This paper presents an experimental investigation on surface and subsurface characterisation of micro-machined single-crystal silicon with (100) orientation. Full immersion slot milling was conducted using solid cubic boron nitride (CBN) and diamond-coated fine grain tungsten carbide micro-end mills with a uniform tool diameter of 0.5 mm. The micro-machining experiments were carried out on an ultra-precision micro-machining centre. Formal design of experiments (DoE) techniques were applied to design and analysis of the machining process. Surface roughness, edge chipping formation and subsurface residual stress under varying machining conditions were characterised using white light interferometry, SEM and Raman microspectroscopy. Tens of nanometre-level surface roughness can be achieved under the certain machining conditions, and influences of variation of cutting parameters including cutting speeds, feedrate and axial depth of cut on surface roughness were analysed using analysis of variance (ANOVA) method. Raman microspectroscopy studies show that compressive subsurface residual stress and amorphous phase transformation were observed on most of the micro-machined subsurface, which provides evidence of ductile mode cutting. Surface and subsurface characterisation studies show that the primary material removal mode is ductile or partial ductile using lower feedrate for both tools, and diamond-coated tools can produce better surface quality. Silicon brain implants were fabricated with good dimensional accuracy and edge quality using the optimised machining conditions, which demonstrated that micro-milling is an effective process for fabrication of silicon components at a few tens to a few hundreds of micron scale.

Lithium niobate (LiNbO 3 ) is a crystalline material which is widely applied in surface acoustic wave, microelectromechanical systems (MEMS), and optical devices, owing to its superior physical, optical, and electronic properties. Due to its low toughness and chemical inactivity, LiNbO 3 is considered to be a hard-to-machine material and has been traditionally left as as an inert substrate upon which other micro structures are deposited. However, in order to make use of its superior material properties and increase efficiency, the fabrication of microstructures directly on LiNbO 3 is in high demand. This paper presents an experimental investigation on the micro machinability of LiNbO 3 via micro milling with the aim of obtaining optimal process parameters. Machining of micro slots was performed on Z-cut LiNbO 3 wafers using single crystal diamond tools. Surface and edge quality, cutting forces, and the crystallographic effect were examined and characterized. Ductile mode machining of LiNbO 3 was found to be feasible at a low feed rate and small depth of cut. A strong crystallographic effect on the machined surface quality was also observed. Finally, some LiNbO 3 micro components applicable to sensing applications were fabricated.

No abstract

This article presents the research on the effect of crystallographic orientation and different cutting tool effect during micro-milling of (001) silicon wafer. Excessive generation of undesirable surface and subsurface damages often occurs when machined at thick depth of cut of several hundreds of microns. Up-milling operations along <100> and <110> directions were performed on a (001) wafer, and the results show that machining surfaces along <100> were of better quality than those of <110> and are in agreement with previous studies. In addition, comparative studies of diamond-coated, chemical vapour–deposited and single crystal diamond end-mills were performed along [Formula: see text] at 150 µm deep. Results have shown that diamond-coated tool generates the least edge chipping. This might be due to the large negative rake angle creating highly compressive hydrostatic pressure in the cutting zone and therefore suppressing the crack propagation. Furthermore, no visible defects were detected on the bottom-machined surface when machined by chemical vapour–deposited and single crystal diamond end-mills. Surface edge chipping however remains a challenge, even though micro-milling were performed along <100> with single crystal diamond end-mill. Apart from milling along <100>, protection to the top silicon surface is required to achieve fracture-free quality micro-milled silicon.

As one of the core applications of computer vision, object detection has become more important in scenarios requiring high accuracy but with limited computational resources such as robotics and autonomous vehicles. Object detection using machine learning running on embedded device such as Raspberry Pi provides the high possibility to detect any custom objects without the recalibration of camera. In this work, we developed a smart and lean object detection model for shipping containers by using the state-of-the-art deep learning TensorFlow model and deployed it to a Raspberry Pi. Using EfficientDet-Lite2, we explored the different cross-validation strategies (Hold-out and K-Fold). The experimental results show that compared with the baseline EfficientDet-Lite2 algorithm, our model improved the mean average precision (mAP) by 44.73% for the Hold-out dataset and 6.26% for K-Fold cross-validation. We achieved Average Precision (AP) of more than 80% and best detection scores of more than 93% for the Hold-out dataset. For the 5-Fold lean dataset, the results show the Average Precision across the three lightweight models are generally high as the models achieved more than 50% average precision, with YOLOv4 Tiny performing better than EfficientDet-Lite2 and Single Shot Detector (SSD) MobileNet V2 Feature Pyramid Network (FPN) 320 as a lightweight model.