We present a rapid and highly reliable glass (fused silica) microfluidic device fabrication process using various laser processes, including maskless microchannel formation and packaging. Femtosecond laser assisted selective etching was adopted to pattern microfluidic channels on a glass substrate and direct welding was applied for local melting of the glass interface in the vicinity of the microchannels. To pattern channels, a pulse energy of 10 μJ was used with a scanning speed of 100 mm/s at a pulse repetition rate of 500 kHz. After 20–30 min of etching in hydrofluoric acid (HF), the glass was welded with a pulse energy of 2.7 μJ and a speed of 20 mm/s. The developed process was as simple as drawing, but powerful enough to reduce the entire production time to an hour. To investigate the welding strength of the fabricated glass device, we increased the hydraulic pressure inside the microchannel of the glass device integrated into a custom-built pressure measurement system and monitored the internal pressure. The glass device showed extremely reliable bonding by enduring internal pressure up to at least 1.4 MPa without any leakage or breakage. The measured pressure is 3.5-fold higher than the maximum internal pressure of the conventional polydimethylsiloxane (PDMS)–glass or PDMS–PDMS bonding. The demonstrated laser process can be applied to produce a new class of glass devices with reliability in a high pressure environment, which cannot be achieved by PDMS devices or ultraviolet (UV) glued glass devices.

We present the selective laser-induced etching (SLE) process and design guidelines for the fabrication of three-dimensional (3D) microfluidic channels in a glass. The SLE process consisting of laser direct patterning and wet chemical etching uses different etch rates between the laser modified area and the unmodified area. The etch selectivity is an important factor for the processing speed and the fabrication resolution of the 3D structures. In order to obtain the maximum etching selectivity, we investigated the process window of the SLE process: the laser pulse energy, pulse repetition rate, and scan speed. When using potassium hydroxide (KOH) as a wet etchant, the maximum etch rate of the laser-modified glass was obtained to be 166 μm/h, exhibiting the highest selectivity about 333 respect to the pristine glass. Based on the optimized process window, a 3D microfluidic channel branching to three multilayered channels was successfully fabricated in a 4 mm-thick glass. In addition, appropriate design guidelines for preventing cracks in a glass and calibrating the position of the dimension of the hollow channels were studied.

We introduce a new 3D impeller micromixer built within a single glass substrate using ultrafast laser process and it shows high mixing efficiency up to 99% and throughput of 30 mL min ⁻¹ with a short mixing channel length of 0.98 mm.

We propose a new packaging process for an implantable blood pressure sensor using ultrafast laser micro-welding. The sensor is a membrane type, passive device that uses the change in the capacitance caused by the membrane deformation due to applied pressure. Components of the sensor such as inductors and capacitors were fabricated on two glass (quartz) wafers and the two wafers were bonded into a single package. Conventional bonding methods such as adhesive bonding, thermal bonding, and anodic bonding require considerable effort and cost. Therefore CO₂ laser cutting was used due to its fast and easy operation providing melting and bonding of the interface at the same time. However, a severe heat process leading to a large temperature gradient by rapid heating and quenching at the interface causes microcracks in brittle glass and results in low durability and production yield. In this paper, we introduce an ultrafast laser process for glass bonding because it can optimize the heat accumulation inside the glass by a short pulse width within a few picoseconds and a high pulse repetition rate. As a result, the ultrafast laser welding provides microscale bonding for glass pressure sensor packaging. The packaging process was performed with a minimized welding seam width of 100 μm with a minute. The minimized welding seam allows a drastic reduction of the sensor size, which is a significant benefit for implantable sensors. The fabricated pressure sensor was operated with resonance frequencies corresponding to applied pressures and there was no air leakage through the welded interface. In addition, in vitro cytotoxicity tests with the sensor showed that there was no elution of inner components and the ultrafast laser packaged sensor is non-toxic. The ultrafast laser welding provides a fast and robust glass chip packaging, which has advantages in hermeticity, bio-compatibility, and cost-effectiveness in the manufacturing of compact implantable sensors.

Low TNF production and its association with TNF gene restriction fragment length polymorphism (RFLP) was demonstrated in (NZW/NZB) F₁ mice. However, little is known about the significance of TNF production in association with TNF gene polymorphism in human SLE. This study was designed to evaluate the role of TNF production of peripheral blood mononuclear cells (PBMC) and its association with TNFB gene polymorphism in SLE, particularly lupus nephritis. TNFB gene polymorphism was defined by PCR- NcoI RFLP. TNF productions of phytohemagglutinin (PHA)- stimulated PBMC and T cells were examined by bioassay using L929 cell line and ELISA. The PBMC stimulated by PHA from patients with SLE ( n = 60) tended to secrete less amounts of TNF by bioassay (1032±184pg/ml vs 1524±224pg/ml, P=0.094), and TNF-β by ELISA ( P = 0.0082) than that from normal controls ( n = 38). The low TNF-α producer was more frequent in nephritis than non-nephritis (34.4% vs 7.1% respectively, P<0.01). TNF-β also revealed similar results (53.1% vs 21.4%, P<0.05). In SLE, mean production of TNF-β was decreased in TNFB*2 homozygote ( n=18) than that in TNFB*1 homozygote ( n = 9) (1126.3±145 pg/ml) vs 642 ± 118.4 pg/ml, respectively, P = 0.021), whereas TNF-α production showed little difference between the two groups (710.1 ± 56.4 vs 542.4 ± 71.1 pg/ml, respectively, P = 0.149). Our results demonstrate that decreased TNF production of PBMC, which was significantly associated with TNFB*2 homozygosity, could be an important predisposing factor of lupus nephritis in Koreans.