1. Introduction

In recent years, two-photon photopolymerization has been shown to be a promising technique for micro- and nanofabrication [1]. With its intrinsic three-dimensional (3D) fabrication capability [1–3] and sub-diffraction-limited spatial resolution [4], it has been used to fabricate different kinds of microstructures such as photonic crystals [5–7], micromechanical devices [8], and other elements [5, 9–12]. These structures might be made with different scanning procedures, but most of them consist of lines such as those of the logpile photonic crystal. To control the lineshape, we need to investigate the influence of incident energy, scan speed, and focal position. In addition, it is easier to investigate the polymerization process with lines rather than single voxels as the elemental unit [13, 14], because single voxels are too small to survive developing. Although many researchers have studied the dependence of line shapes on the exposure and other parameters, to our knowledge no results on wavy lines have yet been reported. In this paper, we present our experimental results on the observation of wavy lines in two-photon photopolymerization microfabrication. Femtosecond laser pulses were focused onto two-photon absorption resin on a glass substrate. As the sample was moved closer to the focusing lens, the height of the focus increased, and we found that the polymerized lines would remain straight, become wavy, and even float away. The dependence on the focus height, the incident laser energy, and the scan speed was studied. Furthermore, the longitudinal size of a voxel was estimated from the focus height where the lines become wavy and float away.

2. Experiments and results

A regeneratively amplified Ti:sapphire laser (Spitfire, Spectra Physics) was used as an excitation source. The 120-fs pulses were at 800 nm with a repetition rate of 1 kHz. The incident energy was attenuated with a half-wave plate before a polarizer. Then the laser beam was filtered and focused by a numerical-aperture- (NA-) tunable objective lens (UPlanAPO, Olympus) onto a liquid resin sample (SCR 500) on a glass substrate mounted upon a 3D computer-controlled piezoelectric translator. The whole process was in situ monitored with a CCD camera.

Figure 1 shows two typical wavy lines in two-photon photopolymerization. The two lines were fabricated with the same incident energy and scan speed but with different focus heights. The focus height for the lower line in Fig. 1 was higher or the focus was farther from the glass substrate, resulting in a longer period and a larger bend.

 

Fig. 1. Top view of two typical wavy lines in two-photon photopolymerization. The line spacing is 5 μm.

Download Full Size | PDF

To investigate the influence of the focus height, the incident energy, and the scan speed, a ladder-like structure was used as shown in Fig. 2. The near-infrared femtosecond laser pulses were focused onto the sample with a microscope objective under the substrate. In the experiment, the sample was moved by a translator but the objective lens was fixed. The focus height or the center of the focal spot ascended as the the sample was moved closer to the objective, i.e., by lowering the sample in the -z direction. First, two long straight lines (lines M and N) were fabricated along the x direction. They worked as markers and two sides of the ladder-like structure. Then a series of rungs were polymerized along the y direction under the same radiation conditions, but the focus height was increased line by line.

 

Fig. 2. Schematic of a polymerized ladder-like structure. The focus height can be increased by lowering the sample in the -z direction. (a) Side view showing the cross section of the polymerized lines. The near-infrared femtosecond laser pulses were focused into the SCR 500 resin by a microscope objective under the cover glass substrate. Gray parts, cross section of a polymerized line; orange ellipsoids, focal spots at different heights. (b) Top view showing straight and wavy lines. From left to right, the focus height ascended line by line.

Download Full Size | PDF

We also observed wavy lines between the fixing and floating straight lines, which may be attributed to the different voxel shapes. An isolated voxel was like a spinning ellipsoid, but it could be truncated [13]. The cross section of a polymerized line consisting of overlapped voxels would change correspondingly. The gray parts in Fig. 2(a) indicate the cross section of a line. If the focal spot was under the interface between the resin and the substrate, no voxels or lines could be fabricated. When the laser was focused too near the interface, the voxel was truncated. When only a small part of the focal spot was in the resin, a fine line (line 1) was straight and stuck to the substrate. With increased focus height, the voxel became larger and a thick straight line (line 2) was formed. However, when most of the focal spot was in the resin, a slim voxel or line (line 3) tended to topple down so that the solid line began to corrugate from the interaction with the surrounding liquid resin. The higher the focus height, the slimmer the line. Therefore, the adhesion to the substrate became weaker and the wavy line (line 4) had a larger bend. When the focus was far above the interface, the line (line 5) floated away but remained straight until it touched the substrate. In addition, we could obtain the longitudinal information of a voxel, which is difficult to measure. According to the above analysis, the focus height difference between the first visible line and the first floating line is a bit larger than the longitudinal size of an isolated voxel. In other words, we can estimate the longitudinal size of an isolated voxel from the floating height.

First we used an objective lens of NA = 1.35 to study the line changes. With the scan speed of 1 μm/s and incident energy of 4 nJ/pulse, the line width is ~400 nm. Wavy lines appeared at the wavy height h_w of ~0.6 μm and floated away at the floating height h_f ~ 0.9 μm. The wavy height h_w and the floating height h_f were calculated from the focus height where the line became visible from the CCD. When the energy was changed to 4.5, 5.0, and 5.5 nJ/pulse, h_w was approximately 0.8, 1.0, and 1.3 μm and h_f was approximately 1.2, 1.6, and 2.5 μm, respectively. With increased power, the longitudinal and the lateral size of a voxel expanded [14]. As a result, the lines became thicker and at a higher height became wavy.

To observe the phenomena clearly, we used a lower-NA objective (NA = 0.85) to get thicker lines. The line width was ~1 μm. With energy at 8 nJ/pulse, we found that the wavy lines appeared at a height h_w of 3 μm, and the lines floated at a height of 5 μm; with energy at 9 nJ/pulse, h_w and h_f are 5 μm and 8 μm, respectively. No wavy lines were observed at the height of 6 μm with energy at 10 nJ/pulse. Typical optical microscopy and scanning electron microscopy (SEM) images of the straight and wavy lines are presented in Fig. 3. The incident energy was 6.5 nJ/pulse, and the scan speed was 1 μm/s. From lines 1 to 16, the focus height increased 0.3 μm, line by line. The line spacing was 5 μm. The three SEM images on the top row showed parts of the ladder-like structure after development in ethanol for ~2 min.

 

Fig. 3. Straight and wavy lines in two-photon polymerization microfabrication. From lines 1–16, the focus height increased 0.3 μm line by line. The line spacing is 5 μm. The three SEM images on the top show parts of the ladder-like structure after development.

Download Full Size | PDF

The focus height of line 1 ascended 0.9 μm from position where the first visible line was observed. As shown in Fig. 3, line 1 kept straight and adhered to the substrate before and after development. Estimated from line 2, the wavy height h_w was ~1.2 μm. We found pronounced corrugation of line 3. After development, its top part was distorted by the pile of lines floating from other places. From lines 4 to 13, the corrugation became increasingly pronounced. With increased focus height, the line had fewer but larger bends during fabrication. After development, lines 4–6 twisted and partly toppled down. Lines 7 and 8 toppled down except for two twisted ends. However, lines 9–12 did not topple down completely as expected because they huddled together and supported one another. Line 14 floated, and the corresponding focus height was 4.8 μm. During fabrication, lines 15 and 16 floated and altered orientation but kept straight. They even moved again after development. We can estimate that the longitudinal size of a voxel or the cross section size of a line were ~4.8 μm from the floating height. The actual size was ~4.1 μm measured from the SEM image.

Under energy at 8 nJ/pulse, h_w dropped from 3 to 1 μm when we changed the scan speed from 1 to 2 μm/s. With increased scan speed, fewer voxels overlapped and it was easier for the lines to topple down and become wavy. At the same time, when the incident energy decreased, the voxel size was reduced and h_w became smaller.

When the line is tens of micron long, it can become wavy easily. In contrast, the line is hard to bend in a 3D structure in which the period is small and the lines stick to one anther. In general, the structure begins with the first layer that adheres to the substrate. Then the other layers are photopolymerized layer by layer, so each fabricated layer adheres to the previous layer. At the same time, the line spacing is small. In this case, the structure becomes a strong framework and the line can hardly become wavy. Figure 4 shows a four-layered structure consisting of lines at different directions. The lines were parallel in the same layer, but their orientation changed 45 deg from the previous layer. The layer spacing was 1 μm, and line spacing was 2 μm. Before and after the development, no wavy line was observed.

 

Fig. 4. SEM image of a four-layered structure. The layer spacing is 1 μm, and line spacing in the same layer is 2 μm.

Download Full Size | PDF

3. Conclusion

In conclusion, wavy lines in two-photon photopolymerization microfabrication were observed. Under the same irradiation and scan speed, the polymerized lines would remain straight, become wavy, and even float away with increased focus height when the sample was moved closer to the focusing lens. Because a line consists of overlapped voxels and an isolated voxel cannot stand alone, the slim line would topple down and become wavy when the focal spot was high enough above the cover glass substrate. The higher the focal spot, the fewer and larger the bends of the wavy line. With increased scan speed, fewer voxels overlap, and it is easier for the lines to corrugate. When the incident energy decreases, the voxel sizes are reduced and the lines become wavy at a lower focus height. This is applicable to controlling lines in two-photon polymerization micro- and nanofabrication.