapse SEM image of

the Au electrode in the CV measurement. The square region, denoted in red, was displayed in the following SEM images in (b)–(f). The full video of the SEM images with a view area of ~100 µm × 100 µm is available at Supplementary data I. (g) Time evolution chart of the potential and current measured with the SEM images in (a)–(f). The broken lines indicate the times when the respective SEM images were started to be taken for 8.4 sec. (h) Cyclic voltammogram plotted from (g).

of Pb dendrites on the Au electrode in an electrolyte of 1.5 M Pb(NO3)2 aqueous solution via in-situ SEM observation (Figure 2).

The bias voltage was swept at a rate of 50 mV/s in a triangular waveform in the range from +2V to –2 V between the working and counter/reference electrodes. First, no obvious reaction was observed when the bias voltage decreased from 2 V to 0 V. From 0 V, small islands started to form at the electrode-electrolyte interface and turned into dendritic structures gradually, there was no obvious change to the measured current until the voltage reached –1.25 V. The dendritic structures rapidly grew with increasing current from –1.25 V, and the current peak appeared at –2 V. After the direction of sweeping voltage was reversed, the dendritic structures drastically dissolved with decreasing current up to –1.25 V. Afterward, the current showed no obvious change whereas the dendritic structure dissolved completely around 1.5 V. With increase in the sweeping rate, the peak value of the current and the redox potential difference increased. When the concentration and composition of electrolyte were changed, the reduction potential shifted toward the negative voltage direction and the oxidation potential shifted toward the positive direction.

We observed in-situ consecutive SEM images of electro-deposition and stripping of Pb on the Au electrode through the SiN membrane window while measuring the cyclic voltammogram. The SEM images clearly showed that the electro-deposition and stripping of Pb on a nanoscale, which were consistent with the cyclic voltammogram curves. From the SEM observation, we distinguish the growth modes of Pd deposits, exhibiting the particulate band structures and dendrite strucutres, depending on the roughness of the electrode. Also the brightness of SEM images in the electrolyte region changed in the course of the cyclic voltammetry, which implied the change of Pb concentration of the electrolyte near the electrode. This study demonstrated the potential of in-situ high-resolution SEM combined with the electrochemical cell.

Part 2: Research Purpose

Most of the previous work for in-situ TEM observation, the current density must be very high when taking high resolution TEM images, which causes H2 bubbling in the liquid electrolyte due to electron beam irradiation. Practically, magnification should be reduced to avoid such irradiation damage, since the current density depends on the magnification. In this sense, I think that in-situ SEM observation seems to be better than TEM, since the current density is relatively low. The sample preparation for in-situ TEM observation of specimens in liquid is time consuming and challenging. The thickness of liquid specimens has to be thinner than 100 nm in order to achieve high spatial resolution imaging. As for the in-situ SEM observation, there is no such requirement for a thin specimen. As concerned with industrial application, nanomaterials such as nanoparticles or nanowires with around 100 nm in size have been synthesized mainly by electrochemical methods. The spatial resolution of SEM is high enough to investigate the electrochemical reaction for nanomaterials of such a size. The equipment for in-situ SEM observation has been commercialized by some companies, but they have some

disadvantage for laboratory research such as high financial cost, experimental limitation and so on. Dendritic structure would appear in electrochemical technique devices like lithium ion battery and fuel cell, it would determine the power density, lifetime of these devices negatively. Through this work, intend to provide a practical experimental method to study the energy storage devices in industry.

Part 3: Research Accomplishment Submitted paper:

He, Gada; Oshima, Yoshifumi; Tomitori, Masahiko

In-situ high-resolution scanning electron microscopy observation of electro-deposition and stripping of lead in an electrochemical cell

Japanese Journal of Applied Physics

Conferences:

1. 14th International conference on Atomically Controlled Surfaces, Interfaces and Nanostructures (ACSIN-14) Dates: October 21 – 25, 2018

Venue: Sendai International Center, Sendai, Japan

Oral presentation: In-situ scanning electron microscopy observation of electro-plating and stripping of lead dendrites in an electrochemical cell

2. The 66th Japan Society of Applied Physics Spring Meeting, 2019 Dates: March 9 – 12, 2019

Venue: Tokyo Institute of Technology, Ookayama Campus, Tokyo, Japan

Oral presentation: Electro-plating and stripping of lead dendrites observed by in-situ scanning electron microscopy with an electrochemical cell

3. The 80th Japan Society of Applied Physics Autumn Meeting, 2019 Dates: September 18 – 21, 2019

Venue: Sapporo Campus, Hokkaido University, Sapporo, Japan

Oral presentation: In-situ scanning electron microscopy observation of lead dendrites grown in an electrochemical cell

4. Annual Meeting of the Japan Society of Vacuum and Surface Science 2019 Dates: October 28 – 30, 2019

Venue: Tsukuba International Congress Center, Tsukuba, Japan

Oral presentation: In-situ scanning electron microscopy observation of electrode–electrolyte interfaces in an electrochemical cell

Abstract

A study of atomic scale mechanics by in situ transmission electron microscopy with a quartz length-extension resonator

Zhang Jiaqi (s1820021) Oshima Lab Doctor of Science (Materials Science)

1. Research Content Background

The mechanical properties of atomic-scale materials are novel not only for development of nanoelectromechanical systems (NEMS) such as sensors but also for development of novel functional devices that utilize the fact that the physical and chemical properties of nanocontacts are sensitive to structure and shape. It has been mainly investigated by theoretical calculation, and the size effect or crystal orientation dependence have been pointed out such that Young’s modulus depends on the size due to surface effect below 10 nm in size. However, such dependence has not been clarified in the experiment. However, no consensus has been obtained due to problems such as different results being obtained depending on the pseudopotential used in the calculation.

Aim

In this thesis, I developed a TEM holder incorporating a length extension resonator (LER) and applied frequency modulation method to the LER in order to directly measure spring constants of nanomaterials with high accuracy. Fig. 1 (A) is an overview view of the developed holder. Since the LER has a high Q factor, the amplitude of the oscillator required to measure the equivalent spring constant is only about 20 pm, which is a big advantage because the plastic deformation of the measured nanomaterial can be avoided and atomic resolution image can be obtained. Therefore, it is possible to measure the equivalent spring constant of the metal nanocontact with high accuracy. Fig. 1 (B) shows time evolution of electrical conductance and equivalent spring constant of a Pt nanocontact in the process of slowly stretching it until breaking, and TEM images taken from the video.

As mentioned above, the structure, conductance, and equivalent spring constant of nanocontacts could be obtained using the developed TEM holder. From these data, we aim to investigate the mechanical properties of the nanomaterials such as the Young’s modulus.

Figure 1. (A) design of the TEM holder, (B) Conductance and spring constant trace and snapshot TEM images during the stretching process of a Pt nanocontact.

Experimental results

Using our developed holder, we precisely measured the spring constant (stiffness) of Pt monatomic chains have 2-5 atoms as shown in figure 2(A). And the individual bond stiffness in tip-suspended platinum (Pt) monatomic chains was determined by statistically analyzing the stiffness of about 150 Pt monatomic chains consisting of 2–5 atoms. The single bond stiffness at the middle of chain was estimated to be 25 N/m, while that of the bond to the suspending tip was 23 N/m. These stiffness values are obviously higher than the bulk counterpart of 20 N/m. The chain spring kept the constant stiffness during stretching process, similarly to the quantized conductance. In addition, the bond length at equilibrium and at maximum breaking point has been estimated to be 0.25 nm and 0.31 nm, respectively, in the monatomic chain. It indicates that the maximum elastic strain of one bond in the chain is as large as 24%. These characteristics including the breaking point could be explained by the configurations minimizing the calculated string tension.

In addition, in order to clarify the size effect on the Young’s modulus, the Au nano junctions with the axis of the [111] direction was evaluated below 3 nm in diameter by our developed TEM holder. The nanocontacts became thinner under stretch while the apex angle of the pyramidal shape kept constant. I found that the volume of the nanocontact was not constant in such deformation process, indicating and the atoms migrated on the surface around the contact area. The Young’s modulus, which was calculated from the spring constant and corresponding geometrical information, was found to decrease when the size was reduced below about 2 nm in diameter. Such size dependence of the Young’s modulus was analyzed by assuming that the Young’s modulus of the surface atom layer, the sub-surface atom layer and the core in Au [111] nanocontact were different. I found that the Young’s modulus of the surface atom layer exhibited very low and one of the core,

the similar value to one of bulk [111] Au crystal. This analysis indicates that the surface atom layer is softer, which is in consistent of the observed surface atom diffusion. The size dependence of the Young’s modulus can be explained by surface effect.

Figure 2. (A) histogram of atomic chains with 2-5 atoms, (B) the variation of Young’s modulus with diameter of the Au nanodisks. The red dots are experiment result and the blue dots is the fitting line use the core shell model with equation (9), the insert shows the atomic model.

2. Research significance

Nanoscale or atom-scale materials are well known to show various physical or chemical properties due to quantum confinement effect or surface effect. However, most studies of such nanoscience have been based on theoretical calculation. Since the physical or chemical properties in nanoscale are sensitive to the atomic structures, the confirmation of the atomic structure is required at the same time as measuring the physical or chemical properties. But such method has been rarely established. In this study, I developed in-situ TEM holder for understanding the elastic properties of Au and Pt nanocontacts. The originality of our experimental equipment is measuring the spring constant precisely by using quartz length extension resonator (LER). By LER, I confirmed that the equivalent spring constant with about 1 N/m in accuracy was obtained based on frequency modulation method with an extremely small oscillation amplitude of about 30 pm, indicating that the spring constant of atomic bond is measurable. I believe that our experimental equipment is the most suitable for estimating the mechanical properties of atom-scale materials in the world.

The most important results in my study are that the spring constants of gold or platinum nanowires are obtained precisely. I estimated the individual atom bond stiffness in platinum monatomic chains, which is obviously higher than the bulk counterpart. Also, I clarified the size and crystal orientation dependence of Young’s modulus for gold and platinum nanowires or

nanocontacts in this study which has been controversial for a long time and confirmed that Young’s modulus of gold nanocontacts gradually decreased below 2 nm in diameter due to surface effect. Such new characterization will open a new field of “atom-mechanics”, leading to atom-scaled electromechanics and chemicalmechanics.

3. Research Accomplishment

• List of Publications

1. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, “Atomic scale mechanics explored by in situ transmission electron microscopy with a quartz length-extension resonator as a force sensor” Nanotechnology. 2020, 31, 205706.

2. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai, Kenta Hongo, Ryo Maezono, Erio Tosatti, Yoshifumi Oshima, “Atom-mechanics in platinum monatomic chains.” (submitted) 3. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai, Yoshifumi Oshima, “Size

dependent young’s modulus of sub-2nm gold nanocontacts” (in preparation)

4. Chunmeng Liu , Jiaqi Zhang, Muruganathan Manoharana, Hiroshi Mizuta, Yoshifumi Oshima, Xiaobin Zhang, “Origin of nonlinear current-voltage curves for suspended zigzag edge graphene nanoribbon.” Carbon. 2020, 165, 476-483.

5. Chunmeng Liu , Jiaqi Zhang, Xiaobin Zhang, Muruganathan Manoharana, Hiroshi Mizuta, Yoshifumi Oshima, "In-situ electrical conductance measurement of suspended ultra-narrow graphene nanoribbons observed via transmission electron microscopy", Nanotechnology 32, 025710 (2020) (8 pages).

• Presentation

1. Jiaqi Zhang, Yuki kobori, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, "Mechanical Properties of Pt Nano-contacts Measured by TEM Combined with a Frequency-Modulation Force Sensor", 10th International Conference on Materials for Advanced Technologies (ICMAT) No. 191177, 25th, June, 2019, Singapore.

2. Jiaqi Zhang, Yuki kobori, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, "Mechanical Properties of Metal Nano-contacts Measured by TEM Combined with Frequency Modulation Force Sensor" (7P-005), Asian Conference on Nanoscience and Nanotechnology 2018 October 18-21th, Qingdao (China)

3. Jiaqi Zhang, Yuki kobori, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, “Mechanical properties of Pt nano-contacts measured by TEM combined with a frequency-modulation force sensing system” The 66th JSAP spring meeting, 9p-PA5-3, 9th, March, 2019, Tokyo Institute of Technology

Oshima, "Mechanical properties of Pt nano-contacts measured by TEM combined with frequency modulation force sensor" The 74th Annual Meeting of the Japanese Society of Microscopy, P-M_48, 8th, June, 2019, Nagoya convention center

5. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, "Mechanical properties of Pt atomic chain measured by TEM combined with a frequency-modulation force sensing system" The 80th JSAP autumn meeting, 19th, September, 2019, 19a-C310-4, Hokkaido University

6. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima "Mechanical properties of Pt atomic chain measured by TEM combined with a frequency-modulation force sensing system" Annual meeting of the Japan Society of Vacuum and Surface Science 2019, 3Ea05S, 31th, October, 2019, Tsukuba

7. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, "Mechanical properties of Pt-Pt bond in an atomic chain measured by TEM combined with a frequency-modulation force sensing system" The 67th JSAP Spring meeting, 13th, March, 2020, 13p-D519-1~12, Sophia University

8. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima, "Clarification of Fe ordering in FexTiS2 structure using transmission electron microscopy" The 76th Annual Meeting of the Japanese Society of Microscopy, 2020, 3pmB_S11-6, On-line. 9. Jiaqi Zhang, Keisuke Ishizuka, Masahiko Tomitori, Toyoko Arai and Yoshifumi Oshima,

"Atomic scale mechanics measured by TEM holder combined with a frequency-modulation force sensor", The 81st Autumn Meeting of the Japan Society of Applied Physics, 2020, 10p-Z06-9 C001981, On-line

• Grant

1. “A study of mechanical properties of Pt atomic contacts by transmission electron microscopy” Research Report Form for JAIST Research Grant (Fundamental Research) 2019

1,300,000 JPY

2. 「周波数変調型力センサーを組み込んだ TEM 観察法による金や白金ナノメカニック

スの解明」(2020-2006)、2020 年度笹川科学研究助成、65 万円

("Elucidation of gold and platinum nanomechanics by TEM observation method incorporating frequency modulation type force sensor" (No.2020-2006), 2020 Sasakawa Scientific Research Grant, 650,000 JPY)

3. 「原子スケール材料のその場 TEM 計測による弾性特性の研究」、一般財団法人 丸文

財団 第２４回（令和２年度）交流研究助成金、100 万円

("A study of elastic properties of atomic scale materials by in-situ TEM measurement", The 24th Marubun Research Foundation (2020) Exchange Research Grant, 1,000,000JPY)

1 論文題目： 繊維状フィラーを利用した高分子複合材料の高機能化 希望取得単位： マテリアルサイエンス 研究室名： 山口政之研究室 学生番号： 1820025 氏名： 西川理穂 1. 研究内容 本研究では、繊維状フィラーを用いた結晶性高分子の分子配向制御と結晶構造の詳細の解明、 および非相溶ポリマーブレンド中のナノフィラーの分散制御を目的とした。以下に、各章の概要を 簡単にまとめる。 (1) 高分子溶融体中におけるカーボンナノチューブのブラウン運動とネットワーク構造 (第 2 章) 高分子溶融体中における多層カーボンナノチューブ (MWCNT) の、ブラウン運動による配向 緩和をレオロジー測定、走査型電子顕微鏡による観察、電気抵抗率の測定により評価した。熱処 理中、ブラウン運動により MWCNT の粒子間相互作用の寄与が高まり、溶融状態の動的弾性率は 時間とともに成長した。この弾性率成長は、低い温度よりも高い温度において顕著であった。ブラウ ン運動によって形成された MWCNT のネットワーク構造は、効率的な導電パスである。したがって、 より高い温度で成形した MWCNT 含有複合材料は、より高い導電性を示した。 本章の結果は、熱処理温度と時間を制御することにより試料中のナノフィラーのネットワーク構造 や配向のコントロールが可能であることを示す。 (2) カーボンナノチューブからの結晶化を利用した高剛性材料の設計 (第 3 章) MWCNT の添加が流動場において高密度ポリエチレン (HDPE) の分子配向に与える影響、お よび HDPE の結晶構造の詳細を調べた。HDPE にごく少量の MWCNT を添加し、キャピラリーレ オメータにて押出成形することで試料を得た。MWCNT 無添加の HDPE ではほとんど配向を示さ ない条件において、わずか 0.1 wt.%の MWCNT 添加で HDPE 鎖は分子配向 (シシカバブ構造) を示すことが明らかになった。これは、流動方向に配向した MWCNT が HDPE の結晶核剤として 働くことにより、冷却過程で MWCNT が HDPE 鎖のシシとして作用し、結晶化速度を速めたことに 起因する。さらに、配向度の向上に伴い引張弾性率も飛躍的に増加した。 本成果は、ごく少量の繊維状結晶核剤を添加することで、汎用的な成形条件にて高分子鎖の 分子配向を大きく向上できることを示唆している。 (3) 共連続ポリマーブレンド中におけるカーボンナノチューブの界面局在化 (第 4 章) ポリカーボネート (PC) /ポリエチレン (PE) 共連続相の界面に MWCNT を局在化させることで、

2 ポリマーナノコンポジットに良好な導電性を与える手法を提案した。PE 成分として超高分子量 PE (UHMWPE) を用いることで PE 相の粘度を高く設定したところ、MWCNT は PC 相と UHMWPE 相 の界面に局在化した。これは、当初 PC 相に存在した MWCNT が混練中に UHMWPE 相へと移 行しようとするが、MWCNT の低い拡散係数のため UHMWPE 相中には入り込めなかったことに起 因する。さらに、本ブレンドは共連続構造を形成するため、MWCNT は効率的な導電パスとして機 能し、試料は良好な導電性を示した。 この章では、濡れ係数を用いた予測では不可能であったフィラーの界面局在を実現した。した がって、これらの成果は、ナノフィラーを用いた導電性ポリマーブレンドの新たな知見であり、材料 設計の幅を広げると考えられる。 (4) 高分子溶融体中における繊維形成を利用したポリビニルアルコール繊維強化プラスチックの 設計 (第 5 章) 溶融ポリプロピレン (PP) 中におけるポリビニルアルコール (PVA) の繊維化を達成するために、 以下の二つの方法を試みた。一つ目は、PP と低分子量の PVA を溶融混練した後、溶融延伸を行 う方法である。二つ目は、PVA 水溶液を溶融 PP に添加し脱気を行いながら二軸押出機にて混練 する方法である。両方法ともに、繊維状の PVA を得ることができ、さらに、PVA は PP に対して結晶 核剤能力を示すことが判明した。続いて行われた射出成形過程において、配向した PVA 繊維に 沿って PP 鎖が結晶化することにより、成形後の試験片は高い分子配向度と力学特性を示した。 PVA 繊維はガラス繊維よりも軽量であり、さらに近年着目されているカーボンナノチューブやセ ルロースナノファイバーよりも安価である。したがって、本手法は自動車部品などへの応用が期待さ れる。さらに、本技術は他の高分子へも応用可能と考えられるため、汎用性が高い。 2. 研究の意義 プラスチック成形加工は、そのほとんど全てにおいて溶融時の流動場を利用して行われている。 高分子複合体における流動条件はフィラーの分散状態に影響し、さらに繊維状フィラーを用いた 際はその配向の異方性にも影響を及ぼす。これらは成形後の性能や物性に大きく関係する。近年、 ナノスケールのフィラーの複合化研究が盛んであるにもかかわらず、高分子溶融体中におけるナノ フィラーの挙動を詳細に研究した例はほとんどない。本研究では、熱処理温度やマトリクスポリマー の粘度に着目し、MWCNT のブラウン運動を制御することで効率的にネットワーク構造やパーコレ ーションを得る方法を提案した。繊維状ナノフィラーの分散状態および配向は、流動場による影響 とブラウン運動による拡散の競合によって決定されるためである。また、ナノフィラーは、高分子鎖 のサイズに近いためマトリクスポリマーの構造に影響を与えやすい。しかし、複合材料中における マトリクスの結晶構造の詳細など未だ不明な点が多い。本研究では、MWCNT を結晶核 (伸長鎖

3 結晶／シシ) とした HDPE の構造を詳細に調査した。これらの成果は、ナノコンポジット設計の有 用な知見となり得る。 また、工業的な応用には、軽量かつ原料が汎用性に優れた材料が必要とされる。そのような繊 維として、例えばPVA 繊維が挙げられる。しかし、PVA のような水溶性高分子の、汎用高分子中に おける均一分散は困難とされてきた。本研究では、分散相である PVA の粘度に着目することによ り、溶融マトリクス中におけるPVA の繊維化および均一分散を実現した。 本研究で得られた成果が、次世代の材料設計や成形加工技術の向上に役立つこととなれば幸 いである。 3. 研究業績 以下に研究業績を記す。なお、学会発表に関しては主要なものを抜粋して記載する。 原著論文 (査読有り)

1. Nishikawa, R.; Yoon, H.; Yamaguchi, M.

“Rheological evaluation of carbon nanotube redistribution in polymer melt”

Journal of Society of Rheology Japan (Nihon Reoroji Gakkaishi) 2019, 47 (3), 105–110.

2. Nishikawa, R.; Yamaguchi, M.

“Effect of carbon nanotube addition on structure and properties for extrudates of high-density polyethylene”

Journal of Applied Polymer Science 2019, 136 (40), 48010.

3. Nishikawa, R.; Tamaki, K.; Notoya, O.; Yamaguchi, M.

“Carbon nanotube localization at interface in co-continuous blends of polyethylene and polycarbonate”

Journal of Applied Polymer Science 2019, 137 (10), 48676.

4. Nishikawa, R.; Aridome, N.; Ojima, N.; Yamaguchi, M.

“Structure and properties of fiber-reinforced polypropylene prepared by direct incorporation of aqueous solution of poly(vinyl alcohol)”

Polymer 2019, 199, 122566.

学会発表

(国際学会ポスター発表、査読有り) 5. Nishikawa, R.; Yamaguchi, M.

“Effect of CNT addition on structure of HDPE”

4 Chiang Mai, Thailand, December 2018. 6. Nishikawa, R.; Yamaguchi, M.

“Impact of carbon nanotube addition on structure of high-density polyethylene”

Annual Technical Conference of Society Plastics Engineers 2019, Detroit, America, March 2019. 7. Nishikawa, R.; Yamaguchi, M.

“Structure change of high-density polyethylene by addition of carbon nanotube” European Polymer Congress 2019 Crete, Greece, June 2019.

(国内学会口頭発表、査読無し) 8. 西川理穂; 山口政之 カーボンナノチューブ添加による高密度ポリエチレンの配向状態制御 第64 回 レオロジー討論会, 彦根, 2019 年 10 月 9. 西川理穂; Panitha Phulkerd; 山口政之; 小島直紀; 有留憲文 ポリビニルアルコール水溶液添加による高性能ポリプロピレンの設計 プラスチック成形加工学会 第 27 回 秋季大会, 高松, 2019 年 11 月 10. 西川理穂; 山口政之; 小島直紀; 有留憲文 ポリビニルアルコール水溶液の添加により調製したポリプロピレン系コンポジットの構造と物性 第69 回 高分子討論会, オンライン, 2020 月 9 月 11. 西川理穂; 山口政之; 小島直紀; 有留憲文 水溶液添加法により調製したポリプロピレン/ポリビニルアルコールの構造と物性 第65 回 レオロジー討論会, オンライン, 2020 年 10 月 12. 西川理穂; 山口政之 カーボンナノチューブを利用したポリエチレンの配向制御 プラスチック成形加工学会 第 28 回 秋季大会, オンライン, 2020 年 12 月 (国内学会ポスター発表、査読無し) 13. 西川理穂; 山口政之 ポリマー溶融体中におけるカーボンナノチューブのネットワーク形成 プラスチック成形加工学会 第 26 回 秋季大会, 浜松, 2018 年 11 月 14. 西川理穂; 山口政之 カーボンナノチューブ添加による高密度ポリエチレンの構造変化 第68 回 高分子学会 年次大会, 大阪, 2019 年 5 月 受賞等 15. 日本学術振興会 特別研究員 DC2 採用 (令和 2 年度)

Ab initio assessment of computational

thermodynamics applied to magnetic alloys

School of Materials Science

Maezono Laboratory

Adie Tri Hanindriyo

1820027

Research Content

Permanent magnets are widely used in society, from small-scale, everyday applications (fridge

magnets, paper holders) to large-scale, industrial needs (electric generators, motors). The rare

earth compound Nd

2

Fe

14

B is well-known as the base material for the strongest permanent

magnets available today, colloquially known as “neodymium magnets”.

1

This is due to the high level

of magnetization (strength of magnetic field) and coercivity (resistance to demagnetization by

external field) exhibited by this ternary rare earth compound.

Even so, improvements upon the base

material are often desired to fit a variety of

applications. One famous example is the

effort to improve upon the Curie

temperature (T

c

), a thermodynamic property

describing the temperature at which

magnetic materials lose their magnetic

ordering. Increasing T

c

is desirable for

applications in which permanent magnets

face a high working temperature, which can

compromise magnetic field strength. An

example of this effort is the alloying of

Nd

2

Fe

14

B with cobalt (Co), which substitutes

several Fe sites with Co (Nd

2

Fe

14-x

Co

x

B).

2

Further alloying with dysprosium (Dy),

substituting Nd sites, is made to make up for

some coercivity lost upon alloying with Co.

3

However, due to dependence on China for global Dy production, it is desirable to find alternative

ways to improve upon the Nd

2

Fe

14

B compound.

Atomic substitution and alloying for property control requires proper information on phase

1 _{Sagawa, et al. Jpn. H. Appl. Phys. 26 785 (1987)} 2 _{Bolzoni, et al. J. Appl. Phys. 61 5369 (1987)} 3 _{Chen, et al. J. Magn. Magn. Mater. 261 222 (2003)}

Figure 1 Crystal structures of Nd-Fe-B ternary system

constituent compounds investigated in this work, (a) NdB6,

stability, which is embodied within

the phase diagram. Phase diagrams

show the conditions (temperature,

pressure, etc.) on which certain

phases can be stably realized. The

computational

thermodynamics

method CALPHAD (CALculation of

PHAse Diagrams) is often employed

to collate meaningful experimental

information on constituent phases

and compile the information to draw

the phase diagram. The core of

CALPHAD is the mathematical

modelling of Gibbs energy of

constituent phases in a system, from

which reliable predictions of phase

transitions can be made (leading to

the complete phase diagram). In

order to reliably model the Gibbs

energy, accurate information of

thermodynamic properties is in turn

required, which is why information on thermodynamic properties of constituent phases is highly

valuable for CALPHAD researchers.

Several CALPHAD assessments of the Nd-Fe-B ternary system have been performed,

456789

but

the lack of experimental information on thermodynamic properties of Nd-Fe-B constituent phases

have limited the reliability of the Gibbs energy models used within these assessments. While the

constituent Fe-B binary system is relatively well-documented (due to its role in hardened steel

alloys), the Nd-B, Nd-Fe, and Nd-Fe-B systems contain phases for which experimental data is not

readily available. In order to compensate for this fact, this work attempts to use ab initio methods

to calculate and predict relevant thermodynamic information of the constituent phases of the

Nd-B, Nd-Fe, and Nd-Fe-B systems (shown in Figure 1) for use as input data for CALPHAD.

Two primary thermodynamic properties, enthalpy of formation (ΔH

f

) and specific heat in

constant pressure (C

p

), corresponding to the enthalpy and entropy terms of the Gibbs energy

model, respectively, are calculated using the ab initio Density Functional Theory (DFT) with the

Generalized Gradient Approximation (GGA) as a conventional choice of exchange-correlation

functional. It is observed that, likely due to the incomplete self-interaction error well-known to

occur with conventional exchange-correlation functionals and localized electron orbitals (3d/4f

4 _{Matsuura, et al. Jpn. J. Appl. Phys. 24 L635 (1985)}

5 _{Hallemans, Wollants, and Roos. J. Phase Equilibria 16(2) 137 (1995)} 6 _{Raghavan. J. Phase Equilibria 24(5) 451 (2003)}

7 _{Van Ende and Jung. J. Alloys Compd. 548 133 (2013)} 8 _{Zhou, Luo, and Zhou. J. Electron. Mater. 45(1) 418 (2016)} 9 _{Chen, et al. Calphad 66 101627 (2019)}

N d 2_B 5 N dB 4 N dB 6 N dB 66 1127.8 1279.0 2272.9 2880.6 2673.6 1388.0 2395.1 2348.0 T [ K ] 500 1000 1500 2000 2500 3000 X(Nd) 0 0.2 0.4 0.6 0.8 1

Figure 2 Calculated phase diagram of the binary Nd-B system in this

orbitals in transition metals and rare earth compounds), the enthalpy of formation calculated for

several phases possess significant discrepancy with experimental results. To address this issue, the

Hubbard U correction according to the ab initio framework of Cococcioni and de Gironcoli

10

was

employed, which reduced the discrepancy to a tolerable level.

Phonon calculations using the finite difference frozen phonon approximation was used to

calculate phonon frequencies of phases in the harmonic approximation to calculate specific heat

in constant volume (C

v

). Subsequently, the quasi-harmonic approximation was employed to

introduce volume dependence and calculate C

p

. The enthalpy of formation and specific heat

information was then used in subsequent CALPHAD method to reassess the phase diagram of the

binary Nd-B system, one of the important constituent systems of the ternary Nd-Fe-B system.

Figure 2 shows the end result of CALPHAD reoptimization described in this work for the binary

Nd-B system.

Research Purpose

The DFT-CALPHAD framework employed in this work is a recent possibility explored to

complement experimental and empirical data, with several important advantages. One of the most

important of which is the investigation of microstructures and phases which might be very costly

(in both time and energy) to synthesize, when ab initio methods can reliably provide an accurate

prediction at a fraction of the cost. This is the first such work to apply the DFT-CALPHAD framework,

and also the DFT+U technique within the same framework, to the Nd-Fe-B system, and the

intriguing results discovered suggest that this will also not be the last application of this framework

for this system, as it is a flexible framework which can be used to investigate quarternary and

quinary systems as well (for example, the Nd-Fe-Co-B quarternary and Nd-Dy-Fe-Co-B quinary

systems relevant for Nd

2

Fe

14-x

Co

x

B phase and (Nd,Dy)(Fe,Co)B magnet materials, respectively).

The resulting binary Nd-B phase diagram is a more reliable prediction of phase stability as a

result of incorporating ab initio data to the Nd-Fe-B database. Significant improvement was made

specifically toward the Nd-rich region of the composition space, particularly relating to the liquidus

line which separates the liquid phase from a multitude of stable solid phases. As a particularly

model-dependent area of the phase diagram, the liquidus line of the Nd-B phase diagram has seen

substantial changes across multiple reassessments relying on a relatively small amount of

experimental thermodynamic data. By compensating for this lack of input information with ab

initio assessments of the Nd-B constituent phases, this work has more reliably predicted the

liquidus line of the Nd-B phase diagram.

In particular to the Nd

2

B

5

system, this work has also shown that the discrepancy of ab initio DFT

calculations with respect to experimental data (as found in preceding work of Colinet and

Tedenac

11

) is not improved by the use of the Hubbard U correction, and is likely caused by more

than the suspected self-interaction error. As this discrepancy is also noted for the isomorphic Gd

2

B

5

compound, it is likely that an aspect of the crystal structure is improperly described by the GGA

exchange correlation functional.

10 _{Cococcioni and de Gironcoli. Phys. Rev. B 71 035105 (2005)} 11 _{Colinet and Tedenac. Calphad 62 49 (2018)}

Research Accomplishment

Peer reviewed

[1] A. T. Hanindriyo, TB. M. Y. Y. Prawira, M. K. Agusta, and H. K. Dipojono. “Computational Design

of Ni-Zn Based Catalyst for Direct Hydrazine Fuel Cell Catalyst using Density Functional Theory”.

Procedia Engineering

170

148-153 (2017)

[2] A. T. Hanindriyo, S. Sridar, K. C. Hari Kumar, K. Hongo, and R. Maezono. “Ab initio

thermodynamic properties of certain compounds in Nd-Fe-B system”. Computational Material

Science

180

109696 (2020)

[3] A. Hermawan, A. T. Hanindriyo, E. R. Ramadhan, Y. Asakura, T. Hasegawa, K. Hongo, M. Inada,

R. Maezono, and S. Yin. “Octahedral morphology of NiO with (111) facet synthesized from the

transformation of NiOHCl for the NO

x

detection and degradation: experiment and DFT

calculation”. Inorganic Chemistry Frontiers

7

3431-3442 (2020)

[4] T. Hasegawa, A. Shigee, Y. Nishiwaki, M. Nagasako, A. T. Hanindriyo, K. Hongo, R. Maezono, T.

Ueda, and S. Yin. “New layered perovskite family built from [CeTa

2

O

7

]

-

layers: coloring

mechanism from unique multi-transitions”. Chemical Communications

56

8591-8594 (2020)

[5] A. Raghav, A. T. Hanindriyo, K. Utimula, M. Abbasnejad, R. Maezono, and E. Panda. “Intrinsic

electronic defect states of anatase using density functional theory”. Computational Materials

Science

184

109925 (2020)

Non peer reviewed

[1] A. T. Hanindriyo, A. K. S. Yadav, T. Ichibha, R. Maezono, K. Nakano, and K. Hongo. “Diffusion

Monte Carlo evaluation of disiloxane linearization barrier”, arXiV:2010.06767

[physics.chem-ph] (2020)

Studies on CuPd Bimetallic Catalysts for

Selective Hydrogenation of Succinic Acid

Doctoral Degree Nishimura’s Laboratory

Le Dinh Son (1820036)

1

Research Content

1.1

Research background and objective

Succinic acid (SA) was identified as one of the most potentially bio-derived platform chemicals which can be converted to various value-added products. Hydrogenation is by far the most investigated transformation due to the impor-tance of its products including γ-butyrolactone (GBL), tetrahydrofuran (THF), and 1,4-butanediol (BDO). However, the selective hydrogenation of SA is gener-ally a challenging reaction due to the low electrophilicity of the carbonyl group and the complexity in its reaction pathways, which have provided a strong spur for chemists to design effective catalysts for this transformation. Despite that, the heavy dependence on precious metals such as Pd, Pt, Re, Ir, Ru, and Rh in previously reported catalysts is economically disadvantageous, which possi-bly limits them from industrial applications. Therefore, the studies embodied in this thesis aim to develop efficiently earth-abundant metal-based bimetallic catalysts for selective hydrogenation of SA to GBL, BDO, and THF (Figure 1).

Figure 1: Selective hydrogenation of SA to GBL, BDO, and THF

1.2

Results and discussion

In the initial attempt to search for a suitable catalyst system, hydroxyapatite (HAP) supported CuxPdy (x+y = 10 wt%) was found to be potential bimetallic catalysts for the production of BDO from SA. The effect of metal ratio was ex-amined and the Cu8Pd2/HAP was found to be the best catalyst, affording a high selectivity of BDO (>80%) at a quantitative conversion of SA. A strong Cu–Pd interaction resulted from alloying formation led to an enhanced catalytic activity to the intermediate GBL, compared to that over the Cu10/HAP monometallic catalyst. While on the other hand, the Cu-rich CuPd nanoparticles (NPs) sup-pressed the over-reactivity of Pd, preventing the side reaction to butyric acid (BA), which is typically encountered in the Pd10/HAP monometallic catalyst. Subsequently, the Cu species that existed adjacently to CuPd alloying NPs pro-moted further hydrogenation of GBL toward BDO.

The influences of various supports, i.e., SiO2, TiO2, and γ-Al2O3, on the constructions of Cu-rich CuPd alloy nanoparticles (NPs) were then investi-gated. In-depth characterizations revealed that randomly homogeneous CuPd NPs were prevalently constructed on TiO2and SiO2, whereas the heterogeneous CuPd alloy NPs with a great extent of Cu segregation were dominantly formed on γ-Al2O3. Accordingly, the activity and selectivity are distinctly different among these catalysts. Particularly, a selectivity of GBL (90%) can be attained over the CuPd/TiO2at 73% conversion of SA, which was attributed to the pres-ence of large CuPd NPs preventing further hydrogenation of GBL and lowering the catalytic activity. On the other hand, higher activity and selectivity toward BDO of CuPd/SiO2 were ascribed to its small CuPd NPs and the presence of isolated Cu species which promoted the formation of BDO at a high yield of 86%. Notably, the strong Lewis acid sites in the CuPd/γ-Al2O3 was revealed as the decisive factor in the formation of highly selective THF with 97% at a quantitative conversion of SA.

To broaden knowledge in the γ-Al2O3 supported CuPd catalysts, the influ-ence metal ratio on the catalytic performance has been extended. Excellent cat-alytic performance toward THF was achieved over the Cu-rich Cu6Pd4/γ-Al2O3 and Cu8Pd2/γ-Al2O3 catalysts, achieving the product yield and selectivity of 85–90%. In addition, the present catalyst can maintain its high activity and selectivity for several recycling runs under high temperature and pressure con-ditions. The strong interaction in CuPd alloy NPs resulted in the enhanced reactivity compared to that of the monometallic Cu, while the Cu-rich com-ponent helped to restrain the strong reactivity of Pd species which favors the formation of BA. Subsequently, the Cu-rich CuPd NPs promoted the formation of the intermediate BDO which was easily converted to THF via cyclodehydra-tion under the influence of strong Lewis acid sites in the support γ-Al2O3.

Finally, the influence of the capping agent on the catalytic performance of CuPd NPs was studied. A highly efficient PVP-capped CuPd NPs constructed on HAP was discovered for selective hydrogenation of SA to GBL. The inhi-bition effect of the capping agent PVP was revealed to play a key role in the formation of GBL with excellent selectivity. The catalyst was able to proceed at extremely low hydrogen pressure from 1 MPa while maintaining high selectivity of GBL (>90%). Besides, the catalyst showed remarkable reusability, offering the catalyst with enormous potential for industrial applications.

2

Key findings and contributions

2.1

Tunable catalysts for selective SA hydrogenation

The importance of SA hydrogenation which generates the key intermediates including GBL, BDO, and THF for the polymer industry is unquestionable. The discovery of CuPd bimetallic catalysts in the present research, therefore, provides not only cheaper catalysts but also effective and feasible methods to tune the product selectivity for the SA hydrogenation (Table 1). Furthermore, in terms of productivity, the present catalyst system showed excellent perfor-mances, which closely approach or even greatly superior to the highest records for BDO or THF and GBL productions so far. On the other hand, in the con-text of reaction conditions, the HAP supported PVP-capped Cu40Pd60catalyst

promoted the SA hydrogenation to highly selective GBL even at low hydro-gen pressure, from 1 MPa. In addition, the extremely low concentration of metal loading (0.1 mmol) in combination with the excellent reusability made this catalyst suitable for flow reactor and thus open up a possibility for indus-trial applications.

Table 1: Catalytic hydrogenation of SA

Catalyst Temp. /℃ P H2 /MPa Conv. /% Selectivity /% BDO GBL THF Cu8– Pd2/HAP or SiO2 200 8 100 82–86 <5 ≤3 Cu8– Pd2/γ-Al2O3 200 8 100 0 0 97 Cu8– Pd2/rutile-TiO2 200 8 73 2 90 3 Cu40– Pd60– PVP/HAP 200 1–4 68–92 0 92–96 0 Ir – Re/C 240 8 100 <5 0 75 Pd – Cu/AX 190 7 72 0 94 0 Pd/Al2O3 170 3 <70 <1 95 0 Re – Ru/C 160 8 99 70 6 6 Re3 – Ru/C 240 8 99 3 5 60 Pd – Re/AC 180 10 100 67 0 14 Re – Ru/BMC 200 8 100 65 34 3 Re – Pd/SiO2 140 8 100 89 3 0 Pd – 5 FeOx/C 200 5 100 70 - -Pd – Re/C 240 8 89 4 - 73 Pd – Re/TiO2 160 15 100 83 0

-The red color texts indicate the present study, while the remaining were reported in the open literature in the last decade.

2.2

Mechanism of SA hydrogenation

(1) SA→ GBL: According to the catalytic performances of monometallic catalysts, both Cu and Pd can reduce SA to GBL. However, Cu catalysts showed poor catalytic performance for SA hydrogenation which can be attributed to the limited capability of Cu for H2 activation. On the other hand, Pd holds a great potential for H2 dissociation, which turned out not to be suitable for SA hydrogenation since it directed the reaction to the side product BA. The strong Cu–Pd alloying formation resulted in better GBL yield in the Cu8Pd2/TiO2 and Cu40Pd60– PVP/HAP catalysts. The reactions were terminated at GBL without further hydrogenation because the absence of isolated Cu and/or large CuPd NPs of the former catalyst and the inhibition effect of PVP in the latter. (2) SA → BDO: The reaction using GBL as a starting material showed that Cu can reduce GBL to afford BDO with excellent yield and selectivity. Therefore, the presence of isolated Cu in Cu8Pd2/HAP and Cu8Pd2/SiO2 can further reduce the intermediate GBL to BDO with high yield and selectivity. It should be emphasized that the close interaction between Cu and Pd is impor-tant because the physical mixture of Cu8/HAP and Pd2/HAP were unable to produce BDO with the comparable yield to that over the bimetallic catalyst.

(3) SA → THF: The reaction using BDO as starting material revealed that strong Lewis acid sites in γ-Al2O3 is one of the crucial factors for the production of THF. However, CuPd alloy is the prerequisite factor for direct hydrogenation of SA to THF since the reaction over the bare support was failed to produce any THF. The hydrogenation of GBL over Cu8Pd2/γ-Al2O3 indi-cated that in the preliminary step of this reaction, BDO was formed as an intermediate, which immediately converted to THF under the influence of the strong Lewis acid site. Therefore, the highly selective of THF via SA hydro-genation requires the combination and working in concert of the Cu-rich CuPd NPs and the strong Lewis acid sites of γ-Al2O3.

2.3

Conclusion

The present thesis provides feasible and versatile methods to design effective catalysts for selective hydrogenation of SA. The SA hydrogenation could be di-rected toward a specific product including GBL, BDO, and THF by adjusting the metal ratio, changing the catalyst support, and stabilizing with capping agent. The roles of metals, support, and capping agent have been revealed by extensive characterizations. In addition, the reaction pathways and optimized reaction conditions for each integrated reaction have been clarified and exam-ined. The mentioned understanding of the current catalyst system allows us to control the SA hydrogenation and other related reactions from various angles.

3

Research Accomplishment

Academic journals – peer-reviewed

1. Son Dinh Le, Shun Nishimura, Hydroxyapatite Supported Polyvinylpyrro-lidone-Capped CuPd Nanoparticles for Highly Selective Lactonization of Succinic Acid. (In preparation)

2. Son Dinh Le, Shun Nishimura, Influence of Metal Ratio on Alumina-Supported CuPd Catalysts for the Production of Tetrahydrofuran from Succinic Acid. (Submitted, under review)

3. Son Dinh Le, Shun Nishimura, Effect of support on the formation of CuPd alloy nanoparticles for the hydrogenation of succinic acid, Appl. Catal. B: Environ., 2021, 282, 119619.

4. Son Dinh Le, Shun Nishimura, Highly selective synthesis of 1,4-butanediol via hydrogenation of succinic acid with supported Cu–Pd alloy nanopar-ticles, ACS Sustain. Chem. Eng., 2019, 7, 18483–18492.

Patent applications

5. Shun Nishimura, Son Dinh Le, Catalysts for 1,4-butanediol, butyric acid and tetrahydrofuran productions, Japanese Patent Application Nos. 2020-115838 (filed July 3, 2020) and 2019-148506 (filed August 13, 2019).

1

Studies on Degradation Mechanism of Fluorescent Organic Light Emitting Diodes Based on Triplet-Triplet Annihilation by Time-resolved Luminescent Spectroscopies

(Materials Science)

Murata laboratory ID: 1720433

Name: LE Cong Duy

Organic light emitting-diodes (OLEDs) are an important technology for high-quality flat panel displays as well as lighting owing to their self-emission, high contrast ratio, wide viewing angle, and fast response. New attractive applications are enabled by OLEDs, such as flexible,1 _wearable,2 _{and transparent diplays.}3

Blue fluorescent OLEDs that utilize triplet-triplet annihilation (TTA) are employed for blue pixels of small OLED displays used for smart phones.4 _{TTA based blue OLEDs, however, demonstrates inferior operation}

lifetime compared with those of green and red phosphorescent OLEDs and requires further improvement.4

To overcome the short operation lifetime, understanding degradation mechanisms of TTA based OLEDs are essential. Emission of TTA based OLEDs is composed of prompt electroluminescence (EL) from singlet exciton that is directly generated by charge recombination and delayed EL from additional singlet exciton generated via TTA process. Hence, degradation mechanism of the OLEDs would be ascribed to the reduction of the density of singlet exciton during the device operation.

In this research, we aim at investigating the degradation mechanisms of TTA based fluorescent OLEDs with time-resolved luminescent spectroscopies. Degradation owing to the decrease in prompt EL was investigated by time-resolved photoluminescence (TRPL) spectroscopy. In addition, degradation ascribed to the decrease in delayed EL was investigated by time-resolved electroluminescence (TREL) spectroscopy. Degradation of green fluorescent OLEDs

Figure 1 demonstrates EL performance of pristine and degraded green OLEDs using tris(8-hydroxyquinolinolato) aluminum (III) (Alq3) as an emitter. Emission of Alq3-based

OLED uses both singlet exciton formed by charge recombination and TTA.5 _{External quantum efficiency (EQE)}

at 50 mA/cm2 _{of the degraded device shows 1.04%, which is}

90% of pristine one (1.15%).

The device degradation was investigated in terms of parameters of the EQE expressed as follows

𝐸𝑄𝐸 = 𝛾𝜂𝑜𝑢𝑡𝜙𝑃𝐿(𝜂𝑆+ 𝜂𝑇𝑇𝐴), (1) where  is charge balance factor, out is out-coupling

efficiency, PL is PL quantum yield, S and TTA are fractions

of singlet exciton formed by charge recombination and by TTA, respectively. Assuming  and out are constant as devices

age, the decrease in the EQE is attributed to the decrease in PL

and/or (S + TTA).

To investigate change in PL after the degradation, we have

measured TRPL of pristine and degraded devices. Singlet lifetime (τ) of Alq3 was obtained from fitting TRPL curves

(Fig. 2). Ratio of PL(0.9) over PL(1.0) corresponds to ratio of

τ(0.9) (13.5 ns) over τ(1.0) (14.3 ns) that is 94%. This indicates

that device degradation is partly ascribed to a reduction in PL

owing to singlet quenching with neutral quenchers (Fig. 4).

Figure 1. External quantum efficiency (EQE), EL spectrum (in set) of pristine (1.0) and degraded (0.9) devices.

Figure 2. Transient PL of pristine (1.0) and degraded (0.9) devices. Fitting equation:

2

Neutral quenchers may act as hole traps during device operation, which is confirmed by capacitance-voltage measurement