2-Photon Polymerized IP-Dip 3D Photonic Crystals for Mid IR Spectroscopic Applications

Photonic Crystals (PhC) are periodically structured dielectric materials that have been subject to extensive research efforts over the last two decades. PhC are known for the slow-light phenomenon, which increases the interaction time between light and the target gas, thereby enhancing sensitivity when applied to sensors. Slow light is best realized using three-dimensional (3D) PhC with a complete photonic band-gap (PBG). However, they are inherently difficult to fabricate with planar microfabrication techniques. In this work, we present two-photon polymerized (2PP) stereolithographically fabricated 3D PhC targeting mid-infrared (MIR) spectroscopic range. Finite Element Analysis (FEA) was performed on a two-dimensional (2D) PhC waveguide (PhCW) to analyze the significance of PBG and slow light properties. Additional FEA was conducted and validated experimentally using fabricated 3D PhC. The ability to tune PhC to a desired wavelength along with their repeatability and feasibility is experimentally demonstrated. This letter presents the first, to our knowledge, 3D PhC fabricated using 2PP stereolithography operating at this wavelength.

Abstract-Photonic Crystals (PhC) are periodically structured dielectric materials that have been subject to extensive research efforts over the last two decades. PhC are known for the slow-light phenomenon, which increases the interaction time between light and the target gas, thereby enhancing sensitivity when applied to sensors. Slow light is best realized using three-dimensional (3D) PhC with a complete photonic band-gap (PBG). However, they are inherently difficult to fabricate with planar microfabrication techniques. In this work, we present two-photon polymerized (2PP) stereolithographically fabricated 3D PhC targeting midinfrared (MIR) spectroscopic range. Finite Element Analysis (FEA) was performed on a two-dimensional (2D) PhC waveguide (PhCW) to analyze the significance of PBG and slow light properties. Additional FEA was conducted and validated experimentally using fabricated 3D PhC. The ability to tune PhC to a desired wavelength along with their repeatability and feasibility is experimentally demonstrated. This letter presents the first, to our knowledge, 3D PhC fabricated using 2PP stereolithography operating at this wavelength.

I. INTRODUCTION
P HOTONIC crystals (PhC) are metamaterials with periodically structured dielectric layers designed to obtain an energy band-structure [1], [2]. The band-structure either allows or prevents the transmission of a few wavelengths of light through the PhC. The range of frequencies prevented from transmission is called Photonic band gap (PBG). When a defect is engineered in the crystal, a guided mode appears in the PBG which transmits the aligned wavelength of light at a lower group velocity, called the slow light phenomenon, which results in an enhanced light-analyte interaction, enabling high sensitivity at a reduced path lengths [3]. This makes PhC suitable for gas sensing platforms.
Existing gas sensing technologies include cavity ring-down spectroscopy (CRDS) which utilizes an optical detection scheme to conduct fast measurements [4], [5]. Complex architecture and large form factor make CRDS expensive and disproportionate for lab-on-a-chip applications. Metal-oxide (MOx) chemo-resistive sensors are widely reported [6], [7]. Continuous heating is necessary for MOx sensors to initiate the surface chemisorption process resulting in large power consumption. Sensors based on catalytic combustion using a wheatstone bridge circuit are also used for gas detection [8], [9]. The sensor, when exposed to combustible gas, ignites it, and the wheatstone equilibrium is disturbed. This technique is only applicable for combustible gases at lower explosion limits (LEL) as it relies on their combustion, limiting widespread use.
Researchers have reported 2D material-based gas sensors [10], [11]. These sensors can exhibit structural defects and material decay which adversely impact their reliability. Infrared (IR) spectroscopy has also been reported as a method for detecting gas during gathering, transmission and distribution [12], [13].
The absorption bands of hydrocarbon gases overlap creating interference as major disadvantage for IR spectroscopy. Also, long absorption path is required to measure gases at low concentrations, limiting its utility in lab-on-a-chip ubiquitous sensing applications. In general, PhC can be one, two, or three-dimensional corresponding to the symmetry. A complete PBG is possible only in a 3D PhC [14]. However, despite this advantage 3D PhC are not commonly reported in literature due to the inherent fabrication challenges. In particular, it is difficult to fabricate a complex periodic structure with adequate precision. Therefore, successfully stacked structures using traditional fabrication have only been reported up to a few layers, limiting the potential of this technology [15], [16].
Two-photon polymerization (2PP) has previously been used to fabricate PhC to overcome fabrication challenges [17], [18], [19]. However, most of the fabricated PhC are tested for near-infrared (NIR) range with resins like SU-8. The fundamental absorption fingerprints in the mid-infrared (MIR) region have more distinct line intensities than their overtones in the NIR [20] which, combined with specialized resins for optical applications [21], [22] results in improved sensing capabilities.
In this work, we present a direct-laser written 3D PhC operating at MIR spectroscopic range fabricated using two-photon polymerization (2PP) microstereolithography. FEA is performed to model the band-structure of 3D PhC (COMSOL multiphysics 5.6a). The advantage of slow light for micro-scale pathlengths is demonstrated using FEA on 2D PhCW. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ In addition, an experimentally obtained transmission signal for 3D PhC is presented for PhC with two different lattice constants. To our knowledge, this is the first report on 2PP IP-DIP fabricated PhC for MIR spectroscopic gas sensing applications.

II. FINITE ELEMENT ANALYSIS A. 2D PhC Analysis
For pillar-type 2D PhC with square lattice, a PBG is visible only in Transverse-Magnetic (TM) mode as the analysis for 2D PhC is restricted to in-plane propagation [23]. In 3D PhC the restriction is removed due to out-of-plane periodicity and a PBG is visible in both TM and Transverse-Electric (TE) modes. To mitigate the high computational needs for complex 3D PhC FEA, slow-light enhanced sensing is analyzed on 2D PhC.
FEA on 2D dielectric pillar-type PhC is used to investigate the significance of PBG. The dielectric pillar (ϵ = 12) radius was set to 0.3 α, where α is the lattice constant of 800 nm to model the band-structure in air. The band-structure depicts PBG in the normalized frequency region of 0.28-0.42 as highlighted in figure 1. The modeled unit cell is shown in the inset. The normalized frequency (α/λ) is further used to determine the stop band. Further, a wavelength sweep was performed on the PhC. The modeling results for three distinct wavelengths above, in, and below PBG are shown in figure 2. The modeling results demonstrate that when the wavelength is in the PBG, light is inhibited from passing through the PhC, while light passes through the PhC otherwise.
Next, a PhCW targeting methane absorption at λ = 7.625 µm, created by removing a row of pillars, is modeled to enable a guided mode in the PBG. The guided mode permits a specific wavelength in PBG to pass through the PhCW, leading to a bandpass operation. The lattice constant is chosen such that the guided mode aligns with the target wavelength. The background was modeled as air with 700 ppm methane.
Methane's ability to trap heat in the atmosphere makes it a potent greenhouse gas associated with the causes of climate change [24]. In addition, high levels of methane pose a significant explosion hazard to the public leading to growing need of ubiqutous methane sensing. To model the effect of slow light, the average power (P av ) was modelled at the end of the PhCW and gas chamber to calculate absorbance along the arc. The FEA depicts higher light absorption in the PhCW, indicating better light-methane interaction in PhCW due to slow light phenomenon, as also shown in figure 2. The FEA results from 2D PhC can be extended to 3D PhC.

B. 3D PhC Analysis
Because complete PBG is only achievable in 3D PhC, 3D PhC are preferred over 2D PhC for sensing applications. In this work we describe a 3D PhC fabricated using a negative tone UV-curable resin (IP-DIP, Nanoscribe GmbH). Although the refractive index of IP-DIP (1.552) is not sufficient to achieve complete PBG as the dielectric contrast is low, however a partial PBG is achievable for sensing applications in both  TE and TM modes [25]. The band-structure for 2PP stereolithographically printed 3D PhC is modeled and is shown in figure 3. The 3D PhC FEA shows the presence of partial PBG along the Brillouin Zone Boundary -X, which is the preferred boundary for transmission measurements for free space testing. The partial PBG is further used to align structures to a particular wavelength.

III. FABRICATION
Two-photon stereolithography (Photonic Professional GT, Nanoscribe GmbH) was used to fabricate 3D woodpile-PhC tuned to target an absorption peak of 4.0 µm. The 3D PhC were fabricated by exposing IP-DIP drop-casted on a 2-inch silicon wafer (Wacker Chemitronic GmbH). A 780 nm wavelength femtosecond laser was focused on the surface by a high numerical aperture (1.4 NA, 63X magnification, Carl Zeiss). The sample was mounted on a 3-axis piezoelectic scanning stage for precise control of movements. A total of 12 alternating layers were printed with the second nearest  neighbor displaced by half of the lattice constant α. Afterward, samples were immersed in SU-8 developer for 24 hours without stirring, and subsequently rinsed in isopropanol for 30 minutes with continuous stirring, before being blow dried in N2.
For the PhC structure, α is 3.5 µm and the rod length is 144.5 µm with a fill factor of 28 %. The surface characteristics were studied using Scanning Electron Microscopy (SEM) shown in figure 4. A displacement of a/2 is observed for second nearest neighbour. A tilted SEM at the edge shows all 12 layers printed as expected in figure 6A. top view of printed structure confirms homogeneous structures with large-scale periodicity. An array of 6 × 6 PhC was printed to enhance results in free-space testing.
Distinct gases exhibit absorption spectra at different wavelengths [26]. To detect gases using PhC, PBG should be tunable for wavelengths of light to target different gases. To confirm tunability of PBG to desired wavelengths in MIR range, another PhC with α 4.1 µm was also printed and tested to target a different wavelength of light.

IV. EXPERIMENTAL RESULTS
Fourier-transform Infrared Spectrometer (FTIR, Thermofisher Scientific Nicolet IS50) equipped with a Polaris IR broadband light source was used with spectral ranges 400 cm −1 to 4000 cm −1 . All the measurements were performed at 24 • C with constant dry purge air supplied from purge gas generator (Parker Balston). An area equal to the tested array of PhC is defined by introducing a laser cut (Tykma Electrox EMS 300) aperture to avoid excess light. Every spectrum is an average of 32 scans at a resolution of 4 cm −1 . The transmission spectra are normalized to a Silicon wafer as background.
Experimentally obtained transmission spectra for both PhC is shown in figure 6. The expected area for partial PBG corresponding to the FEA is highlighted for both PhC. A tooling factor τ is added accounting for polymer shrinkage and inhomogeneous fill factor in the subsequent layers not visible with SEM. Note that experimentally obtained absorbance due to partial PBG matches FEA prediction. The successful peak shift for the PhC with 4.1 α compared to 3.5 α indicates that the PhC can be tuned to target different wavelengths corresponding to different gases. Polymer shrinkage could be reduced in further experiments by designing an anchor wall around the PhC using the same technique and using Critical Point Drying after isopropanol rinse to reduce stress induced in drying stiction.
Both the PhC were replicated twice on identical substrates. The PhC were tested under FTIR to obtain the respective transmission spectra. Figure 7 shows the experimentally obtained results for both PhC, confirming repeatable structures with overlapping PBGs. However, a slight offset is observed outside the PBG. We assume this offset is due to the alignment of the aperture with the light beam. Further improvement in Fig. 7. Experimentally obtained transmission for repeatability test of 3D PhC. Copies of both PhC crystals were printed and tested under identical experimental conditions. the aperture alignment or using different techniques can limit the alignment errors. The total time to print twelve layers to constitute one PhC was 6 minutes, using a custom optimized laser path algorithm. The 6 × 6 array was printed in roughly 4 hours, proving feasibility for rapid fabrication. A single 12−layered PhC can be used standalone for applications involving optical fibres to limit the requirement for a large surface area, further speeding up the fabrication process.

V. CONCLUSION
In this work we demonstrated the experimental fabrication of 2PP micro-stereolithographically printed tunable 3D PhC using IP-DIP for applications in MIR spectroscopic sensing. FEA of a 2D PhC was used to illustrate the significance of slow light for gas sensing purposes. Bandstructures were calculated and plotted to show partial PBG in 3D PhC fabricated using IP-DIP. Transmission spectra were experimentally obtained and compared to the FEA model, with good agreement. The fabrication time of under 10 minutes and enhanced sensing make this a feasible technology for gas sensing. The future work includes defect engineering and refractive index increment by direct dielectric infiltration and coating of IP-DIP 3D PhC which will enable complete PBGs in the MIR range leading to calculation of group velocity, group velocity dispersion and normalized delay bandwidth product for 3D PhCW. This will result in a tuneable, rapid, and cheap fabrication of 3D PhC for on-chip sensing applications.

ACKNOWLEDGMENT
The fabrication of PhC was conducted at the Nanotechnology Core Facility (NCF) at the University of Illinois Chicago.