Interfacial electron transfer (ET) between semiconductor nanomaterials and molecular adsorbates is an important fundamental process that is relevant to applications of these materials. Using femtosecond midinfrared spectroscopy, we have simultaneously measured the dynamics of injected electrons and adsorbates by directly monitoring the mid-IR absorption of electrons in the semiconductor and the change in adsorbate vibrational spectrum, respectively. We report on a series of studies designed to understand how the interfacial ET dynamics depends on the properties of the adsorbates, semiconductors, and their interaction. In Ru(dcbpy)2(SCN)2 (dcbpy = 2,2‘-bipyridine-4,4‘-dicarboxylate) sensitized TiO2 thin films, 400 nm excitation of the molecule promotes an electron to the metal-to-ligand charge transfer (MLCT) excited state, from which it is injected into TiO2. The injection process was characterized by a fast component, with a time constant of <100 fs, and a slower component that is sensitive to sample condition. Similar ultrafast electron injection times were measured in TiO2 films sensitized by Ru(dcbpy)2(X)2 (X2 = 2CN- and dcbpy). Electron injection in these systems was found to compete with the vibrational energy relaxation process within the excited state of the molecules, leading to an injection yield that depends on the excited-state redox potential of the adsorbate. The injection rate from Ru(dcbpy)2(SCN)2 to different semiconductors was found to obey the trend TiO2 > SnO2 > ZnO, indicating a strong dependence on the nature of the semiconductor. To understand these observations, various factors, such as electronic coupling, density of states, and driving force, that control the interfacial ET rate were examined separately. The effect of electronic coupling on the ET rate was studied in TiO2 sensitized by three adsorbates, Re(L n )(CO)3Cl [L n is a modified dcbpy ligand with n (=0, 1, 3) CH2 units between the bipyridine and carboxylate groups]. We found that the ET rate decreased with increasing number of CH2 units (or decreasing electronic coupling strength). The effect of driving force was investigated in Ru(dcbpy)2X2 (X2 = 2SCN-, 2CN-, and dcbpy) sensitized SnO2 thin films. In this case, we observed that the ET rate increased with the excited-state redox potential of the adsorbates, agreeing qualitatively with the theoretical prediction for a nonadiabatic interfacial ET process.

No abstract

To explore the potential for use of ligand-conjugated nanocrystals to target cell surface receptors, ion channels, and transporters, we explored the ability of serotonin-labeled CdSe nanocrystals (SNACs) to interact with antidepressant-sensitive, human and Drosophila serotonin transporters (hSERT, dSERT) expressed in HeLa and HEK-293 cells. Unlike unconjugated nanocrystals, SNACs were found to dose-dependently inhibit transport of radiolabeled serotonin by hSERT and dSERT, with an estimated half-maximal activity (EC(50)) of 33 (dSERT) and 99 microM (hSERT). When serotonin was conjugated to the nanocrystal through a linker arm (LSNACs), the EC(50) for hSERT was determined to be 115 microM. Electrophysiology measurements indicated that LSNACs did not elicit currents from the serotonin-3 (5HT(3)) receptor but did produce currents when exposed to the transporter, which are similar to those elicited by antagonists. Moreover, fluorescent LSNACs were found to label SERT-transfected cells but did not label either nontransfected cells or transfected cells coincubated with the high-affinity SERT antagonist paroxetine. These findings support further consideration of ligand-conjugated nanocrystals as versatile probes of membrane proteins in living cells.

No abstract

Spoken language understanding system is traditionally designed as a pipeline of a number of components. First, the audio signal is processed by an automatic speech recognizer for transcription or n-best hypotheses. With the recognition results, a natural language understanding system classifies the text to structured data as domain, intent and slots for downstreaming consumers, such as dialog system, hands-free applications. These components are usually developed and optimized independently. In this paper, we present our study on an end-to-end learning system for spoken language understanding. With this unified approach, we can infer the semantic meaning directly from audio features without the intermediate text representation. This study showed that the trained model can achieve reasonable good result and demonstrated that the model can capture the semantic attention directly from the audio features.Index Terms-Spoken language understanding, end-toend training, recurrent neural networks