• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2009.06a
• /
• pp.137-137
• /
• 2009

Thin-film transistors (TFTs) that can be prepared at low temperatures have attracted much attention due to the great potential for flexible electronics. One of the mainstreams in this field is the use of organic semiconductors such as pentacene. But device performance of the organic TFTs is still limited by low field effect mobility or rapidly degraded after exposing to air in many cases. Another approach is amorphous oxide semiconductors. Amorphous oxide semiconductors (AOSs) have exactly attracted considerable attention because AOSs were fabricated at room temperature and used lots of application such as flexible display, electronic paper, large solar cells. Among the various AOSs, a-IGZO was considerable material because it has high mobility and uniform surface and good transparent. The high mobility is attributed to the result of the overlap of spherical s-orbital of the heavy pest-transition metal cations. This study is demonstrated the effect of thickness channel layer from 30nm to 200nm. when the thickness was increased, turn on voltage and subthreshold swing were decreased. a-IGZO TFTs have used a shadow mask to deposit channel and source/drain(S/D). a-IGZO were deposited on SiO2 wafer by rf magnetron sputtering. using power is 150W, working pressure is 3m Torr, and an O2/Ar(2/28 SCCM) atmosphere at room temperature. The electrodes were formed with Electron-beam evaporated Ti(30nm) and Au(70nm) structure. Finally, Al(150nm) as a gate metal was evaporated. TFT devices were heat treated in a furnace at $250^{\circ}C$ in nitrogen atmosphere for an hour. The electrical properties of the TFTs were measured using a probe-station to measure I-V characteristic. TFT whose thickness was 150nm exhibits a good subthreshold swing(S) of 0.72 V/decade and high on-off ratio of 1E+08. Field effect mobility, saturation effect mobility, and threshold voltage were evaluated 7.2, 5.8, 8V respectively.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.