1. Introduction

The deep ultraviolet is an important wavelength range for a variety of astronomy, cosmology, planetary studies, communications, and biological applications [1–4]. Recent years have seen significant improvements in UV detectors [5, 6], with an increasing interest in the use of these components for the development of UV optical systems. For imaging and spectroscopy in particular, the development of wavelength-selective filters that can be directly integrated on silicon detectors and operate in the 200–400 nm range is essential [7].

In previous work, UV bandpass filters have been demonstrated using metal-dielectric, thin film stacks on silicon [7, 8]. The center wavelength of the filter was varied over the deep UV range, with a passband width on the order of 100 nm. However, this approach does not straightforwardly allow for fabrication of multiple filters on the same substrate: different center wavelengths would require different sets of film thicknesses to be deposited in the stack.

Two-dimensional (2D) nanohole arrays in metal films provide an alternative approach to filter design [9], in which wavelength response is provided by excitation of surface plasmon modes. Since metallic nanohole arrays can be made by standard lithographic techniques, multiple filter designs can straightforwardly be patterned on the same substrate. Nanohole arrays further offer the advantages of polarization independence, relatively narrow linewidth, and good out-of-band rejection. However, most work to date has focused on visible and infrared filters [10–14]. In the UV, the most common plasmonic materials, gold and silver, are subject to losses and/or material degradation [15–18]. Aluminum (Al) has been suggested as an improved candidate for UV plasmonics, one which benefits from chemical and thermal stability, low cost, and CMOS compatibility [18]. Al nanohole arrays are thus attractive candidates for UV filter design. Yet, while one paper has demonstrated UV filters in Al nanohole arrays [19], these filters were designed to use the waveguide cut-off of the holes, rather than the surface plasmon modes. As a result, varying the nanogrid dimensions primarily shifted the upper edge of a broad passband. For UV spectroscopic and imaging applications, it is instead desirable to select a narrow band within the UV, with the capability to tune the filter design to shift the center wavelength.

In this paper, we design, fabricate, and characterize narrow-linewidth, bandpass filters using the surface plasmon modes of Al nanohole arrays on silicon and show that the center wavelength can be tuned across the UV. We use the finite-difference time-domain method (FDTD) to design filters with a single, dominant reflection dip and a linewidth of approximately 30 nm. We investigate the effect of non-normal incidence, showing that it results in mode splitting. Moreover, the higher-wavelength of the split modes narrows, with a linewidth down to 10 nm. We then fabricate and characterize the designs experimentally, demonstrating good agreement between design and measurement. We experimentally demonstrate tuning of the filter response across the 200–400 nm range by varying the lattice constant and/or angle of incidence. These results should further enable the design of UV spectroscopic and imaging systems for varied application fields.

2. Numerical modeling

Figure 1(a) shows a schematic of an Al thin film of thickness h = 40 nm, deposited on a Si substrate and patterned with a periodic array of holes of diameter 2r = 0.5a and a period a. We use Lumerical FDTD simulations to calculate the reflection and transmission spectra, as shown in Figs. 1(b) and 1(c). For each lattice constant, a single, dominant reflection dip is obtained. As the lattice constant a is swept from 150 to 250 nm, the wavelength of the dip shifts from 210 to 280 nm. These dips correspond to transmission peaks, as illustrated in Fig. 1(c). The transmission value increases from approximately 10% to 20% with increasing lattice constant. Note that the transmission is calculated at the Al-Si interface and so represents the fraction of light that enters the Si substrate. The normal incidence transmission into a bare silicon substrate was also numerically calculated, and it ranges between 24 and 40% for the wavelength range of 150–300 nm. Thus, the Al nanohole array provides high wavelength selectivity at the cost of reducing the carriers generated due to direct absorption in silicon. This total throughput is competitive with commercial stand-alone filter options in this wavelength range which rely on metal-dielectric multilayers to achieve transmissions on the order of 10–30%. Without direct integration, additional reflection losses at the surface of a silicon sensor platform would further reduce overall throughput. High-efficiency, all-dielectric filter structures are not typically available in the deep UV due to absorption losses in the most common optical oxide materials.

 

Fig. 1 (a) Schematic of a periodic array of Al holes on a Si substrate, (b) and (c) are the simulated reflection and transmission spectra at normal incidence for lattice constant a = 150, 200 and 250 nm.

Download Full Size | PDF

The dominant peak in Fig. 1 arises from the fundamental surface plasmon at the Al-air interface. In Fig. 2(a), we replot the reflection spectrum for a = 250 nm and h = 40 nm (red line). The inset shows a side view of the nanohole array. Figure 2(b) shows the squared intensity of the electric field, which is concentrated at the upper edge of the Al holes. At normal incidence, the surface plasmon resonances in a periodically patterned metallic film are governed by [9]:

 

Fig. 2 (a) Reflection spectra of patterned Al thin film on a Si and a glass substrate, with a = 250 nm and h = 40 nm. Inset is a side view of the nanohole array. (b–e) Respective mode profiles for reflection dips in (a).

Download Full Size | PDF

In contrast, the same device fabricated on a commonly-used glass substrate [10, 12] would have multiple reflection dips within the wavelength range of interest. The filter response for the Al nanohole array on glass is shown by the blue line in Fig. 2. In addition to the surface plasmon at the Al-air interface (λ_g3), additional dips appear at higher wavelengths (λ_g2, λ_g1). These dips correspond to the first and second order surface plasmon resonances at the Al-glass interface. The field profiles are shown in Figs. 2(c) –2(e). For the Si substrate, the higher permittivity of Si shifts these modes to higher wavelengths, above the spectral region shown.

The filter response varies with the angle of incident light. Figure 3 shows the reflection for the Al on silicon structure. The fundamental mode for normal incidence splits into two modes for 35° incidence angle. The higher-wavelength mode (λ₂) for 35° incidence is about three times narrower than the normal-incidence peak (λ₁). The angle sensitivity of the filter is a fundamental consequence of the periodicity. For light incident at θ with respect to the surface normal, we can derive (see Appendix) that:

 

Fig. 3 (a) Reflection spectra of patterned Al thin film on Si for normal (0°) and non-normal incidence (35°), with a = 250 nm and h = 40 nm. (b–d) respective mode profiles for reflection dips in (a).

Download Full Size | PDF

Using the results above, we selected several designs for fabrication on Si substrates and characterized the optical response in reflection. The eventual integration of such structures with an active detector platform is required to accurately estimate transmission, but comparison of the measured reflectance with calculations can provide confidence in the fidelity of the designed spectral response. Such an approach has been shown to be an accurate predictor of transmission in 1D metal-dielectric structures integrated directly onto Si photodiodes [7].

3. Experimental results

Devices were fabricated using thin film deposition, e-beam lithography, and etching to form hole arrays within a 200x200 μm square field. The Al film was deposited by electron beam evaporation onto a Si substrate to a target thickness of 40 nm as monitored by quartz crystal microbalance. The system base pressure was less than 5x10⁻⁸ Torr, and film was evaporated at a rate of 8 nm/s. A high evaporation rate was used to ensure good optical quality of the evaporated layer, particularly in the deep UV. The Al nanohole array geometry was defined using ebeam lithography with a ZEP520A resist. After development, the Al was patterned via reactive ion etching in a plasma of BCl₃/Cl₂/Ar. Samples were immediately submersed in water following the etch to reduce the corrosive action of residual chlorine byproducts. Figures 4(a) and 4(b) show scanning electron microscope (SEM) images of the patterned Al film on a Si substrate.

 

Fig. 4 (a,b) SEM pictures of fabricated devices with a = 200 nm and a = 250 nm. (c,d) Measured and corresponding FDTD simulations of reflection spectra of Al nano-holes array on silicon, with (c) a = 200 nm and (d) a = 250 nm, θ = 35° and h = 40 nm.

Download Full Size | PDF

The spectral response of the Al nanohole arrays were measured as a function of increasing lattice constant using a fiber-fed, phase-modulated spectroscopic ellipsometer operating with a xenon lamp. Measurements of unpolarized (average) reflectance were made using an in-line mask to slightly under fill the patterned field. Figures 4(c) and 4(d) show good agreement between the simulations and the experimental results with a measured 10 nm linewidth of the rightmost peaks. In this case measurements were taken at an incident angle of 35° which is the minimum value enforced by the mechanical limits of the proximity between the input and output fiber paths.

The filter response can be tuned by varying the lattice constant, as shown in Fig. 5. The experimental measurements show that the narrow reflection dip can be shifted between 250 and 500 nm by varying the lattice constant a between 150 nm and 250 nm. As the lattice constant increases, higher order modes start to appear at lower wavelengths. In a final implementation these modes could be rejected by additional longpass filtering if a single dominant bandpass is desired.

 

Fig. 5 Measured reflection spectra of Al on silicon, with a = 100, 200, 250, 300 nm, θ = 35° and h = 40 nm.

Download Full Size | PDF

The filter response can also be tuned by varying both lattice constant and angle of incidence. In Fig. 6, we use two lattice constants and three angles of incidence to span the 250–400 nm range. In the Appendix, we compare the wavelength of the measured reflection dips with theory and obtain good agreement. This result can be exploited in a spectroscopic system with an additional mechanism to rotate the detector with respect to the incident light path.

 

Fig. 6 Measured reflection spectra of Al on silicon, with a = 150, 200 nm, θ = 35°, 45°, 55° and h = 40 nm.

Download Full Size | PDF

4. Conclusions

In conclusion, we studied the surface plasmonic resonances in nanohole arrays of Al on Si at deep UV wavelengths. The filter wavelength can be tuned by adjusting the lattice constant of the array, providing a straightforward route to integration of multiple filters on a single lithographically-defined chip. In particular, we have shown that UV filters can be designed throughout the UV wavelength range, from 200 to 400 nm, with a linewidth below 30 nm. At off-normal incidence, the intrinsic angle sensitivity of the filter splits the fundamental plasmonic mode and narrows the higher-wavelength dip, which can readily be tuned over the wavelength range of interest. These performance demonstrations pave the way to develop UV devices and applications for spectral filtering, UV optoelectronics, and structural color. Because of the high energy of the photons in the wavelength range of 200–400 nm, the hot-electron induced charge injection into the substrate could enhance the photocurrent [20], which makes an interesting subject for a future study that implements the Al filter on a CCD photodetector.

5 Appendix A

For light incident at θ with respect to the surface normal, the wavevector at the metal/ dielectric interface is represented by:

(3)$$k_x=\frac{2\pi }{\lambda }\sin \theta \widehat{x}$$

Due to periodicity in the x and y directions, the surface plasmon wavevector can be written as:

(4)$$k_{sp}=\frac{2\pi }{\lambda }\sin \theta \widehat{x}\pm \frac{2\pi }{a}i\widehat{x}\pm \frac{2\pi }{a}j\widehat{y}$$

where,

(5)$$\left|k_{sp}\right|=\frac{2\pi }{\lambda }\sqrt{\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}}$$

Equating the magnitude of both sides of Eq. (4) and solving for λ, we get:

(6)$${\left(\frac{2\pi }{\lambda }\right)}^2\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}={\left(\frac{2\pi }{\lambda }\sin \theta \pm \frac{2\pi i}{a}\right)}^2+{\left(\frac{2\pi j}{a}\right)}^2$$

(7)$$\frac{1}{{\lambda }^2}\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}=\frac{{\sin }^2\theta }{{\lambda }^2}+\frac{i^2}{a^2}\pm \frac{2i\sin \theta }{a\lambda }+\frac{j^2}{a^2}$$

(8)$$\frac{\left(i^2+j^2\right)}{a^2}{\lambda }^2\pm \frac{2i\sin \theta }{a}\lambda +\left({\sin }^2\theta -\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}\right)=0$$

(9)$$\lambda =\frac{\pm \frac{2i\sin \theta }{a}\pm \sqrt{\frac{4i^2{\sin }^2\theta }{a^2}-\frac{4\left(i^2+j^2\right)}{a^2}\left({\sin }^2\theta -\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}\right)}}{2\frac{\left(i^2+j^2\right)}{a^2}}$$

Equation (9) can then be reduced to:

(10)$$\lambda =\pm \frac{ia\sin \theta }{\left(i^2+j^2\right)}\pm a\sqrt{\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}\frac{1}{\left(i^2+j^2\right)}-\frac{j^2{\sin }^2\theta }{{\left(i^2+j^2\right)}^2}}$$

For the fundamental mode (i = 1, j = 0), Eq. (10) reduces to:

(11)$$\lambda =a\left(\sqrt{\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}}+\sin \theta \right)$$

We can compare the analytical prediction from Eq. (11) to the experimental results in the main text, above. Since ε_d = 1 and |ε_m| > 10 for the Al-air interface mode, the square root term in Eq. (11) is approximately equal to 1. This reduces the maximum wavelength Eq. (11) to:

(12)$$\lambda \approx a\left(1+\sin \theta \right)$$

The experimental results are compared to Eq. (11) in Fig. 7, and show good agreement.

 

Fig. 7 Measured (scattered points) and calculated (solid lines) effect of angle of incidence on the maximum wavelength of surface plasmon mode (SPR) at the Al air interface. Measure samples are Al on silicon, with a = 150, 200, 250, 300 nm, θ = 35°, 45°, 55° and h = 40 nm.

Download Full Size | PDF

Funding

Jet Propulsion Laboratory, California Institute of Technology (National Aeronautics and Space Administration contract); Jet Propulsion Laboratory (SURP award).

Acknowledgments

Computational resources were provided by the University of Southern California Center for High-Performance Computing and Communications (www.usc.edu/hpcc). AM and MLP were funded in part by a SURP award.

References

1. S. L. Olson, E. W. Schwieterman, C. T. Reinhard, A. Ridgwell, S. R. Kane, V. S. Meadows, and T. W. Lyons, “Atmospheric seasonality as an exoplanet biosignature,” Astrophys. J. 858(2), L14 (2018). [CrossRef]  

2. P. A. Scowen, T. Tripp, M. Beasley, D. Ardila, B.-G. Andersson, J. M. Apellániz, M. Barstow, L. Bianchi, D. Calzetti, M. Clampin, C. J. Evans, K. France, M. G. García, A. G. de Castro, W. Harris, P. Hartigan, J. C. Howk, J. Hutchings, J. Larruquert, C. F. Lillie, G. Matthews, S. McCandliss, R. Polidan, M. R. Perez, M. Rafelski, I. U. Roederer, H. Sana, W. T. Sanders, D. Schiminovich, H. Thronson, J. Tumlinson, J. Vallerga, and A. Wofford, “Finding the UV–visible path forward: proceedings of the community workshop to plan the future of UV/visible space astrophysics,” Publ. Astron. Soc. Pac. 129(977), 076001 (2017). [CrossRef]  

3. Z. Xu and B. M. Sadler, “Ultraviolet communications: potential and state-of-the-art,” IEEE Commun. Mag. 46(5), 67–73 (2008). [CrossRef]  

4. B. Lin, S. Urayama, R. M. G. Saroufeem, D. L. Matthews, and S. G. Demos, “Characterizing the origin of autofluorescence in human esophageal epithelium under ultraviolet excitation,” Opt. Express 18(20), 21074–21082 (2010). [CrossRef]   [PubMed]  

5. G. A. Shaw, A. M. Siegel, J. Model, A. Geboff, S. Soloviev, A. Vert, and P. Sandvik, “Deep UV photon-counting detectors and applications,” Proc. SPIE 7320, 73200J (2009). [CrossRef]  

6. S. Nikzad, M. Hoenk, A. D. Jewell, J. J. Hennessy, A. G. Carver, T. J. Jones, T. M. Goodsall, E. T. Hamden, P. Suvarna, J. Bulmer, F. Shahedipour-Sandvik, E. Charbon, P. Padmanabhan, B. Hancock, and L. D. Bell, “Single photon counting UV solar-blind detectors using silicon and III-nitride materials,” Sensors (Basel) 16(6), 927 (2016). [CrossRef]   [PubMed]  

7. J. Hennessy, A. D. Jewell, M. E. Hoenk, and S. Nikzad, “Metal-dielectric filters for solar-blind silicon ultraviolet detectors,” Appl. Opt. 54(11), 3507–3512 (2015). [CrossRef]   [PubMed]  

8. J. Mu, P.-T. Lin, L. Zhang, J. Michel, L. C. Kimerling, F. Jaworski, and A. Agarwal, “Design and fabrication of a high transmissivity metal-dielectric ultraviolet band-pass filter,” Appl. Phys. Lett. 102(21), 213105 (2013). [CrossRef]  

9. H. F. Ghaemi, T. Thio, D. E. Grupp, T. W. Ebbesen, and H. J. Lezec, “Surface plasmons enhance optical transmission through subwavelength holes,” Phys. Rev. B 58(11), 6779–6782 (1998). [CrossRef]  

10. Q. Chen and D. R. S. Cumming, “High transmission and low color cross-talk plasmonic color filters using triangular-lattice hole arrays in aluminum films,” Opt. Express 18(13), 14056–14062 (2010). [CrossRef]   [PubMed]  

11. S. Yokogawa, S. P. Burgos, and H. A. Atwater, “Plasmonic color filters for CMOS image sensor applications,” Nano Lett. 12(8), 4349–4354 (2012). [CrossRef]   [PubMed]  

12. V. R. Shrestha, S.-S. Lee, E.-S. Kim, and D.-Y. Choi, “Aluminum plasmonics based highly transmissive polarization-independent subtractive color filters exploiting a nanopatch array,” Nano Lett. 14(11), 6672–6678 (2014). [CrossRef]   [PubMed]  

13. Y.-W. Huang, W. T. Chen, W.-Y. Tsai, P. C. Wu, C.-M. Wang, G. Sun, and D. P. Tsai, “Aluminum plasmonic multicolor meta-hologram,” Nano Lett. 15(5), 3122–3127 (2015). [CrossRef]   [PubMed]  

14. I. J. H. McCrindle, J. P. Grant, L. C. P. Gouveia, and D. R. S. Cumming, “Infrared plasmonic filters integrated with an optical and terahertz multi-spectral material,” Phys. Status Solidi., A Appl. Mater. Sci. 212(8), 1625–1633 (2015). [CrossRef]  

15. G. Maidecchi, G. Gonella, R. Proietti Zaccaria, R. Moroni, L. Anghinolfi, A. Giglia, S. Nannarone, L. Mattera, H.-L. Dai, M. Canepa, and F. Bisio, “Deep ultraviolet plasmon resonance in aluminum nanoparticle arrays,” ACS Nano 7(7), 5834–5841 (2013). [CrossRef]   [PubMed]  

16. A. Ono, M. Kikawada, R. Akimoto, W. Inami, and Y. Kawata, “Fluorescence enhancement with deep-ultraviolet surface plasmon excitation,” Opt. Express 21(15), 17447–17453 (2013). [CrossRef]   [PubMed]  

17. T. F. Villesen, C. Uhrenfeldt, B. Johansen, and A. Nylandsted Larsen, “Self-assembled Al nanoparticles on Si and fused silica, and their application for Si solar cells,” Nanotechnology 24(27), 275606 (2013). [CrossRef]   [PubMed]  

18. M. W. Knight, N. S. King, L. Liu, H. O. Everitt, P. Nordlander, and N. J. Halas, “Aluminum for plasmonics,” ACS Nano 8(1), 834–840 (2014). [CrossRef]   [PubMed]  

19. W.-D. Li and S. Y. Chou, “Solar-blind deep-UV band-pass filter (250 - 350 nm) consisting of a metal nano-grid fabricated by nanoimprint lithography,” Opt. Express 18(2), 931–937 (2010). [CrossRef]   [PubMed]  

20. T. Gong and J. N. Munday, “Aluminum-based hot carrier plasmonics,” Appl. Phys. Lett. 110(2), 021117 (2017). [CrossRef]