Microlight 3D study

In order to help other users to tune their recipes to achieve the desired profile a parameter study was conducted. This study aims to give a better understanding of the effects that different parameters have in the thickness and profile of photolithography depositions.

Table 1: Parameters measured

Expose Dose (mJ)	130, 150, 170, 190, 210, 230, 250, 270, 290, 310
Spinner speed (rpm)	1500 ,3000, 4500
Developing time (sec)	20, 45, 70
Bake time	1min (No Lid), 1 min (with Lid), 5min(No lid), 5 min (with Lid)

 

Expose dose:

Exposure dose quantifies the light intensity used in the evg 620 and is measured in mJ. 

In the 7µm and 10µm lines Exposure dose seems to have no effect on the thickness of the photoresist. For 5µm lines thickness stays close to the 7µm and 10µm thickness, but doses higher than 190mJ seem to have a sudden decrease in height. This is due to the breaking of the top part of out profile (5µm line defects), leading to inconsistent heights and lower than expected heights.

The angle between the silicon substrate and the photoresist is a representation of the geometrical shape of our lines. Angles close to 90 will look completely square while angles close to 120 will result in triangular profiles. As seen in the graph developing dosage has no effect in the angle, which remains in the 110-105 degree region.

10 $\mu$m

130mJ

150 mJ

170 mJ

 

190 mJ

210 mJ

230 mJ

 

250 mJ

270 mJ

290 mJ

 

310 mJ

7 $\mu$m

130mJ

150 mJ

170 mJ

 

190 mJ

210 mJ

230 mJ

 

250 mJ

270 mJ

290 mJ

 

310 mJ

5 $\mu$m

130mJ

150 mJ

170 mJ

 

190 mJ

210 mJ

230 mJ

 

250 mJ

270 mJ

290 mJ

 

310 mJ

 

Developer time:

Developing removes the uncured photoresist from our sample, leaving the desired cured pattern intact. Underdeveloped samples tend to have visible photoresist left, while overdeveloping might damage or features and cause them to detach from the wafer.

Developing time seems to have no effect on 7µm and 10µm lines but the 5µm displayed a broken top part only at 45sec and 70 sec, while being intact at 20 sec.

10 $\mu$m

20sec

45 sec

70 sec

7 $\mu$m

20 sec

45 sec

70 sec

5 $\mu$m

20 sec

45 sec

70 sec

 

Spinner Speed:

As we increase spin speed less photoresist will remain in our wafer and thus our film thickness will decrease. Again, both 10µm and 7µm lines display a perfect linear relationship between spin speed and thickness, but the 5µm line displayed a broken top porting at 1500rpm and 3000rpm, giving the results we see below.

Spinner speed looks to have a significant effect in the angle between the silicon substrate and the photoresist lines. Higher speeds yield a angles of 98 degrees, close to a square profile as seen in the images. Thus increasing spinner speed past 3000rpm has a beneficial effect in the square shape of our lines.

10 $\mu$m

1500 rpm

3000 rpm

4500 rpm

7 $\mu$m

1500 rpm

3000 rpm

4500 rpm

5 $\mu$m

1500 rpm

3000 rpm

4500 rpm

 

Bake Time:

After depositing the liquid layer of photoresist it is baked in a hot plate at 110 C° in order to remove the solvent present in the photoresist. Higher bake times result in harder, more brittle photoresist while lower bake times result in softer more ductile photoresist. An aluminum lid can be used to enclose the sample and better distribute the heat. As seen, bake time seems to have no effect on PR thickness.

5 $\mu$m

1min no lid

1 min with lid

5 min no lid

 

5 min with lid

7 $\mu$m

1min no lid

1min with lid

5 min no lid

 

5 min with lid

10 $\mu$m

1min no lid

1 min with lid

5min no lid

 

5 min with lid

 

5 µm Line defects:

Looking at the data we see that the thickness of the 5µm lines tends to suddenly drop in developing time at 45 sec and exposure dosage at 210 mJ. This is due to a crack of the top part of our Photoresist lines as seen below:

Here the Left PR line shows a common defect at the very end of it, while the line next to it preserves its original shape. In Exposure dosages over 190, the whole line shows this defect from end to end making the measurements suddenly drop.

 

Line Wall Defects:

At high exposures or high developing times we also see a serrated edge defect in out patterns as shown below:

This defect could be due to overexposure of the photoresist, making it develop irregularly along the edges. This pattern decreases in magnitude but it is still visible at low doses (130mJ).