Micro-patterning of 3D colloidal photonic crystals via solvent-assisted imprint lithography

Abstract

Scalable patterning of photonic crystals is critical in building advanced photonic devices. In this communication, we applied solvent-assisted imprinting technique to define the patterns of pre-formed colloidal photonic crystal (CPC) films. The minute solvent within the PDMS stamp along with the conformal pressure partially melts and relocates the polymers, thereby forming large-scale micropatterns in the CPC.

---

Photonic crystals, with prominent light trapping and waveguiding properties, are recognized as powerful tools for controlling the flow of light¹ and have become one of the promising candidates for advanced photonic devices of the next generation.² To achieve such advanced and highly integrated photonic devices, the state-of-the-art nano/microfabrication (nano/micropatterning) techniques have been strongly applied for low-cost and large-scale production. However, the real challenge towards this end is the difficulty in realising complete photonic bandgap and the unintentional defects generated during fabrication.³ The main effort to date, as it seems, has been diverted to sensing and image displays.⁴

Traditional top-down approaches (lithography) have mostly been applied by the electronic industry for integrated circuits and chips. With the help of these microfabrication technologies, we can build photonic cavities, optical waveguides and photonic crystal patterns at designated locations.⁵ However, most of these fabrication processes are complicated and time-consuming, and the photonic band gaps usually fall within near infrared or microwave ranges, which is limited by the resolution of photolithography. Moreover, expensive and sophisticated instruments are always needed to achieve the desired photonic structures. On the contrary, bottom-up (self-assembly) approach provides a unique opportunity for cost-effective and large-scale production of colloidal photonic crystals (CPCs) with band gaps falling in the visible range. However, the controllability of desired patterns is poor, and unintentional defects may be introduced. In order to make a complement to the drawbacks, it is more effective to combine both approaches wisely towards various hierarchical micro/nanopatterns and complex nanostructures.⁶

The typical combination for the patterning of CPCs is based on the template-directed self-assembly.⁷ The templates can be physical or chemical templates, which are fabricated with lithography techniques. Under the direction of the templates, the colloidal particles are self-assembled at desired locations, which are predefined by the templates. These locations are usually physically confined spaces or hydrophilic domains such that the colloidal particles can be self-assembled only within these regions via capillary forces. However, the dimension of the templates should be properly designed to fit the packing geometry of the spheres; otherwise it may induce packing defects within the CPCs.⁸ Such a sophisticated design has severely limited the generality of the templates for colloidal particles with different sizes, thereby limiting the robust patterning of photonic crystals with different band gaps.

Alternatively, the patterns can be created on the surface of the colloidal crystal film via lift-up soft lithography.⁹ The patterned poly-(dimethylsiloxane) (PDMS) stamps can be functionalized as a sticker for the colloidal particles on the surface. When lifting up the stamp, the colloidal patterns can be transferred onto another substrate, and the negative patterns can be formed on the surface of the colloidal crystal film. Because the colloidal crystal films were formed without any interference from the templates, the quality of the film is relatively high and uniform. However, the lift-up lithography only created the monolayer patterns on the surface of the colloidal crystals, and therefore no obvious patterns of the photonic structure colour can be observed. Herein, we applied the solvent-assisted imprinting technique to define the patterns in the CPCs. Instead of lifting up the polystyrene (PS) beads on the surface layer of the colloidal crystal films with PDMS, we attempted to imprint the colloidal crystal films with patterned PDMS stamps, which contain only a minute amount of toluene. The polymer beads (poly(St-MMA-AA) in our case) in contacted regions were partially dissolved, and the melted polymers were re-localised within the colloidal crystal films. As a result, the structure colour in the contacted regions faded away due to the lower refractive index contrast, whereas in the non-contacted regions, the poly(St-MMA-AA) beads were softened and could easily be squeezed into bumps because of the lateral deformation in the neighbouring regions caused by normal compression of the PDMS stamp. Due to this, reasonably, compared to the template-directed micropatterning techniques, solvent-assisted imprinting technique can minimize the non-uniformity of arrays and thickness of the colloidal crystals localised in different patterning regions. Moreover, in contrast to the traditional thermal imprint lithography, no instruments are required in this method.¹⁰ Thus, this facile and versatile micropatterning technique can be very effective for large scale patterning of photonic crystals.

It is critical to modulate the amount of solvent within the PDMS matrix; otherwise the whole colloidal crystal films will be melted into homogeneous ones with no photonic band gap. We also adopted the solvent evaporation method to fine-tune the amount of toluene within the PDMS matrix and minimize the swelling effect of the mould. Although it is challenging (and also not so necessary) to quantify the exact amount of toluene left within the PDMS matrix that applied for etching, we can control the evaporation time of toluene to fix the proper time window for the patterning with just several rounds of trial and error.

It is feasible to direct the pattern formation by imprinting (partial melting and relocating of the polymers within the colloidal crystal films), but it is also risky in melting the entire photonic nanostructures. Therefore, it is crucial to tune the evaporation time to control the amount of toluene applied for etching. We started from the evaporation time of 2.5 min, and then the toluene swelled PDMS stamp was pressed against the colloidal crystal films for 5 min. Consequently, the surface of the colloidal crystal films was moulded with patterns of the PDMS stamp as shown in Fig. 1. The original polymer colloidal crystals composed of poly(St-MMA-AA) particles with a diameter of ∼240 nm showed a green structure colour (see ESI the inset of Fig. S1†). After the solvent-assisted imprinting process, the structure colour disappeared (see ESI Fig. S2†). The SEM images in Fig. 1 further revealed that the poly(St-MMA-AA) particles were merged into solid polymer film with micropatterns moulded by the stamp. This is helpful in explaining the disappearance in structure colour because there was no refractive index contrast in the system. However, we could still find some submicropatterns of spherical dots on each micropattern (see the inset in Fig. 1a and c). It is noteworthy that there was no capillary lithography¹¹ in the case of moulding the colloidal crystal films, which is mainly because of the porous features of the colloidal crystal films. The nanopores in the colloidal crystals is considerably smaller than the holes in the PDMS stamp, thereby generating stronger capillary forces, which sucked the partially dissolved polymers into the interstice of the colloidal crystals instead of the holes of the PDMS stamp. As a result, the regions contacted with the stamp would be softened and compressed, whereas the non-contacted regions would protrude out to form the micropatterns with no capillary lithography effect. The statistic measurement (Fig. 1d) of the sizes of the micropatterns also presented good monodispersity (relative standard deviation, height: 8.4%, diameter: 6.4%) with an average height of ∼900 nm and diameter of ∼2.33 μm.

Fig. 1 SEM images of the moulded colloidal crystal films after 2.5 min solvent evaporation, (a) top view, (b) side view, (c) magnified side view. Scale bar in (a) 20 μm, inset 2 μm; in (b) 5 μm; in (c) 1 μm. (d) Size distribution of the micropatterns on the surface of the colloidal crystal films.

In order to maintain the photonic structure, as well as produce the micropatterns, it is important to further decrease the amount of toluene applied for micromoulding lithography. When the evaporation time of toluene was extended, the micropatterns could be formed with the presence of structure colour as shown in Fig. 2. Fig. 2a is the micropatterns formed after toluene evaporation for 3 min. The structure colour (inset of Fig. 2a) in the nonpatterned regions (the regions contacted with PDMS mould) changed to dark green, and the colour of patterned regions (the regions that are not contacted with PDMS mould) turned orange. From the SEM images in Fig. 2a and b, we can observe the patterned regions had colloidal photonic structures, while the rest are merged into continuous film. However, this continuous film formed only on the top layer of the colloidal crystals, whereas the layers below still preserved the colloidal photonic structure, which can be seen from the cross section view in ESI Fig. S3a.† It was possible only when the toluene released from the PDMS matrix was enough to dissolve the top layer of the poly(St-MMA-AA) particles into continuous film but not sufficient to melt the whole colloidal crystals. Therefore, the structure colour can be preserved.

Fig. 2 SEM images of micromoulded poly(St-MMA-AA) colloidal crystal films with solvent evaporation time of 3 min (a) and 2 min (c). (b and d) are the magnified views of the individual micropattern. Scale bars in (a) and (c) are 5 μm, and in (b) and (d) they are 1 μm. The insets in (a) and (c) are the corresponding optical images of the colloidal crystal micropatterns, and scale bars are 20 μm.

We can qualitatively predict the change of structure colour with Bragg's law:⁹

mλ = 2nd₁₁₁ (1)

where m is the order of Bragg diffraction; λ is the wavelength of the stop band, which also indicates the reflection colour; d₁₁₁ is the interlayer spacing of the (111) crystalline plane, and n is the refractive index of the dielectric materials, which is also proportional to the dielectric constant. There are two possible factors that may attribute to the red-shift of the structure colour in the patterned regions. First, the partially dissolved polymers from the poly(St-MMA-AA) spheres in the nonpatterned regions were sucked into the interstice of the colloidal crystals by capillary forces, which may also lead to the minor increase in the dielectric constant of the patterned regions. Second, because of the conformal pressure, the poly(St-MMA-AA) spheres in the contacted regions were deformed laterally, which could extrude the patterned regions into bumps and the distance between the crystalline planes along the vertical direction were enlarged (see ESI Fig. S3a and c†). Overall, the structure colour in the patterned regions was red-shifted from green to orange. On the contrary, colour in the nonpatterned regions turned dark green, which was mainly because of (1) decrease in the lattice distance caused by the conformal pressure and (2) reduced refractive index contrast caused by back-filling of partially dissolved poly(St-MMA-AA). The decrease in lattice distance might offset the increase in the dielectric constant. Therefore, the structure colour might still stay green. Because the poly(St-MMA-AA) spheres were fused together with partially dissolved polymer, the filling ratio of the CPC was increased. As a result, the refractive index contrast was decreased in the system, and the structure colour was not as brilliant as the original colour (see ESI Fig. S1†). If we further decreased the evaporation time to 2 min, the surface profile of the colloidal crystals would disappear (see Fig. 2c and d), and the layers below were merged into a homogenous film with no photonic structures (see ESI Fig. S3b†). Therefore, the structure colour could hardly be seen as shown in the inset of Fig. 2c.

With this method, we can also fabricate colloidal crystal micropatterns with different sizes and geometries. For example, we adopted the PDMS stamp with cubic cavity micropatterns (20 × 20 μm², wall thickness 5 μm) and allowed toluene to evaporate for 3 min. The same moulding procedure was applied, and the photonic structure could be well preserved as shown in Fig. 3a. The cracks in the colloidal crystals were formed during the drying process, which currently cannot be totally avoided.¹² Fig. 3b is the corresponding SEM image of the micropatterned colloidal crystals, and the magnified view (see Fig. 3c–e) of the patterned regions and nonpatterned regions further revealed that the minute amount of toluene could only melt the surface of poly(St-MMA-AA) spheres, which infiltrated into the interstice of the poly(St-MMA-AA) spheres. It is noteworthy that the contacted regions with PDMS were more severely melted than the non-contacted regions (see Fig. 3b), which resulted in the difference in structure colours in the colloidal crystals. The patterned regions of the colloidal crystals showed green-yellow structure colour, which was mainly caused by the increase of the dielectric constant after the minor infiltration of the polymers. It is different from the case (orange structure colour) when the stamp with patterns of circular holes (2 × 2 μm²) was used. This is mainly because (1) a lower amount of poly(St-MMA-AA) was melted and infiltrated into the interstice of the poly(St-MMA-AA) spheres because toluene released from the stamp with micropatterns of cubic cavities (20 × 20 μm²) was even less; (2) there was almost no increase in lattice distance because the area of square patterns was very large, and the pressure could not result in the formation of bumps as the case in the circular holes. On the other hand, the non-patterned regions showed a dark green colour, which is mainly because of the back-filled polymer. It lowered the refractive index contrast of the non-patterned regions and lead to a less brilliant dark green colour. Further, if the solvent was only allowed to evaporate for 1 min, the excess amount of toluene would melt the entire colloidal crystals into a homogeneous film with surface topology replicating the deformed mould (see Fig. S4†).

Fig. 3 The optical image (a) and SEM image (b) of the micromoulded CPCs with square micropatterns (20 × 20 μm²) after 3 min solvent evaporation, and (c) is the magnified view of the patterned region in (b). The white dash lines serve as the guide for the eyes only. (d and e) are the side view of the contacted and non-contacted regions with the PDMS stamp. Scale bars in (a) 50 μm, (b) 25 μm, (c) 5 μm, (d) and (e) 1 μm.

In summary, by utilizing the solvent as a melting agent for the polymers, we demonstrated the easiness, cost-effectiveness and versatility of the modified solvent-assisted imprinting technique for the patterning of CPC films. The concept here is based on the controlled release of a limited amount of etchant from the PDMS stamp to a certain location of the poly(St-MMA-AA) colloidal crystal films for partial dissolution and relocation under the conformal pressure. We adopted the solvent evaporation method with a tightly controlled evaporation time to fine-tune the amount of toluene left for etching. After several trials, we can fix the time window for the patterning of CPCs with distinctive structure colours. This approach showed advantages of simplicity, uniformity and scalability over traditional patterning techniques (template-directed approach) of CPCs. Moreover, the micropatterns obtained in this manner can integrate two different structural colours into a single chip. In addition, it is expected that this concept of microfabrication can be extended for the patterning of metallic films for plasmonic applications.

Experimental

Fabrication of the PDMS moulds

A replication process of micropillar structures on silicon wafer using PDMS was applied to form the elastic stamps. First, the silicon wafers patterned with pillar structures were fabricated by photolithography and deep reactive-ion etching (DRIE) techniques. Second, the PDMS part A and part B (Sylgard 184, Dow Corning) were mixed in a 10 : 1 (w/w) ratio and stirred for 5 minutes. The mixed PDMS was placed in vacuum desiccators for degassing for 15 minutes. The transparent and gas free PDMS solution was poured onto the patterned silicon substrates and cured at 90 °C for 3 hours. The negative PDMS model was obtained by cooling down to room temperature and carefully peeling.

Fabrication of colloidal photonic crystal patterns via solvent-assisted imprinting

The monodispersed poly(St-MMA-AA) colloids were from the same batch used in previous reports.¹³ The colloidal crystal films made of poly(St-MMA-AA) were fabricated with convectively self-assembly method.¹⁴ A volume of 20 μl toluene was dropped on the surface of the patterned PDMS mould to allow it to slowly evaporate (relative humidity 30%; temperature 25 °C). Next, it dried on the surface of the PDMS mould, there was still some residue toluene in the PDMS matrix. An extra drying time (3 minutes in our case) was allowed before pressing the PDMS mould to colloidal crystal films, which are clipped between two parallel glass slides. The pressure was maintained for 5 min to ensure complete evaporation of the toluene and the fixation of the micropatterns on the colloidal crystal films.

Characterization

The patterned micro/nanostructures were characterized with both an optical microscope (Olympus BX51) and a scanning electron microscope (Hitachi, S4300). The sizes of the nanostructures were measured with a NanoMeasure 1.2.

Acknowledgements

The authors thank the 973 Program (2013CB834505) of China for financial support.

Notes and references

1. E. Yablonovitch, Phys. Rev. Lett., 1987, 58, 2059; S. John, Phys. Rev. Lett., 1987, 58, 2486.
2. J. D. Joannopoulos, P. R. Vileneuve and S. Fan, Nature, 1997, 386, 143.
3. T. Ding, Y. Long, K. Zhong, K. Song, G. Yang and C.-H. Tung, J. Mater. Chem. C, 2014, 2, 4100.
4. L. Wang, J. Wang, Y. Huang, M. Liu, M. Kuang, Y. Li, L. Jiang and Y. Song, J. Mater. Chem., 2012, 22, 21405; H. S. Lee, T. S. Shim, H. Hwang, S.-M. Yang and S.-H. Kim, Chem. Mater., 2013, 25, 2684; D. Yang, S. Ye and J. Ge, Adv. Funct. Mater., 2014, 24, 3197.
5. A. Birner, R. B. Wehrspohn, U. Gösele and K. Busch, Adv. Mater., 2001, 13, 377; P. V. Braun, S. A. Rinne and F. García-Santamaría, Adv. Mater., 2006, 18, 2665; K. Aoki, D. Guimard, M. Nishioka, M. Nomura, S. Iwamoto and Y. Arakawa, Nat. Photonics, 2008, 2, 688.
6. J. Henzie, J. E. Barton, C. L. Stender and T. W. Odom, Acc. Chem. Res., 2006, 39, 249; J. Zhang, Y. Li, X. Zhang and B. Yang, Adv. Mater., 2010, 22, 4249.
7. A. van Blaaderen, R. Ruel and P. Wiltzius, Nature, 1997, 385, 321; C. A. Fustin, G. Glasser, H. W. Spiess and U. Jonas, Adv. Mater., 2003, 15, 1025; S. Arpiainen, F. Jonsson, J. R. Dekker, G. Kocher, W. Khunsin, C. M. S. Torres and J. Ahopelto, Adv. Funct. Mater., 2009, 19, 1247; T. Ding, L. Luo, H. Wang, L. Chen, K. Liang, K. Clays, K. Song, G. Yang and C.-H. Tung, J. Mater. Chem., 2011, 21, 11330.
8. Y. Yamauchi, J. Imasu, Y. Kurod, K. Kurod and Y. Sakka, J. Mater. Chem., 2009, 19, 1964.
9. J. M. Yao, X. Yan, G. Lu, K. Zhang, X. Chen, L. Jiang and B. Yang, Adv. Mater., 2004, 16, 81.
10. T. Ding, Q. Zhao, S. Smoukov and J. Baumberg, Adv. Opt. Mater., 2014, 2, 1098; T. Ding, S. Smoukov and J. Baumberg, Nanoscale, 2015 10.1039/c4nr05934d.
11. K. Y. Suh, Y. S. Kim and H. H. Lee, Adv. Mater., 2001, 13, 1386; C. M. Bruinink, M. Péter, P. A. Maury, M. de Boer, L. Kuipers, J. Huskens and D. N. Reinhoudt, Adv. Funct. Mater., 2006, 16, 1555.
12. A. T. Skjeltorp and P. Meakin, Nature, 1988, 335, 424.
13. T. Ding, K. Zhong, Y. Long, K. Song, G. Yang and C.-H. Tung, CrystEngComm, 2014, 16, 7617; T. Ding, K. Song, K. Clays and C.-H. Tung, Langmuir, 2010, 26, 4535.
14. P. Jiang, J. F. Berton, K. S. Hwang and V. L. Colvin, Chem. Mater., 1999, 11, 2132.

---

Footnote

† Electronic supplementary information (ESI) available: Additional optical and SEM images of the colloidal photonic crystals and the patterns. See DOI: 10.1039/c4ra12958j

---