1. Introduction

Research based on nanotechnology covers a vast variety of scientific disciplines which implies potential applications in many commercial fields, including electronics, energy production, biomaterials or medicine. Innovative functionalities are being developed based on the particular properties of the materials at the nanometer scale, e.g. metamaterials [1]. Critical to the progress of this technology is the maturation of the fabrication techniques to achieve nanoscale control of the structuring materials [2]. Thus, in many nanotechnological applications, the development of nanoscale structures is not only demanded in 2-dimensions (2D) but also in 2.5- and 3-dimensions (2.5D and 3D), which implies a big challenge for current fabrication techniques [3]. For example, in nanophotonics, which is an emerging area of this technology, fundamental understanding of the unique effects induced by the interaction of UV-visible-NIR light with sub-wavelength objects requires the 2.5D and 3D fabrication of nanometer scale patterns [4].

Among the next-generation nanofabrication techniques, a recent cost-effective alternative is direct laser writing (DLW) lithography, which has attracted a lot of attention due to its ability to fabricate real 3D structures based on a multiphoton absorption process by using tightly focused femtosecond-pulsed lasers [5]. This technique has experienced a tremendous development in the last few years accomplishing a minimum feature size on the order of 50-150 nm in the x-y plane (depending on the resist material), which renders this technique a promising candidate to overcome the limitations of current manufacturing techniques [6]. To improve even more the lateral resolution, a two-beam approach has been recently implemented in DLW lithography, generating polymerized structures with lateral feature down to a few tens of nanometers [7,8]. This approach, inspired by stimulated emission depletion (STED) microscopy, is based on the use of a second beam to selectively deplete the intermediate state of the photoinitiator, thereby confining the polymerization process to a volume of lateral sub-diffraction size [9]. Regarding axial resolution, DLW lithography can typically generate features 2 to 3 fold larger than in the lateral dimension. This decrease in resolution along the optical axis is caused by the inherent properties of focused Gaussian beams, with an axial extent as large as three times the lateral beam waist [7,10]. Although STED-like approaches have also been used to achieve an axial feature size as small as λ/20 (~40 nm) [11], technical aspects make this concept difficult to implement in practice. Therefore, DLW lithography is currently unsuitable for applications requiring real nanometer (<100 nm) axial control.

In this communication, we report a simple one-pot method to precisely control the height of 2.5D photopolymerized structures fabricated by DLW lithography. The approach is based on the thermal effect provoked by the interaction between high peak-power femtosecond (fs) pulsed lasers and the photopolymerized feature, resulting in a controlled axial expansion of the resin. This method enables to combine the nanometer feature size in the x-y plane of the DLW lithography with the precise thickness control provided by the expansion of the resin. An axial height control of 40 ± 10 nm (mean ± s.d.), which corresponds to λ/20, has been accomplished by careful selection of the excitation power of the fs laser and the irradiation time. To validate the potential of such axial nanometer control, two chessboard-like patterns containing squares with different controlled heights were fabricated and the holographic signal was measured.

2. Experimental Section

2.1. Materials

The resin used for the polymerization was pentaerythritol triacrylate (PETA, Sigma Aldrich), a negative resin. The photoinitiator employed for triggering the polymerization was isopropyl thioxanthone (ITX, Sigma Aldrich). The ATTO 565 carboxy-derivative was purchased from Sigmal Adrich. All the chemicals were used without further purification. The resist employed for polymerization was prepared by dissolving ITX (0.16 wt%) and ATTO 565 (0.025 wt%) in PETA (2 g). In a typical lithography experiment, a glass cover slip was placed on an oil-immersion objective lens and a drop of the resin was deposited on top. After irradiation with the laser, the cover slip was immersed in methanol for 5 minutes and then rinsed with isopropanol to wash away the liquid monomer.

2.2. Lithography optical setup

The optical setup is based on a custom-made optical microscope which has been previously described [12]. The two-photon polymerization beam is obtained from a mode-locked Titanium-Sapphire laser (Chameleon Vision II, Coherent, USA), which delivers pulses of ~150 fs, with a wavelength of 800 nm and at a repetition rate of 80 MHz. The beam is deflected by two fast galvanometric scanning mirrors (6215HM40B, Cambridge Technologies, USA), toward the objective lens (PL APO 100x/0.7-1.4, oil immersion, Leica Microsystems, Germany). The power of the beam is controlled with an acoustic optic modulator (AOM, AA Optoelectronics, France). The software package Imspector (Max-Planck Innovations, Germany) ensures the full control of the optical setup.

2.3. AFM Imaging

The AFM measurements were performed by using a Nanowizard III (JPK Instruments, Germany) mounted on a Nikon optical microscope (Nikon A1R MP, Nikon Instruments Inc., Japan). Monolithic silicon cantilevers (NCHR AFM probes, NanoWorld, Swtizerland), with a force constant ranging from 21 to 78 N/m, with a resonance frequency in air ranging from 250 to 390 kHz and tip with typical curvature radius of less than 8 nm were used. In the indentation measurements, a rectangular shaped cantilever with a 3 N/m elastic constant (FMG01, NT-MDT) was used.

2.4. Holography

The chessboard-like structures were placed at the back focal plane of a 100 mm lens and illuminated with a 488 nm Argon laser (35-LAP-431-230, Melles Griott) at a power of 60 mW. A CMOS camera (DCC1545M, Thorlabs) located in the focal plane of the 100 mm lens was used to detect the diffraction patterns. A spatial filter in contact with the lens was used to block the zeroth diffraction order.

2.5. Simulations

An algorithm to solve the diffraction of light using the Fraunhofer approximation was implemented in MATLAB. The main input parameters were the height profiles measured experimentally, the refractive index (a value of 1.49 was considered, from Sigma Aldrich), the wavelength of light and the focusing power of the lens. The phase differences introduced by the fabricated structure were implemented following indications from [13].

3. Results and discussion

Our method consists of two consecutive steps that are easily assembled in a one-pot process using a conventional experimental setup needed for DLW lithography. Figure 1(A) illustrates the two processing steps before the sample is developed with a washing solvent to eliminate the liquid monomer. In the first step, the desired photopolymerized structure is fabricated following the requirements of the DLW lithography [6]. In particular, a high peak-power fs pulsed excitation beam, in combination with a suitable pixel dwell time (writing speed), is needed to generate a concentration of radicals above the polymerization threshold. However, if the excitation power used for the polymerization is too high, micro-explosions due to the local increase of the temperature are observed which eventually produce the volatilization of the resin and consequent damage in the fabricated structure [14]. In the second step, the same fs pulsed beam used for the photopolymerization, but at a much lower power (below the polymerization threshold), is focused into the structure to induce a controlled expansion of the previously polymerized resin. Importantly, the axial expansion is only constrained by the power and number of exposures of the beam and not by diffraction. It is worth remarking that a fluorescent dye (ATTO 565) with high one- and two-photon cross-section was added into the resin (pentaerythritol triacrylate, PETA) with the aim of increasing the light absorption capability of the system.

 

Fig. 1 A) Schematic illustrations showing the two processing steps involved in the one-pot fabrication method to precisely control the height of the resin structures. B) Non-contact AFM image (height) of a square (15x15 μm) fabricated with two-photon lithography (100 mW) with subsequent irradiation with the same femtosecond laser at lower power (60 mW, two exposures) to generate a central elevated small square (5x5 μm). The inset shows a zoom of the elevated area with a fine detail of the line-by-line fabrication process. The image size is 20 x 20 μm. C) Height profiles along the white line drawn in panel B. D) Height of different central elevated squares obtained upon irradiation with the femtosecond laser in a photopolymerized square with an increasing number of exposure at five different powers (25, 36, 50, 60 and 76 mW).

Download Full Size | PDF

Figure 1(B) shows an atomic force microscopy (AFM) image of a simple structure fabricated with our two-step approach. Firstly, a 15x15 μm square was fabricated by DLW lithography, using the fs pulsed beam with a power of 100 mW measured at the back aperture of the objective lens. In this case, the writing speed, controlled by galvanometric mirrors, was 270 μm/s. The same fs beam but at a lower power, 60 mW, was used to scan four consecutive times the central 5x5 μm area of the initial square. The idle time between consecutive scans was 120 ms. It is worth mentioning that, in both steps, the axial position of the focus remained unchanged – it was located at the surface of the used substrate (cover slip). This assures proper attachment of the photopolymerized structure to the cover slip, avoiding washing away the solid feature during the development process. More importantly, maintaining the focal plane assures modification of the internal composition of the polymerized resin during the second irradiation step. Notably, each step of our process results in significant height differences, as shown Fig. 1(C). Whereas the first step yields to a height of ~500 nm, the second step elevated the central square by 220 ± 15 nm. In both cases the uniformity of the upper surface is determined by the roughness of the polymerized resin, which is ~7 nm.

Because the power of the beam is well below the polymerization threshold, we do not attribute the formation of the elevated area to a conventional polymerization of the remaining monomer resin, but to a controlled expansion of the resin as a consequence of heat accumulation. One can think of other phenomena that cause polymerization at powers well below threshold. For instance, the so called “memory effect” [7], i.e. accumulation of radicals during multiple exposure, or “surface effects” [15], i.e. the decrease of the polymerization threshold when irradiating on a different surface, such as the first polymerized layer instead of the glass substrate. However, both phenomena seem unlikely in our case. The long idle time between consecutive scans (120 ms) makes the accumulation of free radicals – “memory effects” – less probable. Also, the additional polymerization can partially be excluded by a close inspection of the structure. Indeed, the scanning process during the second irradiation step was performed perpendicular to the direction used to photopolymerize the initial square. In the inset of Fig. 1(B), an enlarged image of an area covering the elevated central part (bright area) as well as the initial photopolymerized square (dark area) is presented. This image provides details of the individual lines that are fabricated during the photopolymerization of the resin. The orientation of the lines across the entire figure corresponds to the scanning direction used during the first fabrication step. Such parallelism is maintained even in the elevated central area, created during the second step with the beam scanned perpendicular to the lines. Clearly, there is no interruption of the lines in between both areas which indicates that the second step only affects the height of the structure. This behavior is difficult to attribute to a conventional polymerization process. In addition, we noted that the increase in resin height appears only if, during the second irradiation step, the focus of the beam is positioned inside the structure polymerized during the first step. With the help of a piezo-electric actuator, we lifted the focus of the beam to the upper surface of the polymerized structure during the second step. In this case, we did not appreciate in the AFM images a change in the height of the central square, which suggests the absence of extra polymerization.

On the other side, the hypothesis of thermal effects as the cause for resin expansion is consistent with AFM indentation measurements performed to evaluate the possible change in the stiffness of the elevated area. Figure 2 shows a clear decrease of the Young’s modulus at the doubly irradiated area which indicates this area is softer compared to the initially polymerized structure. With all these arguments in mind, we tend to ascribe the elevation of the doubly irradiated area to the swelling of the resin due to heat accumulation. The same thermal expansion mechanism has been previously described to account for laser swelling of polymers, a process used for optical patterning [16] as well as microlens array fabrication [14,17]. However, this is the first time to use such effect for accurate axial control in DLW Lithography. In our particular experiment, the high repetition rate of our femtosecond laser (80 MHz) prevent heat diffusion throughout the resin between pulses and therefore, a local increase of the temperature occurs. If the temperature reaches a certain threshold, the cross-linked resin can suffer vaporization generating gaseous products. Thus, the combination of high temperature together with the formation of the gaseous products is thought to produce the expansion of the polymerized resin [14]. To further evaluate this assumption, the role in the expansion of the presence of the ATTO 565 dye was studied. The same example as in Fig. 1(B) with same irradiation conditions was reproduced but with the resin lacking the ATTO 565. Identical polymerized structures were fabricated during the first step, which points out the presence of ATTO 565 does not affect the polymerization process. However, after performing the second step, no expansion of the photopolymerized structure was observed. It was necessary to double the number of exposures to eight to observe an elevation of the area of only 50 ± 12 nm. Thus, ATTO 565 turns out to play a key role in the expansion process owing to its absorption of the irradiation light, which is consistent with the idea of local increase in the resin temperature as cause of the observed expansion behavior.

 

Fig. 2 AFM indentation measurements on a photopolymerized structure fabricated by our one-pot method. (A) Topography map of the structure, where the area irradiated twice corresponds to the right side of the dotted line. Heights were measured with respect to the area irradiated only once (left). (B) Corresponding colormap of the Young’s modulus of the structure shown in (A). The image field of view corresponds to 5.2 µm x 5.2 µm

Download Full Size | PDF

Due to the importance of controlling the height of the photopolymerized structures in multiple applications, the effects of the two main technical parameters – laser power and number of consecutive exposures (x-y scans) – on the degree of expansion of the resin were studied. Figure 1(D) depicts the wide variety of heights of the elevated areas (40-250 nm) that are accessible with this method by selecting the right combination of power and number of exposures. The small features sizes are comparable to those obtained with regular DLW lithography or even the two-beam approach but with a more simple procedure – without moving the focus and using a standard lithography station. The increase of height when using high powers and/or several number of exposures is not unlimited – above a certain threshold the resin is damaged. Moreover, Fig. 1(D) shows that upon increasing the number of exposure the rise of the elevated areas is less pronounced, which means that the ability of the resin to be expanded decreases upon irradiation. Such an effect is expected since the process of expansion relies on the capacity of the resin to generate volatile compounds that contribute to the enlargement of the structure. Thus, after every expansion, a lower density of the cross-linked resin remains which makes more difficult the later expansion. In addition, the absence of temperature addition between consecutive irradiations may help to explain the non-linear increase of the height with the number of exposures. A full diffusion of the heat out of the irradiated areas between sequential exposures occurs since the long idle time, ~120 ms, is much longer that the typical time constant for thermal diffusion on the length scale of the writing laser spot, ~100 ns (thermal diffusivity ~10⁻⁷ m²/s).

The versatility of our method to precisely control the height of the polymerized features is demonstrated with the fabrication of three different structures displayed in Fig. 3 (AFM images and line profiles). Figure 3(A) shows sixteen elevated squares fabricated on a large photopolymerized square (25x25 μm). Each expanded square was fabricated by scanning once a 4x4 μm area with a 50 mW laser beam. The AFM image shows that all of them present an almost perfect square shape, without uncontrolled expansion of the resin in the areas in between the squares. Moreover, the height of the elevated squares is quite consistent, 100 ± 8 nm [Fig. 3(D)]. This example proves the good reproducibility of the expanded structures in the axial direction. In the following example, a chessboard-like pattern was generated in a 25x25 μm photopolymerized square by sequential irradiation of different areas with the low-power beam. Figure 3(B) shows four squares fabricated by passing four times with a 36 mW laser beam (height = 75 ± 17 nm) and another four by irradiating two times at 76 mW (height = 270 ± 13 nm), remaining eight squares from the undisturbed resin. Despite producing a latter expansion in an area close to a previously elevated one, the shape and height of the squares are well-preserved. Thus, photopolymerized structures decorated with a variety of motifs at different heights can be fabricated with this method. The last example is a pyramidal pattern fabricated through the multiple irradiation of the same area, obtaining an extra expansion of the resin after each irradiation. Initially, a 15x15 μm polymerized square was fabricated by DLW lithography. Then, the 10x10 μm central part of the square was irradiated once with a 70 mW laser beam obtaining an elevated square with a 170 ± 15 nm height. Later, the same 70 mW beam was used to sequentially irradiate a 5x5 and a 2.5x2.5 μm areas always at the center of the initial large square [Fig. 3(C)]. Since the ability of the resin to be expanded decreases as it is expanded, the heights of the two new smaller squares were lower, 70 ± 12 and 40 ± 10 nm [Fig. 3(F)]. Thus, the same area can be expanded multiple times generating a multi-step pattern, which provides more flexibility to this method to fabricate complex patterns with different heights. These three examples clearly illustrate the ability of this method to generate reproducible, two-step or multi-step elevated features in a polymerized resin.

 

Fig. 3 Non-contact AFM images (height) of three patterns fabricated by expansion of a photopolymerized square with the femtosecond laser: (A) separated square pattern, (B) chessboard-like pattern and (C) pyramidal pattern. Height profiles (D, E and F) along the white lines drawn in panel A, B and C, respectively. The image size is 30x30 μm for image A and B and 20x20 μm for image C.

Download Full Size | PDF

The fine height control of the photopolymerized structures with feature size below the wavelength of UV-visible-NIR light achievable with our method can be exploited in a wide variety of relevant photonics applications, such as holography or transformation optics [18]. As an example to demonstrate the potential of our method, the diffraction signal from two chessboard-like patterns was collected [Fig. 4]. The first pattern was fabricated using the traditional photopolymerization approach (single irradiation), while the second one was prepared using our method. Notably, the height control of the features obtained in the first pattern in this material is limited by the axial feature size of the DLW system to about 500 nm. This is a significant constraint in many photonic applications where subwavelength control of light is required [18]. In particular, considering a wavelength of 488 nm and a refractive index of 1.5, the 500 nm axial feature size implies that phase can only be modified in fixed steps of ~3π. Contrary to traditional DLW, our method enables to fabricate photopolymerized structures with height control down to 40 nm. In this way, the phase of light can be modified in steps more than an order of magnitude smaller – for a wavelength of 488 nm and a refractive index of 1.5, phase can be controlled in steps as small as ~0.2π. The diffraction signal after illuminating the two chessboard-like structures with 488 nm light are presented in Fig. 4(B) and 4(C) for single irradiation DLW and our approach, respectively. In each case, a different diffraction pattern is obtained. This result clearly highlights the importance of fine height control for diffractive optical elements – a difference in height of only 100 nm can already result in distinct diffraction patterns. In addition, the good agreement between experiment and simulation (no fitting parameters) seems to indicate that the refractive index of the polymer has not been significantly affected by the expansion. This can be attributed to the nonlinear nature of the Lorentz-Lorenz equation that relates refractive index and density [19] – even if material rarefaction may occur during expansion, the corresponding change in refractive index is expected to be less important (Appendix A). Therefore, the combination of traditional DLW with a second irradiation step at low power opens the door to controlling light phase in a range spanning over an order of magnitude.

 

Fig. 4 A) Non-contact AFM images (height) of the chessboard-like pattern fabricated for the holography experiments. The image size is 70 x 70 μm. Height profiles of the sample with no further irradiation (blue line) and with femtosecond irradiation (red line) of the photopolymerized squares with 50 mW (one exposure). B) and C) Experimental and simulated diffraction signal of the chessboard-like patterns upon irradiation with λ = 488 nm.

Download Full Size | PDF

4. Conclusion

In conclusion, this communication reports a simple one-pot method to spatially-selective control the height of the 2.5D photopolymerized structures fabricated by DLW lithography with sub-diffraction feature size of λ/20. The basis of the method resides in the change in height exhibited by the resin upon secondary irradiation with a tightly focused laser beam. Through careful selection of the irradiation power and the number of exposures, a 40-250 nm fine control of the height of the photopolymerized structure is achieved. This exceptional nanometer control in the axial direction can be exploited in many technological fields, where structures with dimensions below the UV-visible-NIR wavelengths are required.

Appendix A Relationship between change in refractive index and density

Using the Lorentz-Lorenz [19] equation, it follows that:

(1)$$n=\sqrt{\frac{2f(\rho )+1}{1-f(\rho )}},$$

with f(ρ) being a function of the material density (ρ), which and can be written as:

(2)$$f=\frac{\rho A}{W},$$

where A is the material molar refractivity and W the molecular weight.

In the case of PETA [20], A=72.91 cm³, W=298.28 and ρ=1.18 gr/cm³. A plot of the relationship between the change in refractive index versus density variations is presented in Fig. 5

 

Fig. 5 Plot of the expected change in refractive index as a function of the variation in density for PETA.

Download Full Size | PDF

Acknowledgment

The authors thank Dr. Claudio Canale and Reinier Oropesa for the support with the AFM measurements. G. M. is grateful to the European Commission for a Marie Curie CIG grant (Nº 631316).

References and links

1. S. Larouche, Y.-J. Tsai, T. Tyler, N. M. Jokerst, and D. R. Smith, “Infrared metamaterial phase holograms,” Nat. Mater. 11(5), 450–454 (2012). [CrossRef]   [PubMed]  

2. H.-D. Yu, M. D. Regulacio, E. Ye, and M.-Y. Han, “Chemical routes to top-down nanofabrication,” Chem. Soc. Rev. 42(14), 6006–6018 (2013). [CrossRef]   [PubMed]  

3. B. D. Gates, Q. Xu, M. Stewart, D. Ryan, C. G. Willson, and G. M. Whitesides, “New approaches to nanofabrication: molding, printing, and other techniques,” Chem. Rev. 105(4), 1171–1196 (2005). [CrossRef]   [PubMed]  

4. M. S. Rill, C. Plet, M. Thiel, I. Staude, G. von Freymann, S. Linden, and M. Wegener, “Photonic metamaterials by direct laser writing and silver chemical vapour deposition,” Nat. Mater. 7(7), 543–546 (2008). [CrossRef]   [PubMed]  

5. C. N. LaFratta, J. T. Fourkas, T. Baldacchini, and R. A. Farrer, “Multiphoton fabrication,” Angew. Chem. Int. Ed. Engl. 46(33), 6238–6258 (2007). [CrossRef]   [PubMed]  

6. M. Malinauskas, M. Farsari, A. Piskarskas, and S. Juodkazis, “Ultrafast laser nanostructuring of photopolymers: A decade of advances,” Phys. Rep. 533(1), 1–31 (2013). [CrossRef]  

7. J. Fischer and M. Wegener, “Three-dimensional optical laser lithography beyond the diffraction limit,” Laser Photonics Rev. 7(1), 22–44 (2013). [CrossRef]  

8. Z. Gan, Y. Cao, R. A. Evans, and M. Gu, “Three-dimensional deep sub-diffraction optical beam lithography with 9 nm feature size,” Nat. Commun. 4, 2061 (2013). [CrossRef]   [PubMed]  

9. J. Fischer, G. von Freymann, and M. Wegener, “The materials challenge in diffraction-unlimited direct-laser-writing optical lithography,” Adv. Mater. 22(32), 3578–3582 (2010). [CrossRef]   [PubMed]  

10. S. Juodkazis, V. Mizeikis, K. Seet, M. Miwa, and H. Misawa, “Two-photon lithography of nanorods in SU-8 photoresist,” Nanotechnology 16(6), 846–849 (2005). [CrossRef]  

11. L. Li, R. R. Gattass, E. Gershgoren, H. Hwang, and J. T. Fourkas, “Achieving λ/20 Resolution by One-Color Initiation and Deactivation of Polymerization,” Science 324(5929), 910–913 (2009). [CrossRef]   [PubMed]  

12. B. Harke, W. Dallari, G. Grancini, D. Fazzi, F. Brandi, A. Petrozza, and A. Diaspro, “Polymerization Inhibition by Triplet State Absorption for Nanoscale Lithography,” Adv. Mater. 25(6), 904–909 (2013). [CrossRef]   [PubMed]  

13. J. W. Goodman, Introduction to Fourier Optics (Roberts and Co. Publishers, 2005).

14. T. Meunier, A. B. Villafranca, R. Bhardwaj, and A. Weck, “Fabrication of microlens arrays in polycarbonate with nanojoule energy femtosecond laser pulses,” Opt. Lett. 37(20), 4266–4268 (2012). [CrossRef]   [PubMed]  

15. D. Kunik, S. J. Ludueña, S. Costantino, and O. E. Martínez, “Fluorescent two-photon nanolithography,” J. Microsc. 229(3), 540–544 (2008). [CrossRef]   [PubMed]  

16. E. Mcleod and C. B. Arnold, “Subwavelength direct-write nanopatterning using optically trapped microspheres,” Nat. Nanotechnol. 3(7), 413–417 (2008). [CrossRef]   [PubMed]  

17. J. Shao, Y. Ding, H. Zhai, B. Hu, X. Li, and H. Tian, “Fabrication of large curvature microlens array using confined laser swelling method,” Opt. Lett. 38(16), 3044–3046 (2013). [CrossRef]   [PubMed]  

18. J. B. Pendry, A. Aubry, D. R. Smith, and S. A. Maier, “Transformation optics and subwavelength control of light,” Science 337(6094), 549–552 (2012). [CrossRef]   [PubMed]  

19. M. Born and E. Wolf, Principles of Optics (Cambridge U., 2003).

20. http://www.lookchem.com/Pentaerythritol-triacrylate/