Hiroshi Hiroshima

Control of bubble defects in UV nanoimprint

We studied UV nanoimprint in air and the elimination of bubble defects using pentafluoropropane, which has a vapor pressure of 0.15 MPa at 25 °C. Bubble defects are unavoidable when UV nanoimprint is carried out in air. Pillars fabricated in thin resin film by UV nanoimprint in air contain bubble defects since bubbles are not eliminated by resin squeezing and only a small amount of air is dissolved in the resin. By introducing pentafluoropropane under such experimental condition, bubble elimination by gas condensation was separately investigated. The bubble elimination phenomenon is completed within a few seconds under an imprint pressure of 0.5 MPa in a sufficiently high concentration of pentafluoropropane. No bubble defects are generated in the entire imprint area of 10×10 mm2 with an imprint pressure of 0.5 MPa at a hold time longer then 20 s and a pentafluoropropane flow higher than 150 sccm.

Highly suppressed short-channel effects in ultrathin SOI n-MOSFETs

We have investigated short-channel effects of ultrathin (4-18-nm thick) silicon-on-insulator (SOI) n-channel MOSFETs in the 40-135 nm gate length regime. It is experimentally and systematically found that the threshold voltage (V/sub th/) roll-off and subthreshold slope (S-slope) are highly suppressed as the channel SOI thickness is reduced. The experimental 40-nm gate length, 4-nm thick ultrathin SOI n-MOSFET shows the S-slope of only 75 mV and the /spl Delta/V/sub th/ of only 0.07 V as compared to the value in the case of the long gate-length (135 nm) device. Based on these experimental results, the remarkable advantage of an ultrathin SOI channel in suppressing the short-channel effects is confirmed for future MOS devices.

UV-nanoimprint with the assistance of gas condensation at atmospheric environmental pressure

Nanoimprint should be carried out in an ordinary environment from the standpoint of cost-performance. However, bubble defects arise when the process is done in air and the problem can become worse with large patterns and thin resin films. UV-nanoimprint using resin droplets reduces air capture by expelling air during resin spreading and, in some cases, achieves bubble free UV-nanoimprint by air dissolution in resin. When the authors simulated the smooth spreading of a large resin droplet in wafer warping, resin spreading occurred on mold plateaus but hardly reached into mold recesses where a significant amount of air was sealed and the air in a mold recess was not completely dissolved in resin within 10 min under an imprint pressure of 0.5 MPa. In contrast, a captured gas of concentrated pentafluoropropane (CHF2CH2CF3, 1,1,1,3,3-pentafluoropropane, HFC-245fa, CAS No. 460-3-1) whose vapor pressure is 0.15 MPa at room temperature was completely condensed and dissolved within 2 s, leaving no trace of bubble ...

Elimination of pattern defects of nanoimprint under atmospheric conditions

We studied generation mechanism of bubble defects in air using molds having many recessed square patterns and photo-curable liquid polymer with a viscosity of 64 mPa s. The liquid polymer films with an initial thickness of 620–2000 nm were pressed with an imprint pressure of 0.1–0.5 MPa and cured. We found that air trapped in mold recesses is partially evacuated when liquid polymer flows out during the pressing. Bubble defects decrease with increasing liquid polymer outflow and fewer bubble defects are generated for smaller patterns. The initial polymer thickness exceeding 1 µm produces sufficient outflow to evacuate almost all the bubbles. We propose a method of condensing trapped gas, which should work well even if the liquid polymer outflow amount is small. We demonstrate this methods effectiveness experimentally using trichrolofluoromethane as the ambient gas. No bubble defects are observed in the sample imprinted using this method.

Diamond nanoimprint lithography

Electron beam (EB) lithography using polymethylmethacrylate (PMMA) and oxygen gas reactive ion etching (RIE) were used to fabricate fine patterns in a diamond mould. To prevent charge-up during EB lithography, thin conductive polymer was spin-coated over the PMMA resist, yielding dented line patterns 2 μ m wide and 270 nm deep. The diamond mould was pressed into PMMA on a silicon substrate heated to 130, 150 and 170oC at 43.6, 65.4 and 87.2 MPa. All transferred PMMA convex line patterns were 2 μ m wide. Imprinted pattern depth increased with rising temperature and pressure. PMMA patterns on diamond were transferred by the diamond mould at 150oC and 65.4 MPa, yielding convex line patterns 2 μ m wide and 200 nm high. Direct aluminium and copper patterns were obtained using the diamond mould at room temperature and 130.8 MPa. The diamond mould is thus useful for replicating patterns on PMMA and metals.

Maskless etching of a nanometer structure by focused ion beams

Microfocused heavy ion beams obtained from liquid metal ion sources of gallium, indium, and tin are bombarded onto silicon and gallium arsenide substrates, and the amorphous regions created are selectively dissolved in suitable etchants (ion bombardment enhanced etching). The area exposure doses required to etch to the depth of the calculated projected range of the incident ions are in the region of 5×10−6∼1×10−5 C/cm2 at accelerating voltages of 30∼50 kV,and the dose dependencies of the etched depths show rapid increases in specified dose regions. Widths of etched depths obtained in line delineations depend on a line exposure dose, and the minimum linewidth clearly obtained is 20∼40 nm for all the ion beams. From measurements of the dose dependencies of linewidth, the beam diameters are evaluated for various conditions of the source operation and of a lens acceptance half‐angle. The virtual crossover diameters that are independent of the half‐angle, are found to be 40∼50 nm for the gallium source and 67 ...

Conditions for Fabrication of Highly Conductive Wires by Electron-Beam-Induced Deposition

Conductive wires were fabricated by electron-beam-induced deposition (EBID) using WF6 gas. It was difficult to fabricate highly conductive wires with good reproducibility unless samples were cleaned before EBID. Contamination appears to reduce the conductivity of the wires. O2 plasma cleaning of samples before EBID seems to reduce contamination growth; however, it is not effective for regions in the vicinity of Au patterns. We found that by combining annealing at 300°C and O2 plasma cleaning, highly conductive wires could be fabricated with relatively good reproducibility in such regions. A linear relation was found between wire conductance and linedose at lindoses of more than 70 µC/cm. The change in deposition yield estimated from the conductance was about 12% when the gas flux was halved. Wires with a length of less than 40 nm were less conductive than longer wires because of a shortage in gas supply.

Measurement of virtual crossover in liquid gallium ion source

Gallium ion probes of 50 keV obtained from a liquid metal ion source are recorded on silicon and gallium arsenide substrates by using the method of ion bombardment enhanced etching. The minimum linewidth obtained in line exposures is about 20 nm for both substrates which corresponds to the lateral spread of the implanted ions. For a specified value of the lens acceptance half‐angle, the line dose varies as a square function of the linewidth, which means that the distribution of the current density in the focused beam is Gaussian. From the theoretical evaluation of the full widths at half‐maximum of these distributions, the virtual crossover diameter of the source is estimated to be 40–45 nm independently of the emission current ranging from 4 to 12 μA, and of the value of the needle apex radius of the source in the range 1–3 μm.

Photo-nanoimprinting using SOFT (sample on flexible thruster) stage

In the photo-nanoimprinting, the base layer thickness decreases with imprint pressure and imprint time; however, the decreasing speed becomes very slow when the polymer thickness is below several hundred nm. The simple idea for resulting in thin base layer is supplying enough small amount of polymer (i.e., small droplet(s) or thin film). In the droplets system, a mold is soaked with the polymer by capillary action. The soaking action results air evacuation and the soaking time limits the process speed as presented in Chen et al. (2004). In the thin film system, no soaking time is required but it is quite difficult to bring a mold in full contact with the polymer so as not to catch air since the surfaces of the mold and the polymer are uneven. To solve the problem, we developed a novel sample stage, which equips wafer warp function, compatible to such a thin initial polymer thickness.

Release force reduction in UV nanoimprint by mold orientation control and by gas environment

Release force reduction is necessary for reliable UV nanoimprint. The effect of inclined mold detachment and pentafluoropropane as an ambient gas was investigated using a UV-nanoimprint stepper with a mold orientation control and a gas introduction system. The inclined release was realized by diverting the mold orientation control for directing a mold parallel to a wafer commonly used after mold exchange. A pentafluoropropane environment was generated in UV-nanoimprinting space by a gas introduction system within a UV-nanoimprint stepper. The inclined release showed a reduction in release force in comparison to parallel release at this experiment. However, the inclined release is only modestly beneficial for release force reduction. In contrast, the ambient gas in which UV nanoimprint is carried out has a significant impact on release force. Release force can be reduced to one-third of what is obtained in air by using pentafluoropropane. It is quite beneficial to use pentafluoropropane not only for bubble...