【Research Highlights】

 
Physical Sciences
The laboratory of solid state microstructures of NJU was considered “approaching world class standards in research” by Nature in 1997. There have been numerous achievements in the research history of NJU. However, perhaps the most influential one should associate with Professor Nai-ben Ming. He and his team members, Yong-yuan Zhu, Shi-ning Zhu, Ya-lin Lu and Yan-qing Lu won the First Prize of National Natural Science Award in 2006. It took almost 20 years for Ming’s team to achieve this precious high honor since 1986. 
 
In many news reports around 2007, a photo of a dielectric superlattice laser with three colors was frequently exposed. It was viewed as one of the most representable work of Ming’s award. Actually what his team revealed are much wider and profound, from the fundamental theory of domain engineering and quasi-phase-matching to the design, fabrication, properties, and applications of dielectric superlattice (DSL) materials. As a vivid example, their capability showing in the three-color laser is astonishing, which means they may intentionally control the output spectrum with only a single input wavelength through cascading nonlinear optical interactions. It was always people’s dream to manipulate light’s route, intensity, polarization etc, while it is the DSL’s unique properties that give rise to the simultaneous multi-color generation. The physics behind this phenomenon was proposed by Yong-yuan Zhu then experimentally demonstrated by Shi-ning Zhu both under Ming’s supervision. This work later was extended to an active branch of on-demanded single and multiple wavelength light sources based on domain engineering. Other interesting features of the DSL family including high frequency ultrasonic bulk or surface wave generation, electro-optic polarization rotation, all-optical bistability and even negative dielectric constant due to polariton excitation. The last one, proposed by Yan-qing Lu et al., actually is very close to the embryonic concept of current metamaterial.
 
In recent year, Shi-ning Zhu’s group successfully developed and commercialized a serial of DSL based on solid state lasers, especially the mid-infrared tunable lasers that have massive important applications in noninvasive medical diagnostics, free space communication, laser scalpel, spectroscopy and remote sensing. In the meantime, Shi-ning Zhu et al. also have some cutting-edge work on plasmonics and integrated photonic circuits. Hui Liu et al. even demonstrated Gravitation lensing mimicking and optical trapping in an optical chip, which was highlighted in last year’s APS March meeting images gallery. However, all these work is still in the classical physics regime, while the most giant leap in the DSL study is entering the miracle quantum world.
 
Stimulated by the striking properties and applications of DSLs in nonlinear optics, Shi-ning Zhu and Ping Xu mainly focused on the generation and manipulation of photonic entanglement which acts as the key resource in quantum technologies. Due to the unique properties of DSLs, they may not only increases the photon flux and achievable wavelength but also offers a compact way to engineer the polarization, frequency, spatial entanglement through domain engineering on a monolithic crystal wafer. They fabricated the first Lithium Niobate (LN) quantum chip featuring two efficient photon sources in addition to the manipulation of photons through components such as junctions and wavelength-selective filters. This chip is controlled with on-chip electro-optical modulators resulting in a variety of quantum states. After this work was published in Phys. Rev. Lett., many media reported this new progress including APS Physics and IEEE Spectrum. According to Ping Xu, “This work present a scenario that LN superlattice may become an ideal platform towards fully integrated quantum optics.” 
 
 
 
Chemistry
Nanomaterials and nanostructures have exerted great influence on numerous fields. As one of the earliest groups in China, Professor Hong-Yuan Chen & Professor Jing-Juan Xu’ group from the State Key Laboratory of Analytical Chemistry for Life Sciences has introduced nanotechnology and biotechnology into electrochemistry field, and conducted substantial experimental and theoretical studies in ultramicroelectrodes, electrochemical biosensing and bioelectrochemistry as well as microfluidic chips etc. Interested in broad range of functionalized nanomaterials and patterned nanostructures, they have exploited the unique electronic, mechanical, electrochemical, photoelectrochemical and interfacial properties of these materials for addressing critical issues both in fundamental studies and real-world applications more than 20 years. They have achieved multiple national-class scientific awards for their accomplishments. 
 
Their recent research interest focuses on exploiting the photophysical properties of the nanomaterials for photo-electric mutual conversion-based bioanalytical processes, i.e., electrochemiluminescent (ECL) and photoelectrochemical (PEC) bioanalysis toward various life-related substances. Main contributions of their recent works could be summarized as follows:
 
1. On the Construction of High Efficient Energy Transfer-based ECL Sensing Interface. They were the first to introduce the energy transfer into the ECL of semiconductor nanocrystals system. Utilizing the tenability of sizes and shapes of nanomaterials, they adjusted the ECL emission wavelengths to find suitable energy donor-acceptor pairs, which not only solved the problem from the instability of organic compound acceptors in ECL process but also endowed the energy transfer mode with great diversity. They further presented four types of ECL-RET patterns: (a) They were the first to propose the semiconductor nanocrystals ECL DNA sensor, based on the energy transfer of Au nanoparticles (NPs) caused by local surface plasma resonance (LSPR). Combined with a “DNA machine” for amplification, this approach could sensitively respond DNA down to 5 aM. (b) Based on absorption characteristic of the black body-like NPs, they proposed a high efficient energy quenching mode of ECL sensing. (c) They demonstrated that the distance-depending property of ECL from the diluted magnetic semiconductor CdS:Mn nanocrystals by superparamagnetic Fe3O4 NPs, and applied such opto-magnetic interaction for sensitive immunoassay. (d) They developed the ECL-RET system from CdS QDs to Ru (bpy)32+ and realized the sensitive Cytosensing. Besides, they have also done extensive work on the preparation of high-performance ECL materials, dual-potential ECL ratiometric sensing, wireless bipolar electrode based ECL biosensing system, portable thermos-powered visual ECL detection device and etc. Integrated specific recognition events with these systems, small molecules (ions), proteins and nuclei acids could be detected with high sensitivity and specificity. 
 
2. On Novel PEC Sensing Methodologies Based on the Energy Transfer Processes and Multi-Amplification Signaling Strategies. They first investigated the energy transfer process between CdS QDs and noble metal NPs in the PEC sensing system and realized the sensitive DNA detection with Au NPs and Ag NPs as labels. Using the charge-transfer photochemistry at the surface of noble metals NPs caused by LSPR stimulation, they established the visible light-activated plasmonic PEC biosensing approach. Integrated the functionalized semiconductor electrodes with various biocatalytic systems, a serial of new protocols were developed for enzymatic sensing and PEC immunoassay toward the protein biomarkers. 
 
 
 
Earth Sciences                             
Atmospheric dust plays an important role in the marine and terrestrial geochemical cycles, and impacts global climate at orbital to annual timescales. Asia is the second largest dust source on Earth, and the deserts, sand fields and Gobi in China dominate Asian dust emission. Studies of dust transport based on dust tracers and satellite imagery clearly show that the eolian dust from Asia is transported globally. Some is deposited in the nearby Chinese Loess Plateau, whereas the remaining dust could be carried as far as the central North Pacific Ocean and even the Northern America. In the leeward regions of dust source areas, dust deposition provides an excellent archive for reconstructing the past environmental changes since at least ~22 Ma ago.
 
Over the past twenty years, Professor Jun Chen and his team have investigated the origin, transportation and deposition of Asian dust; they used the loess and Red Clay sediments in North China and the aeolian deep-sea sediments in the North Pacific Ocean to reconstruct past environment changes during the late Cenozoic era. Their scientific contributions focus on three major areas in recent years. First, they established the spatial patterns of Nd-Sr isotope ratios, rare Earth element (REE) compositions, and detrital mineral compositions for the eolian sediments in the deserts, sand fields, Gobi deserts, loess deposition areas, and the Tibetan Plateau in China. Based on the results, they were able to delineate the source regions of Asian dust, and provide direct evidence for the close relationship between the isotopic composition of desert deposits and the tectonic settings of the surrounding mountains. Second, by analyzing Nd-Sr isotopes and geochemistry of detrital monomineral, they found that the direction of short-range transportation of dust in central and eastern Asia was largely determined by the prevailing near surface wind. Source tracing also revealed that the Asian dust had two ultimate material sources: the northern margin of the Tibetan Plateau and the Central Asian Orogen. These findings confirmed the importance of mountain processes in the production of silt eolian particles. On the basis of these observations, and the Nd-Sr isotope analyses on the densely sampled long aeolian silt deposit sequences, they found major changes in the nature and provenance of Asian dust accumulation at ~7 Ma and ~1.2 Ma since the Early Miocene (~22 Ma). In addition, by calculating the Nd isotopic ratio of Asian dust contribution to the deep-sea sediments of the North Pacific Ocean, they found progressive uplifting of the Northern Tibetan Plateau since the mid Miocene (~15 Ma). Third, they found that all of the sand fields in Northern China, an important dust source today, were nearly completely covered by vegetation during the Holocene Optimum, whereas the deserts in northwestern and central northern China were reduced by around 5-20% in area during this time. These findings suggest that the surface conditions of Asian dust source regions changed dramatically during the glacial and interglacial climate variations, and such changes affected regional dust emissions and subsequent feedback to the climate change. Therefore, the dust impact should be coupled with the climate system in paleoclimatic modelling in order to fully understand the forcing mechanism of the past climate changes. 
 
The unique processes of dust deposition and transportation in central and eastern Asia provide a key to understanding the evolution of the Earth’s environmental systems during late Cenozoic. The ongoing ambitious research project undertaken by Professor Chen and his team focuses on investigating the global dust emission, transportation, deposition and their climatic impacts at the orbital to annual timescales. They are using high-resolution and high-precision geochemistry and mineralogy techniques to pinpoint dust sources, and look for new biogeochemistry proxy indices to better reconstruct past climate and environment changes. Their research will significantly improve our understanding of the behavior of the Earth’s surface environmental systems.
 
 
Information Sciences
The LAMDA is a research team led by Professor Zhi-Hua Zhou, who is an ACM Distinguished Scientist and IEEE Fellow with h-index 63, belonging to the National Key Laboratory for Novel Software Technology of Nanjing University. As its name “Learning And Mining from DatA” suggests, this team focuses on research of machine learning and data mining, with the fundamental motivation of constructing computer programs that are able to analyze existing data and generalize to unseen future cases. This kind of research becomes extremely important in the big data era. 
 
Aiming at constructing models with strong generalization, one popular approach is to construct ensembles of multiple models. Professor Zhou and his colleagues proved that it possible to prune a full-sized ensemble to a small-sized one with even better performance; breaks the existing concept that ensemble pruning will sacrifice generalization, and initiated the study of optimization-based pruning. To enable complicated black-box models to be understandable, they proposed effective approaches that can produce comprehensible single models that are even more accurate than the ensembles; this line of research is similar to “knowledge distillation” recently advocated by some deep learning researchers, but ten years earlier. Another approach to strong generalization is to exploit unlabeled data by semi-supervised learning, rather than considering labeled data only. Professor Zhou and his students first proved a theory of sufficient and necessary condition for a mainstream paradigm, and started the study of semi-supervised regression which later became an important branch. He also pioneered the leverage of ensemble and semi-supervised learning, and disclosed that these two different methodologies can be mutually beneficial. Considering that data objects in real tasks often belong to multiple concepts simultaneously, they proposed the MIML (multi-instance multi-label) learning framework and developed a series of learning algorithms. Other important contributions can be perceived in their website (lamda.nju.edu.cn). 
 
In addition to publishing high quality works in top-tier publication venues in their areas, Professor Zhou and his collaborators applied their innovative techniques to various practical tasks including image retrieval, movie annotation, network monitor, CPU design, cancer diagnosis, gene pattern annotation, mobile photo management, etc. Their innovations have also been incorporated into MATLAB and many open-source softwares. Part of their research won the National Natural Science Award of China, 2nd grade, the highest distinctions achieved by China computer scientists in 2013. In that year Professor Zhou also received the Annual IEEE CIS Outstanding Early Career Award, which was presented to only one outstanding young researcher no more than 40 years old over the world. 
 
When Professor Zhou founded LAMDA ten years ago, there were only two faculties; now the team has eleven faculties and research staffs, possibly the strongest in its area in China and known internationally. LAMDA shows warm attitude of attracting oversea Chinese talents, and also exhibits strong ability of talent cultivation, e.g., during the last six years, every year there was a LAMDA student won the CCF Best Ph.d Thesis Award, the premium award established by the China Computer Federation to recognize the ten best Ph.d theses in computer science all over China. Indeed, Professor Zhou himself is a good example of home-grown talent of Nanjing University-- he got all his degrees in this university. The National Key Laboratory for Novel Software Technology, with which LAMDA is affiliated, has been ranked as the No.1 among all the national key laboratories in computer science in China, in the past two rounds of national assessments held in 2007 and 2012, respectively.