The Quek Group

A charge density wave (CDW) state is a macroscopic quantum state exhibiting periodic lattice distortion and a modulation of the electron density modulation, leading to novel phenomena associated with these spontaneously broken symmetries. However, the CDW phase typically exists at very low temperatures. Efforts to increase the CDW phase transition temperature, known as T_CDW, have focused on the impact of interfacial strain and charge dopants. However, the effects of such modifications on T_CDW have not been significant, because the extent to which the CDW phase is stabilised by such modifications isintrinsically limited.

See Reference: ACS Nano 2020, 14, 4, 3917–3926 (2020)

Electronic screening is a process in which the electrons in a material rearrange themselves under the application of an external electric field. Being atomically thin, electronic screening within a 2D material is known to be significantly weaker than a bulk material. The screening of charges adjacent to a 2D material, however, is different. Using first-principles calculations, we found that just a single layer of a 2D material is able to significantly impact the energy levels of an organic molecule placed above it. In fact, a 2D material can screen just as effectively as a 3D one with the same energy gap. This result shows that electronic screening by 2D materials is non-trivial and needs to be carefully investigated in consideration of designing next-generation electronic devices.

Hexagonal boron nitride, a 2D material, has been demonstrated to be an excellent substrate for graphene-based electronics by significantly reducing charge inhomogeneities above it. With our first-principles calculations, we provide evidence that, in general, 2D materials can serve as excellent substrates due to their intrinsic ability to screen out the effect of charge impurities beneath them. (This insight would have been difficult to probe within a laboratory where environmental effects would have to be taken into account.)

See Reference: 2D Materials 6, 035036 (2019)

Two-dimensional (2D) material systems have recently been found to host localized excitons and single photon emitters (SPEs), meaning that light is emitted from colour centers one photon at a time. This has huge implications in quantum optics and quantum information processing, where device integration and up-scaling are currently the biggest challenges to surmount. SPEs in 2D materials such as WSe₂, are intrinsically embedded in a two-dimensional material matrix, allowing for potential device and circuit integration as in conventional semiconductor industry. However, the detailed nature of these experimentally discovered SPEs in WSe₂ are still unknown, which hinders their further utilization and application. We have now unveiled the nature of point defects and localized excitons in WSe₂ single layers through a combination of theory and experiment.    

To exactly pin-point the SPE-related point defects and correlate their nature with the corresponding emission line is a non-trivial task. For the well-established single photon quantum emitters in diamond systems, researchers took decades to finally ascertain the exact energy level structure of the exact defect centers responsible for the SPE in diamond. Similarly, the chemical origins and detailed energy levels of defect centers responsible for SPEs in 2D materials have been controversial so far. The difficulty arises from several different perspectives. The spatially resolution of experimental characterization techniques has not reached single atomic limit, when the photo-activities of those sites need to be determined simultaneously. On the other hand, theoretical descriptions of defect centers are usually quantitatively unsatisfactory or inconclusive.

Here, our group has studied theoretically the point defects and localized excitons in a 2D WSe₂, in collaboration with our experimental collaborators, Professor Andrew Wee and Professor Stephen Pennycook in NUS. We first conducted exhaustive density functional theory (DFT) calculations on intrinsic point defects in WSe₂. The experimental scanning tunneling microscopy and spectroscopy (STM/S) methods are used to characterize the most abundant point defects in the same system. It was found that none of the intrinsic point defects in the system matches the experimental STS spectra. Thus, the researchers then focused on the oxygen-related point defects of WSe₂, which they found could be incorporated easily into the material inadvertently, either during the synthesis process or by ambient passivation. Through comparison of experimental and simulated STM images, the researchers identified oxygen substituted Se vacancies (O_Se) and oxygen interstitials in the lattice (O_ins) as the most abundant point defects in WSe₂. Additionally, scanning transmission electron microscopy (STEM) images identified an additional Se_W anti-site defect. Subsequently, state-of-the-art many body perturbation theory based GW-Bethe Salpeter equation (GW-BSE) calculations are conducted for the three most likely extrinsic defects identified in WSe₂:  O_Se, O_ins and Se_W. It was found that only the O_ins defect hosts localized exciton states close to the experimentally observed spectral positions. O_Se does not possess localized excitons, while Se_W has localized excitons too low in energy. The mechanical strain gradient within the material will also help to tune the spectral position and funnel excitons into the bottom of the strain well.

This work presents a detailed study of point defects in monolayer WSe₂ and predicted the nature and energies of excitons at these defect sites. The implications of the experimentally observed point defects on the optical response have been predicted. The key point here is that of all the experimentally observed point defects, only one candidate (O_ins) is available that gives localized excitons in the energy range observed in recent SPE experiments. These predictions suggest ways to create quantum emitters in other semiconducting TMDs, e.g., through controlled reaction with O₂, as well as their further integration into quantum optical applications.

 See Reference: ACS Nano 13, 6050 (2019)

Many-body perturbation theory in the GW approximation can give accurate quasiparticle levels for the electrons and holes, and therefore accurate energy barriers. However, the conventional GW code is limited to rather small systems due to high computational cost and memory. One key bottleneck is the computation of the polarizability (chi) matrix. Several groups have simplified the computation of chi for interfaces with no hybridization, by taking the chi matrix for the full system to be the sum of the chi matrices of the individual components. However, it is often assumed that this method breaks down in the presence of hybridization between the individual components.

In our group, we have developed a method, which we call XAF-GW (X: eXpand-chi, A: Add-chi, F: Full wavefunctions), that enables GW calculations for large interface systems. We show analytically that even in the presence of interface hybridization to form bonding and anti-bonding states, the chi matrix of the full interface is still a sum of the individual components, up to first order in the overlap matrix elements involved in the hybridization. The approach is validated by showing that the band structure obtained using this method is almost identical to that obtained using a regular GW calculation for bilayer black phosphorus, where interlayer hybridization is significant. Significant savings in computational time and memory are obtained by computing chi only for the smallest sub-unit cell of each component, and expanding (unfolding) the chi matrix to that in the unit cell of the interface. To treat interface hybridization, the full wavefunctions of the interface are used in computing the self-energy.

Using this approach, we have successfully obtained accurate quasiparticle levels and energy barriers for an experimentally relevant organic/2D material system, as shown in the figure below.

The XAF-GW approach is easy to implement and will enable accurate prediction of quasiparticle levels and energy barriers at large interface systems, including other organic/2D material interfaces, and twisted bilayer systems.

See Reference: Journal of Chemical Theory and Computation 15, 3824 (2019)

Molecular electronics involves the use of molecules as the main building block for creating the electronic circuitry. It can potentially be used to develop circuits that are much smaller than those made from conventional silicon processes. Understanding the electronic properties of the interface between the molecules and metal conductors, particularly their associated energy levels, is important for rationalising and optimising device performance. This is central to the development of molecular electronics.

A fundamental property of every molecule is its energy gap, defined as the energy difference between the highest and lowest orbital energy level occupied and unoccupied by electrons respectively. These levels are also the most important energy levels for device performance. The energy gap of a molecule becomes smaller when the molecule is brought close to a metal surface; this will make it easier for charge carriers to move between the molecule and the metal contact. This change in gap is primarily caused by electronic screening effects from the metal surface, and can be as large as several electron-volts. However, this electronic screening effect is missing from the majority of theoretical studies on this topic.

A research team led by Prof Su Ying QUEK, from the Department of Physics, NUS has elucidated the interface electronic structure properties for a number of different molecules on gold surfaces using state-of-the-art theoretical and computational methods that explicitly take into account electronic screening effects from first principles. The researchers carried out computations on molecular systems anchored by common chemical functional groups (amine, pyridine and thiolate groups). The research team found that for a single molecule, the electronic screening effect can be accurately predicted from an image charge model, even in the presence of chemical bonds. The image charge model is a classical electrostatics method which approximates the electronic screening of a test charge by an image charge in the metal. However, in devices with many molecules, the researchers found significant additional electronic screening mechanisms. Besides intermolecular screening effects, substrate-mediated intermolecular interactions are also found to contribute to these additional screening mechanisms. The findings suggest that charge carriers can tunnel more easily across the interface in devices with many molecules.

Prof Quek said, “This work provides valuable insights into the many electron effects at the molecule-metal interfaces involving chemical bonds. The results and findings from this research constitute an important step towards the understanding and manipulation of functional organic systems in the development of molecular devices.”

See Reference:
The Journal of Physical Chemistry C 121, 13125 (2017)

Electronic devices with lower resistance between connections uses less power. These devices are also less likely to have localised heating effects. Graphene, a flexible material with exceptionally high electron mobility and thermal conductivity, is a material for next generation electronics. However, it has a large resistance when used in electrical connections. The construction of smaller and more efficient electronic devices is hindered partly by heat generated within the device. If this resistance can be lowered, electronic devices will be able to operate at a higher speed using less power. An NUS team has now shown how the contact resistance can be reduced in graphene devices by manipulating the spin state of the charge carriers.

Although graphene is a prime candidate material for next generation electronics the large electrical resistance at metal-graphene interfaces is a bottleneck for practical devices. A collaboration between Prof Su Ying QUEK from the Department of Physics, NUS and Prof John THONG from the Department of Electrical and Computer Engineering, NUS has shown that “edge-contacted” device geometries in Ni/Co-graphene interfaces result in some of the lowest contact resistances reported to date. The resistance in such geometries is significantly lower than in “surface-contacted” Ni/Co-graphene interfaces. This is due to the different behaviour of electron spins in these geometries. Temperature-dependent measurements show that electron spin indeed can control the contact resistance at ferromagnetic-graphene interfaces.

There have been significant efforts in graphene research due to its unique material properties including high electron mobility, excellent mechanical strength and flexibility, and high thermal conductivity. This has prompted the use of graphene as a potential material for on-chip interconnects and high-speed transistors for next generation electronic circuits. However, the large electrical resistance at the metal-graphene interface limits the performance of these devices.

In this work, the researchers examined the contact resistance of edge- versus conventional surface-contacted ferromagnetic metal-graphene interfaces. They found that the contact resistance of edge-contacted interfaces is much lower, with resistances comparable to the lowest values reported to date. They showed that the higher contact resistance arises from the spin filtering phenomena at the metal-graphene interface. This is in contrast to previously reported work, where it is widely believed that the higher resistance is a result of weaker coupling strength. By using edge contacts, one essentially allows charge carriers of both majority and minority spin to pass from the electrodes to graphene, resulting in almost twice the number of carriers and reducing the contact resistance significantly. These findings suggest that the contact resistance in graphene devices could be adjusted by tuning the spin state of the magnetic metal.

The research team plans to explore tuning the contact resistance in ferromagnetic metal-graphene contacts by varying the spin state of the metal. They also intend to investigate other device parameters such as spin injection efficiency and spin lifetime at these interfaces for spintronics applications.

See reference:
ACS Nano 10, 11219 (2016)

Two-dimensional (2D) layered materials have taken the world of materials science by storm. Unlike other materials, one would expect that peeling off layers to create a surface in a layered material should not cause much grief to the surface atoms. A team led by Prof QUEK Su Ying from Department of Physics in NUS has found, however, that even in layered materials, surface atoms feel the loss of the missing neighboring layer enough to cling on more strongly to their remaining neighbors in the same layer. And these stronger surface interactions account for previously unexplained anomalous trends in the vibrational frequencies of these prototypical 2D materials.

Vibrational frequencies are routinely used to identify the thickness of 2D materials. For the vibration which involves adjacent layers moving against one another (see Figure). It was thought that the frequency should increase with thickness due to increased friction in thicker materials. By parameterizing high-level calculations into a simple physically-motivated model, the team found this to be indeed the case if the surface force constants are the same as in the bulk material. However, once the stronger surface force constants are accounted for, the frequency decreases with increasing thickness, in excellent agreement with experiments. Similar effects have since been found in other 2D layered materials.

These findings change the mindset that surface effects are not important in layered materials, and have implications for predicting vibrational frequencies and thermal properties in these systems.

See References: Physical Review B 88, 075320 (2013), Nano Letters 13, 1007 (2013), Critical Reviews in Solid State and Materials Sciences 39, 319 (2014)