This section, including all images and pictures, is created by DTU Nanolab staff unless otherwise stated.

Feedback to this page: click here

Exposure technology

Aligner: Maskless 04 is both a direct writer and not a direct laser writer. In Raster mode, it works like MLA 1, where an image is projected onto the substrate surface, and stepped across the substrate in order to produce the entire pattern. In Vector mode, it is a direct laser writer, where a focused laser beam is moved across the substrate surface in order to trace out the pattern.

The writing head of the Aligner: Maskless 04 moves only in the z-direction. Using a focusing system, the maskless aligner is able to do real-time autofocus. The defocus process parameter is used to compensate offsets in the focusing mechanism, and to optimize printing quality in different resists and varying thicknesses.
The stage of the Aligner: Maskless 04 moves only in x and y. It has no theta-axis. All rotation during alignment is thus accomplished by transformation of the input design.

Raster mode

The light source in Raster mode is a 6W 365nm LED. The exposure light is passed through a spatial light modulator in the form of a digital micro-mirror device. The individual mirrors of the DMD are switched in order to represent an area of the design, and are timed in order to yield the desired exposure dose, while taking into account illumination uniformity, soft-stitching, and possibly also sub-pixel interpolation. This image is projected onto the substrate through a lens(system). The projected image yields a pixel size of 0.5µm X 0.5µm at wafer scale. This writing field is stepped across the substrate, in order to expose the entire design, each field overlapping slightly in order to minimize stitching errors.

Vector mode

The light source in Vector mode is a 120mW 405nm diode laser. The laser beam is focused onto the substrate through a lens(system). The size of the focused spot is controlled using a set of pinholes (1, 2, 5, 10, and 25µm). By moving the stage at a constant velocity while switching the laser on and off, the pattern represented in the design is exposed in the resist. The dose the resist will receive is a function of the output power of the laser, the size of the pinhole, and the velocity of the stage.

Process Parameters

Control of the exposure dose differs between the two available exposure modes. In Raster mode, the dose is directly controlled by the on-time of the DMD mirrors during exposure, and the dose process parameter is thus simply given in units of mJ/cm². As the intensity of the exposure light at the substrate surface cannot easily be measured, this dose is indirectly calibrated at the factory, by adjusting the optimal dose to match the known dose of a particular resist.
In Vector mode, however, the exposure dose is determined by combination of four different process parameters:

Pen [1, 2, 5, 10, or 25 µm]; the size of pinhole projected onto the substrate surface
Transmission [30, or 100 %]; the transparency of the filter in front of the pinhole
Laser Power [0-120 mW]; the output power of the diode laser
Exposure Velocity [0-200 mm/s]; the movement speed of the stage during exposure

The Pen determines the size of the spot on the substrate surface and thus the ultimate resolution. It may also affect the intensity available at the substrate. For highly sensitive resists, the Transmission can be used to limit the available intensity at the substrate (lowering the required Exposure Velocity in a more reasonable value). The Laser Power directly affects the intensity available at the substrate. The Exposure Velocity determines the dwell time at each point on the substrate, and thus the effective dose received by the resist at the intensity determined by the three other parameters.

Aligner: Maskless 04 offers not only two exposure modes, but also two autofocus modes; optical or pneumatic. The defocus process parameter is used to compensate for offsets between the autofocus mechanism and the focal point of the exposure light, and simultaneously optimize print quality in different resists and varying thicknesses. The "defoc" parameter us a unitless value, representing ±100% of the available correction available above and below the initial focus point established during loading of the substrate. One defoc step is approximately 0.25µm. Positive defoc is into the resist, i.e. writehead moves down when defoc is increased.

Optical autofocus: Uses red laser light to focus on the substrate surface. Works for substrates down to 3x3 mm.

Pneumatic autofocus: Uses compressed air flowing through the nozzle of the writehead to focus on the substrate surface. Substrates must be at least 5x5 mm to be successfully loaded. The pneumatic AF fails at some distance from the substrate edge, which means that in order to have any useful area in the center of the sample when using the pneumatic AF, the substrate must likely be larger than 5x5 mm. When using the pneumatic autofocus, we recommend a substrate size of at least 10x10 mm.

Exposure dose and defocus

Raster mode

	Date	Thickness	Exposure mode	Dose	Defoc	Resolution	Comments
AZ 1500	2024 FAT/SAT	0.5 µm	Raster (365nm)	80 mJ/cm²	10 (optical)	1 µm	4" mask plate with Cr Development: 60s in 1:4 AZ 351B
AZ 5214E	2024 During installation	1.5 µm	Raster (365nm)	85 mJ/cm²	15 (optical)	>1 µm	4" Si wafer Development: 60s in 1:4 AZ 351B

Vector mode

	Date	Thickness	Exposure mode	Dose	Defoc	Resolution	Comments
AZ 1500	2024 FAT/SAT	0.5 µm	Vector (405nm)	Pen: 1µm Transmission: 100% Laser Power: 120 mW Velocity: 17 mm/s	30 (optical)	1 µm	4" mask plate with Cr Development: 60s in 1:4 AZ 351B

Writing speed

Raster mode

The Raster mode will scan the entire exposure area and write with high fidelity and accuracy. The write time of a given design will only depend on the area and exposure dose, not on the number, shape, or size of structures in the area.

The site acceptance test performed during the installation of the Aligner: Maskless 04 showed an exposure speed of 47mm²/min, which is slightly higher than the 40mm²/min given in the specifications. At such speeds, a full 4" design would take 2:45-3:15 hours to expose.

The exposure time increases linearly with exposure dose and writing area. However, due to the stepped nature of the exposure, the exposure time as a function of fill factor is highly nonlinear. It takes the same time to expose a single pixel as the entire writing field, so the exposure time depends on the number of addressed writing fields, rather than on the fill factor of the design. In practice, there will probably not be much variation in exposure time with fill factor. Exposure tests have shown that the exposure speed is low for small areas, but reaches full speed at areas above 40x40mm². The lower speed for small areas is probably due to the delay in exposure start caused by the generation of the data needed for the exposure (conversion). Exposure tests using a 25mm² design have shown that the exposure time increases linearly with dose, from 37s at 10mJ/cm² to 252s at 2000mJ/cm².

Vector mode

The Vector mode will write structures line by line. Structures will have low edge roughness and there will be no stitching of long lines. However, write time of a given design will strongly depend on the number, shape, or size of structures.

The exposure time is a function of the number of elements (lines) in the design and the stage velocity during exposure. While a very low Exposure Velocity will of course slow down the exposure speed, a high velocity will also lower the exposure speed because the stage has to accelerate longer before the exposure of each element can start. The maximum velocity of the stage is 200mm/s but Heidelberg recommend not to exceed 120mm/s, in order to avoid distortions of the elements due to acceleration effects.
During exposure tests of different sizes of simple square designs, the optimal stage velocity (in terms of exposure time) increased linearly from 15mm/s for 0.5mm elements to 50mm/s for 5mm elements. The highest "exposure speed" was just below 4 elements per second, which was observed for the 0.5mm elements. In practice, exposure can't always be at the optimal Exposure Velocity, as it may be necessary to decrease the velocity in order to achieve the required dose.

Resolution

Raster mode

The pixel-size of the DMD in the Aligner: Maskless 04 is 0.5µm X 0.5µm (at the substrate surface). The resolution specification of the machine is 1µm minimum feature size and 1.5µm minimum lines and spaces, which was demonstrated in the factory acceptance test and the site acceptance test after installation using a resist thickness of 0.5µm. In 1.5µm thick resist, the resolution is expected to be around 2µm.

Vector mode

The smallest pinhole in the Aligner: Maskless 04 is 1µm. The resolution specification of the machine is 1µm minimum feature size, which was demonstrated in the factory acceptance test using a resist thickness of 0.5µm.

Substrate positioning

During load, the machine will focus on the surface of the sample. Then, using the automatic focusing system, it will detect the edges of the sample (this function depends on the substrate template used) in order to determine the center and rotation of the sample. The following results rapport findings using the "Wafer (d=100 mm; Flat)" template on a standard 100mm Si substrate.

Substrate centering

During (4") substrate detection, the sample is scanned diagonally. From these measurements, the diameter of the substrate is calculated, as well as the stage position matching the center of the substrate. This stage position will be the default origin for the subsequent exposure.
The result of testing the centering on a couple of wafers has shown offsets of the printed pattern from the substrate center in excess of 0.5mm. This is worse than the results observed on Aligner: Maskless 01, which has shown a centering accuracy of ±0.2mm.

Flat alignment

At the end of (4") substrate detection, the sample is scanned twice along the flat, in order to determine the substrate rotation. This angle will be presented before exposure start along with the option to expose the design rotated in order to compensate for this angle, i.e. aligned to the flat.
The result of testing the flat alignment on a couple of wafers has shown the alignment to be within ±0.1° (1.7mRad). This matches the results observed on Aligner: Maskless 01.

Repeatedly loading the same wafer, without moving it on the stage, gave values within 0.01mRad for both pneumatic and optical autofocus. A single outlier out of 13 measurements gave an error of 1.6mRad, corresponding to 0.1°. The measured angle, however, differed between pneumatic and optical. Investigating further, measuring the angles using different starting points (i.e., moving the wafer on the stage between tests), showed that the angle measured using optical autofocus was consistently higher than using pneumatic, by 3±1.6mRad, corresponding to 0.2±0.1°.

Alignment

The alignment accuracy of the Aligner: Maskless 04 is a combination of the position accuracy of the stage, the accuracy of the alignment mark detection, and the distortion of the pattern already on the wafer (first print). The alignment specification of the machine is 1µm for both raster and vector mode, which was demonstrated in the factory acceptance test and the site acceptance test after installation.
An alignment test performed shortly after installation, using a 3x3 array of structures with 30mm pitch, showed an alignment error of 0.29±0.18µm in X and -0.49±0.23µm in Y (Raster mode).