Miniature Spectroscopes with Two-Dimensional Guided-Mode ... - MDPI

materials Article

Miniature Spectroscopes with Two-Dimensional Guided-Mode Resonant Metal Grating Filters Integrated on a Photodiode Array Yoshiaki Kanamori *, Daisuke Ema and Kazuhiro Hane Department of Finemechanics, Tohoku University, Miyagi 980-8579, Japan; [email protected] (D.E.); [email protected] (K.H.) * Correspondence: [email protected]; Tel.: +81-22-795-6965 Received: 18 September 2018; Accepted: 5 October 2018; Published: 10 October 2018







Abstract: A small spectroscope with 25 color sensors was fabricated by combining metamaterial color filters and Si photodiodes. The metamaterial color filters consisted of guided-mode resonant metal gratings with subwavelength two-dimensional periodic structures. Transmittance characteristics of the color filters were designed to obtain peak wavelengths proportional to grating periods. For each color sensor, a peak wavelength of the spectral sensitivity could be tuned in the range of visible wavelengths by adjusting each grating period. By performing spectrum reconstruction using Tikhonov regularization, the spectrum of an incident light was obtained from the signal of photodiodes. Several monochromatic lights were made incident on the fabricated device and the spectral characteristics of the incident light were reconstructed from the output signals obtained from the respective color sensors. The peak wavelengths of the reconstructed spectra were in good agreement with the center wavelengths of the monochromatic lights. Keywords: spectroscopes; metamaterial; plasmonics; structural color filters; photodiodes

1. Introduction Color filters, which work as wavelength selective filters, have been used in image sensors and liquid crystal displays. In recent years, since the discovery of extraordinary transmission phenomenon based on surface plasmon by Ebbesen et al. in 1998 [1], plasmonic color filters using metal nanostructures have been actively studied [2–6]. Plasmonic color filters have various advantages over the conventional color filters using pigment, such that various color characteristics can be realized depending on the structural shapes with thicknesses of just tens of nanometers. Therefore, plasmonic color filters for many colors can be fabricated on the same substrate by a single fabrication process, unlike the conventional color filters using pigment. Also, plasmonic color filters have high compatibility with complementary metal–oxide–semiconductor and charge-coupled devices based on semiconductor microfabrication technologies, compared with color filters using conventional pigments. Furthermore, in recent years, a nanoimprint technology in which submicron structures are easily formed using molds has been advanced [7–18], and improved productivity of plasmonic nanostructures can be expected. Spectroscopes are widely used [19] as instruments for measuring the energy intensity for each wavelength of light. Since most spectroscopes use diffraction gratings as wavelength selective elements, a certain propagation distance is required to separate diffracted waves spatially. Therefore, it is difficult to miniaturize the spectroscopes that use diffraction gratings. Moreover, precise optical axis adjustment is required for assembling the numerous optical components in the construction of the spectroscopes, which results in high cost. In industrial fields, colorimeters using spectroscopes are used as devices

Materials 2018, 11, 1924; doi:10.3390/ma11101924

www.mdpi.com/journal/materials

Materials 2018, 11, 1924

2 of 11

Materials 2018, 11, x FOR PEER REVIEW

2 of 11

for measuring color. To meet the increasing demand for spectroscopic devices, such as the color such as the color management products andcost food qualityand control, cost reduction and the management of products and foodofquality control, reduction the downsizing of spectroscopic downsizing spectroscopic devices are strongly required. devices are of strongly required. InInrecent novel spectroscopes fabricated by combining plasmonic colorcolor filter filter array array and recentyears, years, novel spectroscopes fabricated by combining plasmonic photodiode array array have been [20–26]. A color formed on each and and photodiode have reported been reported [20–26]. A filter color isfilter is formed on photodiode, each photodiode, spectral information is calculated using using outputoutput signalssignals obtained from photodiodes. Such filter and spectral information is calculated obtained from photodiodes. Sucharray filter spectroscopes do not do require long propagation distances like the ones diffraction gratings and array spectroscopes not require long propagation distances like theusing ones using diffraction gratings are to have the possibility of expanding the the range of of useuse ofofspectroscopic andthought are thought to have the possibility of expanding range spectroscopicdevices devicesasas ultrasmall ultrasmalland andinexpensive inexpensivespectroscopes. spectroscopes.However, However,the thewavelength wavelengthselectivity selectivityofofcolor colorfilters filtersused used for forthe thefilter filterarray arrayspectroscopes spectroscopesreported reportedare areinsufficient. insufficient.Instead, Instead,increasing increasingthe thenumber numberofoffilters filters and andimplementation implementationofofan animproved improvedcalculation calculationalgorithm algorithmhave havebeen beenattempted attemptedtotoimprove improvethe the spectroscopic [22]. Increasing thethe number of filters leads to an in spectroscopiccharacteristics characteristicsofofdevices devices [22]. Increasing number of filters leads to increase an increase the receiving area,area, contrary to miniaturization. Also, complicated calculation processing leads in light the light receiving contrary to miniaturization. Also, complicated calculation processing toleads the to load of calculation processing. Therefore, to improve filter theand loaddelay and delay of calculation processing. Therefore, to improve filterarray arrayspectroscope spectroscope devices, improvementofofthethe filter characteristics is required. Although guided-mode devices, the the improvement filter characteristics itselfitself is required. Although guided-mode resonant resonant gratings are known as high-efficiency wavelength selective filters [27–36], they function as gratings are known as high-efficiency wavelength selective filters [27–36], they function as reflection reflection type wavelength selective filters. Transmission type wavelength selective filters are type wavelength selective filters. Transmission type wavelength selective filters are necessary for necessary forin application filter array spectroscopes. application filter arrayinspectroscopes. As Asmentioned mentionedabove, above,various variousplasmonic plasmoniccolor colorfilters filtershave havebeen beenreported reportedsosofar. far.However, However,asasfar far asaswe weknow, know,there thereare areonly onlyfew fewreports reportson onfilter filterarray arrayspectroscopes spectroscopesintegrating integratingcolor colorfilters filtersand and photodiodes. onthe theimprovement improvement their characteristics is required. In study, this photodiodes.Therefore, Therefore, research on of of their characteristics is required. In this study, we design and fabricate a filterspectroscope array spectroscope and evaluate the characteristics of the we design and fabricate a filter array and evaluate the characteristics of the fabricated fabricated spectroscopic device integrating newly designed colorand filters and photodiode spectroscopic device integrating newly designed color filters photodiode array. array. There There are no are no reports of filter array spectroscopes combining guided-mode resonant gratings and reports of filter array spectroscopes combining guided-mode resonant gratings and photodiodes, photodiodes, as far We as we know. Wedesigned have newly designedtype transmission type wavelength as far as we know. have newly transmission wavelength selective filtersselective based on filters based onresonant guided-mode resonantwith metal gratings withtwo-dimensional subwavelength (2D) two-dimensional (2D) guided-mode metal gratings subwavelength periodic structures periodic without polarization dependency normal incident light. withoutstructures polarization dependency for normal incidentfor light. DeviceConfiguration Configuration 2.2.Device Figure1a1ashows shows a conceptual diagram of proposed the proposed spectroscopic device. A plurality Figure a conceptual diagram of the spectroscopic device. A plurality of colorof color sensors with different sensitivity characteristics, which metamaterial color filters sensors with different spectralspectral sensitivity characteristics, in whichinmetamaterial color filters having having different spectral characteristics areand stacked and on arranged on each photodiode, are Incident arranged. different spectral characteristics are stacked arranged each photodiode, are arranged. Incident wavelengths the color enter the photodiodes, the output signals of the wavelengths selected byselected the colorby filters enter filters the photodiodes, the output signals of the photodiodes photodiodes are read sequentially read out, and thecharacteristics spectral characteristics of the incident lightobtained are obtained are sequentially out, and the spectral of the incident light are by by calculation processing. Figure 1b shows the cross-sectional view of a color sensor consisting of calculation processing. Figure 1b shows the cross-sectional view of a color sensor consisting of aa metamaterialcolor colorfilter filterformed formedon onan anSiSiphotodiode photodiodethrough throughaaspacer spacerlayer. layer.The Themetamaterial metamaterialcolor color metamaterial filterconsists consistsofofaaguided-mode guided-moderesonant resonantmetal metalgrating gratingcovered coveredwith withaaSiO SiO2 2layer. layer. filter Incident light Cover Metal grating layer Waveguide layer Spacer Photodiode

Bias voltage Photo current

n-Si Al

(a)

n+-Si SiO2

p-Si HfO2

(b)

Figure1.1.Schematics Schematics proposed devices. Perspective view in the × 2 color sensors; Figure of of proposed devices. (a) (a) Perspective view in the casecase of 2of × 22 color sensors; (b) (b) cross-sectional view of one particular color sensor. cross-sectional view of one particular color sensor.

Materials 2018, 11, 1924 Materials 2018, 11,11, x FOR PEER Materials 2018, x FOR PEERREVIEW REVIEW

3 of 11 33ofof1111

grating layer layer formed formedon onan anHfO HfO222 Theguided-mode guided-moderesonant resonantmetal metalgrating grating consists consists of of aa 2D 2D Al Al grating The layer formed on an HfO guided layer. Theguided guidedlayer layerisisformed formedon onthe theSiO SiO222 spacer spacer layer. Since layer. the effective effectiverefractive refractiveindex index guided layer. The Since the theguided guidedlayer layerisishigher higherthan thanthat thatof ofthe thesurroundings, surroundings, the guided ofofthe surroundings, guided layer layer functions functionsas asaaplanar planar waveguideand andstrongly stronglyconfines confinesthe theresonant resonantwavelength. wavelength. Transmitted Transmitted light waveguide light generates generatesonly onlyzeroth zeroth order diffractionbecause becausethe thegrating gratingperiod periodis smaller than than the resonant wavelength. grating period isissmaller smaller than the resonant resonant wavelength. wavelength. The order diffraction Thegrating grating groove isfilled filled with SiO Incident light impinged from SiOlayer. 22 cover layer. groove with 2.2.Incident isis impinged cover layer. Photocurrent Photocurrent groove isisfilled with SiOSiO lightlight is impinged from the SiO2the cover Photocurrent generated 2 . Incident generated in the photodiode is read out as an output signal. By changing the period of metal generated in the photodiode is an read out as an output signal.the Byperiod changing themetal period ofthe thelayer metal in the photodiode is read out as output signal. By changing of the grating of grating layer of each color filter, it is possible to control the resonant wavelength and to extract grating layer of each color filter, it is possible to control the resonant wavelength and to extract each color filter, it is possible to control the resonant wavelength and to extract spectral information spectral information fromeach eachphotodiode. photodiode. spectral information from from each photodiode. Designand andNumerical NumericalAnalysis Analysis 3. 3. Design 3.1. Designof ColorFilters Filters 3.1. Design ofofColor Color Filters 3.1. Design Numerical analysis was carried out using rigorous coupled-wave analysis (RCWA), Numerical outout using rigorous coupled-wave analysis (RCWA), whichwhich yields Numerical analysis analysiswas wascarried carried using rigorous coupled-wave analysis (RCWA), which yields accurate results using Maxwell’s equations in the frequency domain [37,38]. Figure 2 shows a accurate resultsresults using using Maxwell’s equations in the frequency domain [37,38]. yields accurate Maxwell’s equations in the frequency domain [37,38].Figure Figure22shows shows aa calculation modelofofone one colorfilter filter element. The The electric electric field, field, magnetic field, and propagation calculation calculation model model of one color color filter element. element. The electric field, magnetic magnetic field, field, and and propagation propagation direction of incident light are parallel to the x, y, and z axes, respectively. A 50-nm-thick SiO2 cover direction of incident incidentlight lightare areparallel paralleltoto x, and y, and z axes, respectively. A 50-nm-thick SiO2 direction of thethe x, y, z axes, respectively. A 50-nm-thick SiO2 cover layer, 30-nm-thick 2D-periodic Al nanodot array, 100-nm-thick HfO 2 waveguide layer, and a 150-nmcover layer, 30-nm-thick 2D-periodic Al nanodot array, 100-nm-thick HfO waveguide layer, and a 2 layer, 30-nm-thick 2D-periodic Al nanodot array, 100-nm-thick HfO2 waveguide layer, and a 150-nmthick SiO2 spacer are formed on an Sionsubstrate of photodiode. Here, Λ Here, and a Λ are thea grating period 150-nm-thick SiO spacer are formed an Si substrate of photodiode. and are the grating thick SiO2 spacer 2are formed on an Si substrate of photodiode. Here, Λ and a are the grating period and nanodot size, size, respectively. The color filters have polarization independent because period and nanodot respectively. filters have polarization independent becauseofof of2D 2D and nanodot size, respectively. The The colorcolor filters have polarization independent because 2D subwavelength gratings with the same periods for the x and y directions. subwavelength gratings with the same periods for the x and y directions. subwavelength gratings with the same periods for the x and y directions.

y z

y

x

Λ Λ a

x

50 nm 30nm nm 50 100nm nm 30 150 nm nm 100 150 nm

a

z

Al Al HfO2 HfO2

(a) (a)

SiO2 SiO2 Si Si

Λ a Λ a

a a

Λ Λ

(b) (b)

Figure 2. Calculation model of one color filter element. (a) Cross-sectional view; (b) top view. Figure 2. Calculation model of one color filter element. (a) Cross-sectional view; (b) top view. view.

Normally, incident light from outside was made to pass through the cover layer, and Normally, light outside wasthe made through the cover layer, transmittance Normally,incident incident light outside was to pass through the toand cover layer, and transmittance spectrum at from thefrom Si side of SiOto 2made /Sipass interface was calculated obtain spectral spectrum at the Si side of the SiO /Si interface was calculated to obtain spectral characteristics ofand the 2optical transmittance at the side designing of the SiOof2/Si calculated to color obtainfilter spectral characteristicsspectrum of the device. ForSi theinterface structure was integrated with the device. For optical designing of the structure integrated with the color filter and photodiode, spectral photodiode, spectral characteristics, including theofinfluence of reflection andwith interference theand Si characteristics of the device. For optical designing the structure integrated the colorby filter characteristics, including the influence of reflection and interference by the Si substrate, were calculated substrate, were calculated instead ofincluding just obtaining the transmission spectrum only the color photodiode, spectral characteristics, the influence of reflection andofinterference by filter. the Si instead of just theinstead transmission of only the color filter. Thethe incident light polarized The incident light polarized along the spectrum x-axis was normally incident and transmittance substrate, wereobtaining calculated of just obtaining the transmission spectrum of total only the color filter. along the x-axis was normally incident and the total transmittance obtained by combining all the obtained by combining all the transmitted diffracted waves was calculated. Refractive indices of Al, The incident light polarized along the x-axis was normally incident and the total transmittance transmitted diffracted waves was calculated. Refractive indices of Al, Si, and SiO in references [39–41], 2 Si, and SiO 2 in references [39–41], respectively, were used for the calculation. Figure 3 shows real and obtained by combining all the transmitted diffracted waves was calculated. Refractive indices of Al, respectively, were used the calculation. 3 shows real and imaginary parts the refractive parts of thefor refractive index of Figure HfOwere 2 used for for the calculation, which was3of measured a Si,imaginary and SiO2 in references [39–41], respectively, used the calculation. Figure shows realbyand index of HfO used for the calculation, which was measured by a spectroscopic ellipsometer. spectroscopic ellipsometer. 2 imaginary parts of the refractive index of HfO2 used for the calculation, which was measured by a

Refractive index Refractive index

spectroscopic ellipsometer.

3 2.5 3

2.5 2

nReal kImaginary nReal kImaginary

1.5 2

1.5 1 0.5 1

0.5

0 200

0 200

400

600

800

1000

800

1000

Wavelength [nm]

400

600

Figure index HfO calculation. Figure3.3.Refractive Refractive indexof of HfO22 used for calculation. Wavelength [nm]

Figure 3. Refractive index of HfO2 used for calculation.

Materials 2018, 11, 1924

4 of 11

Materials 2018, 11, x FOR PEER REVIEW

4 of 11

Materials 2018, 11, x FOR PEER REVIEW

4 of 11

Figure 44shows showsthe the calculated transmittance spectra penetrating a Si substrate for grating several Figure calculated transmittance spectra penetrating a Si substrate for several grating periods. Here, a/Λ was fixed to be 0.8. Figure 4 illustrates that the peak wavelength Figure shows calculated transmittance spectra penetrating a Si substrate for grating of periods. Here,4a/Λ wasthe fixed to be 0.8. Figure 4 illustrates that the peak wavelength ofseveral the transmission theperiods. transmission spectrum can mainly through theThe grating period. The peak wavelength Here, was fixed tobe becontrolled 0.8. Figurethe 4 illustrates that the peak wavelength of the transmission spectrum can bea/Λ controlled mainly through grating period. peak wavelength is shifted to the is shifted to the longer wavelength side as the period increases, which agrees with the principle of spectrum can be controlled mainly through grating period. The peak wavelength is shifted to the longer wavelength side as the period increases, which agrees with the principle of guided-mode longer wavelength side as the period increases, which agrees with the principle of guided-mode guided-mode resonant gratings. From this result, we could design the structures that exhibit functions resonant gratings. From this result, we could design the structures that exhibit functions of gratings. Froma this result,and we could designeffect, the structures that exhibit of functions of of resonant spectroscopes, including a reflection and interference effect, duetotothe the influence of the interface spectroscopes, including reflection interference due influence the interface spectroscopes, including a reflection and interference effect, due to the influence of the interface between the SiO space and Si photodiode. The output signal actually obtained by the color sensor between the SiO22space and Si photodiode. The output signal actually obtained by the color sensor is the 2 space and Si photodiode. The output signalof actually obtained by spectrum the color sensor is in is between outputted a characteristic signal,which which theproduct product of the transmission transmission spectrum shown in outputted as as aSiO characteristic signal, isisthe the shown outputted as a characteristic signal, which is the product of the transmission spectrum shown in Figure 4 and the spectral sensitivity characteristic of the Si photodiode. Figure 4 and the spectral sensitivity characteristic of the Si photodiode. Figure 4 and the spectral sensitivity characteristic of the Si photodiode. 0.6

0.6

Λ [nm]

Λ [nm] 250 250 300 300 350 350 400 400

0.5

Transmittance

Transmittance

0.5

0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 400 400

500 600 500 600 Wavelength [nm] Wavelength [nm]

700 700

Figure Calculatedtransmittance transmittancespectra spectrapenetrating penetratingaaaSiSi substrate for several grating periods. a/Λa/Λ Figure 4. 4. Calculated transmittance spectra penetrating Sisubstrate substrate for several grating periods. for several grating periods. a/Λ is is fixed to be 0.8. fixed to to bebe 0.8. is fixed 0.8.

Figure 5shows showsthe the peak wavelength as aaasfunction function of period from thethe results Figure thepeak peak wavelength a function ofgrating the grating period extracted from the Figure 55shows wavelength as ofthe the grating periodextracted extracted from results of Figure 4. It can be seen that the peak wavelength shifts in proportion to the grating period thethe results of4.Figure It canthat be the seen thatwavelength the peak wavelength shifts intoproportion theingrating of Figure It can 4. be seen peak shifts in proportion the gratingtoperiod in entireinvisible wavelength range. The linear fitting equation for equation Figure 5 is given by 5the following period the entire visiblerange. wavelength range.fitting The linear fitting Figure given by the entire visible wavelength The linear equation for Figure 5for is given byisthe following equation: following equation:

equation:

λpeak + 163.4 λpeak== 11.24 .24 ××ΛΛ+ 163 .4

(1) (1)

λpeak = 1.24 × Λ + 163.4

(1)

Here, wavelength. Here,λpeak λpeak is is aa peak peak wavelength. Here, λpeak is a peak wavelength.

750

Peak wavelength [nm]

Peak wavelength [nm]

750

700

700

650

650

600

600

550

550

500

500 450 450 400

200

400 200

250

250

300

350

400

300 period 350[nm] 400 Grating

450

450

[nm] from Figure5.5.Peak Peakwavelength wavelengthas as aa function function Grating of grating period, results of of Figure 4. 4. Figure gratingperiod period,extracted extracted fromthe the results Figure Figure 5. Peak wavelength as a function of grating period, extracted from the results of Figure 4. 3.2. Spectrum Examples 3.2. SpectrumReconstruction: Reconstruction:Principle Principle and Calculation Calculation Examples

AnAn output signal from each color sensor is the amount of theofenergy spectrum received output signal from each color sensor is integrated the integrated amount the energy spectrum 3.2. Spectrum Reconstruction: Principle and Calculation Examples byreceived each photodiode and it cannot decomposed into spectral components. Therefore, in order by each photodiode and be it cannot be decomposed into spectral components. Therefore, in to An output signal from each color sensor is the integrated amount of the energy spectrum received by each photodiode and it cannot be decomposed into spectral components. Therefore, in

Materials 2018, 11, 1924

5 of 11

reconstruct the characteristic of the input light from the output signal, calculation 5between Materials 2018, 11, spectral x FOR PEER REVIEW of 11 output signals is required. To solve the inverse problem of determining the input signals (spectrum to reconstruct of the input light from output signal, oforder incident light) fromthe thespectral outputcharacteristic signals, spectral characteristics arethe calculated using calculation the Tikhonov between output signals[42]. is required. Toasolve the inverse problem of determining thesensors. input signals regularization method Consider spectroscopic device composed of b color b is the (spectrum of incident light)The from the output signals, spectral are calculated using the number of color sensors. photocurrent of each color characteristics sensor is obtained as a product of the Tikhonov regularization method [42]. Consider a spectroscopic device composed of b color sensors. spectral characteristic of incident light and the wavelength sensitivity characteristic of each color b is theintegrated number ofover colorthe sensors. The photocurrent of each color sensor is in obtained a product of the sensor, wavelength. The photocurrent is expressed matrixas format as follows: spectral characteristic of incident light and the wavelength sensitivity characteristic of each color sensor, integrated over the wavelength. The photocurrent is expressed in matrix format as follows: (2) O = SI

O = SI (2) Here, O (A) is a 1 × b column vector of photocurrent. S (A/W) is a b × c matrix of wavelength Here,and O (A) is a 1is×ab1column vector vector of photocurrent. S (A/W) is a b × c of matrix of wavelength sensitivity, I (W) × c column of spectral characteristics incident light. c is a sensitivity, and I (W) is a 1 × c column vector of spectral characteristics of incident light. c is a wavelength division number. O and S are measured experimentally, and I is to be determined. division number.problem, O and SIare measured using experimentally, and is to be determined. In Inwavelength order to solve this inverse is calculated the matrix, M,Iobtained by the Tikhonov order to solve this inverse problem, I is calculated using the matrix, M, obtained by the Tikhonov regularization method, as shown below: regularization method, as shown below: I = MO (3) I = MO (3) M is a c × b matrix. Besides, because the spectral characteristic cannot be a negative value, M is a ccondition × b matrix.isBesides, the following added: because the spectral characteristic cannot be a negative value, the following condition is added: I = I (I > 0) or 0 (I ≤ 0) (4)

I = I (I > 0 ) or 0 (I ≤ 0 ) (4) Equations (3) and (4) are solved to obtain the spectral characteristics. Next, several light spectra werethe designed spectra) and compared with the Equations (3)incident and (4) are solved to obtain spectral(designed characteristics. incident light spectra (reconstructed spectra) calculated using Equations (3) compared and (4). Itwith should Next, several incident light spectra were designed (designed spectra) and the be incident light spectra (reconstructed spectra) calculated using Equations (3) and (4). It should benm. noted that b was set to 25 and the calculated wavelength range to be integrated was 400 to 700 noted that bcolor was set to 25with and the calculated wavelength to be210 integrated was 400attointervals 700 nm. of Twenty-five filters different grating periodsrange between and 450 nm colorfor filters with different grating periods between 210 and 450 at intervals of 104.nm 10Twenty-five nm were used the matrix, S. Some of the 25 filter characteristics arenm shown in Figure Here, were used for the matrix, S. Some of the 25 filter characteristics are shown in Figure 4. Here, the the sensitivity of the Si photodiode is ignored for simplicity. sensitivity of the Si photodiode is ignored for simplicity. The calculation results are shown in Figure 6. As shown in Figure 6a,b, the peak positions of the Thespectrum calculationand results shown in Figure 6. As shown in Figure 6a,b,other, the peak of the designed the are reconstructed spectrum coincide with each andpositions the bandwidths designed spectrum and the reconstructed spectrum coincide with each other, and the bandwidths also substantially coincide. However, the sharp edges of the designed spectrum take rounded shapes also substantially coincide. However, the sharp edges of the designed spectrum take rounded shapes like the Gaussian distribution in the reconstructed spectrum. This can be attributed to the coarse like the Gaussian distribution in the reconstructed spectrum. This can be attributed to the coarse resolution because just 25 filters are provided across a wavelength range of 300 nm. It can be improved resolution because just 25 filters are provided across a wavelength range of 300 nm. It can be by increasing the number of filters. Moreover, as can be seen from Figure 6c,d, it is possible to improved by increasing the number of filters. Moreover, as can be seen from Figure 6c,d, it is possible calculate the results for multiple peaks. Figure 6d reveals that as the width of the designed spectrum to calculate the results for multiple peaks. Figure 6d reveals that as the width of the designed becomes narrower, the peak position of reconstructed data is in good agreement with the designed spectrum becomes narrower, the peak position of reconstructed data is in good agreement with the peak position, the peak of the reconstructed spectrum decreases. TheseThese results suggest designed peakbut position, butintensity the peak intensity of the reconstructed spectrum decreases. results that optical spectra in a wavelength range of 400 to 700 nm can be obtained by using the proposed suggest that optical spectra in a wavelength range of 400 to 700 nm can be obtained by using the devices; however, is room forisimprovement. proposed devices;there however, there room for improvement. 1.2

1

1

0.8

0.8

Intensity

Intensity

1.2

0.6 0.4 0.2

0.6 0.4 0.2 0

0 400

500

600

700

400

500

600

Wavelength[nm]

Wavelength[nm]

(a)

(b) Figure 6. Cont.

700

Materials 2018, 11, 1924

6 of 11

Materials 2018, 11, x FOR PEER REVIEW Materials 2018, 11, x FOR PEER REVIEW

6 of 11 6 of 11

1.2 1.2

1.2 1.2 1 1

0.8 0.8

0.8 0.8 0.6 0.6 0.4 0.4

Intensity Intensity

Intensity Intensity

1 1 0.6 0.6 0.4 0.4 0.2 0.2 0 0 400 400

500 500

600 600

Wavelength[nm] Wavelength[nm]

700 700

0.2 0.2 0 0 400 400

(c) (c)

500 500

600 600

Wavelength[nm] Wavelength[nm]

700 700

(d) (d)

Figure Simulation spectral reconstruction red line: Reconstructed Figure 6. 6. Simulation (Blue line:Designed Designedspectrum; spectrum; red line: Reconstructed Figure 6. Simulationofof ofspectral spectralreconstruction reconstruction(Blue (Blueline: line: Designed spectrum; red line: Reconstructed spectrum). (a) Single band spectrum near nm in wavelength; (b) single band spectrum spectrum). (a) Single band spectrum near 700 nm in wavelength; (b) single band spectrum a center spectrum). (a) Single band spectrum near 700 nm in wavelength; (b) single band spectrum at at aat a center center wavelength of 500 nm; (c) multiple band-stop spectrum; (d) multiple peak spectrum. wavelength of 500 nm; (c) multiple band-stop spectrum; (d) multiple peak spectrum. wavelength of 500 nm; (c) multiple band-stop spectrum; (d) multiple peak spectrum.

4. 4. Fabrication 4. Fabrication Fabrication Figure 7 shows the process steps. steps. Ann-type n-type substrate Si substrateaof a 400 thickness µm thickness is etched by a Figure Figure 77 shows shows the the process process steps. An An n-type Si Si substrate of of a 400 400 μm μm thickness is is etched etched by by aa fast fast fast atom beam (FAB) to form alignment marks (Figure 7a,b). Next, to prevent contamination, atom beam (FAB) to alignment marks (Figure 7a,b). Next, to metal contamination, an atom beam (FAB) to form form alignment marks (Figure 7a,b). Next, to prevent prevent metalmetal contamination, an 2 protective film with a thickness of 5 nm is formed by chemical vapor deposition (CVD) (Figure anSiO SiO protective film with a thickness of 5 nm is formed by chemical vapor deposition (CVD) 2 SiO2 protective film with a thickness of 5 nm is formed by chemical vapor deposition (CVD) (Figure 7c). PP ions implanted into contact area electrodes to n+ -Si After (Figure 7c). Then, ions are implanted the contact of electrodes n+ -Si7d). (Figure 7c). Then, Then, ionsPare are implanted into the theinto contact area of of area electrodes to form form to n+form -Si (Figure (Figure 7d). After7d). that, ion implantation of B is performed to form p-Si (Figure 7e). Besides, rapid thermal annealing is After that, implantation is performed to form (Figure Besides, rapid thermal annealing that, ionion implantation of BofisBperformed to form p-Sip-Si (Figure 7e). 7e). Besides, rapid thermal annealing is to activate ions simultaneously with the recovery of crystal from damage due to ion is performed performed to activate ions simultaneously with the recovery of crystal from damage due to performed to activate ions simultaneously with the recovery of crystal from damage due to ionion implanted ions (Figure 7f). Next, as spacer, SiO Then, as aawaveguide implanted ions (Figure 7f). isformed formedby byCVD CVD(Figure (Figure7g). 7g). Then, a waveguide implanted ions (Figure 7f).Next, Next,as asaaaspacer, spacer,SiO SiO222is formed by CVD (Figure 7g). Then, asas waveguide layer, HfO 2 is deposited by electron-beam (EB) evaporation (Figure 7h). layer, HfO deposited (Figure7h). 7h). layer, HfO 2 is depositedby byelectron-beam electron-beam(EB) (EB) evaporation evaporation (Figure 2 is (a) n-Si wafer (a) n-Si wafer

(b) Si etching (FAB) (b) Si etching (FAB)

(c) SiO2 deposition (CVD) (c) SiO2 deposition (CVD)

(g) SiO2 deposition (CVD) (g) SiO2 deposition (CVD)

(h) HfO2 deposition (EB evaporation) (h) HfO2 deposition (EB evaporation)

(l) SiO2 deposition (sputtering) (l) SiO2 deposition (sputtering)

(m) Etching for contact holes (m) Etching for contact holes

(i) Etching for contact holes (i) Etching for contact holes

(d) Ion implantation (P) (d) Ion implantation (P) (j) Al-Si deposition and patterning (j) Al-Si deposition and patterning (e) Ion implantation (B) (e) Ion implantation (B)

(f) Rapid thermal annealing (f) Rapid thermal annealing

(k) Al deposition and patterning (k) Al deposition and patterning

n-Si n-Si nn++-Si -Si p-Si p-Si SiO SiO2

HfO HfO22 Al Al

Al-Si(1%) Al-Si(1%)

2

Figure Figure 7. 7. Process Process steps. steps. Figure 7. Process steps.

After that, SiO 2 and HfO2 etching is performed to make contact holes (Figure 7i). Next, the Al-Si After that, SiO 2 and HfO2 2etching etching is is performed performed to Next, thethe Al-Si After that, SiO HfO to make makecontact contactholes holes(Figure (Figure7i). 7i). Next, Al-Si 2 and (1%) layer with a thickness of 400 nm is formed wet etching is performed to (1%) layer with a thicknessof of400 400nm nmis isformed formed by by sputtering, sputtering, and and to form form (1%) layer with a thickness by sputtering, andwet wetetching etchingisisperformed performed to form electrodes (Figure 7j). After Al film formation by sputtering, nanodot array structures are patterned electrodes (Figure 7j). After Al film formation by sputtering, nanodot array structures are patterned electrodes (Figure 7j). After Al film formation by sputtering, nanodot array structures are patterned by by EB followed by FAB (Figure 7k). Next, an SiO cover layer is formed by EB lithography, lithography, followed FAB etching etching 7k).an Next, SiO2layer 2 cover layer is formed by EBby lithography, followed by FABbyetching (Figure(Figure 7k). Next, SiO2 an cover is formed by sputtering sputtering (Figure 7l). Finally, SiO 2 in the electrode pad portion is etched (Figure 7m). sputtering (Figure 7l). Finally, SiO 2 in the electrode pad portion is etched (Figure 7m). (FigureFigure 7l). Finally, SiO2 in the electrode pad portion is etched (Figure 7m). filters fabricated on a shows reflection image of an of Figure8 88shows showsaaareflection reflectionimage image of of an an optical optical microscope microscope ofofcolor color onon a a Figure optical microscope colorfilters filtersfabricated fabricated photodiode photodiode array. array. The The filters filters are are fabricated fabricated in in 25 25 patterns, patterns, with with grating grating periods periods between between 220 220 and and 460 460 photodiode array. The 10 filters are fabricated in 25 (a) patterns, with grating periods between 220 ×and 460 nm nm nm at at increments increments of of 10 nm, nm, which which correspond correspond (a) to to (y) (y) in in the the figure. figure. All All filters filters are are 150 150 μm μm × 150 150 μm μm at in increments of 10 nm, color which correspond (a) to (y) in the figure.type, All filtersstructural are 150 µm × 150 µm in in size. size. Although Although the the color filter filter is is designed designed as as aa transmissive transmissive type, the the structural color, color, which which size. Althoughthe the color filter is designed as a transmissive type, thethe structural color, which depends on depends depends on on the period period of of the the structure, structure, can can also also be be confirmed confirmed in in the reflected reflected images. images. The The wiring wiring thepattern periodconnected of the structure, can also be confirmed in the reflected images. The wiring pattern connected pattern connected from from each each photodiode photodiode can can also also be be confirmed. confirmed. from each photodiode can also be confirmed.

Materials 2018, 11, x FOR PEER REVIEW Materials 2018, 11, 1924

7 of 11 7 of 11

Figure 9 11, shows scanning Materials 2018, x FOR PEER REVIEWelectron

microscope (SEM) images of the fabricated 7color of 11 filters corresponding to thescanning symbols (a) to (y) microscope of Figure 8. All filters are designed to be 0.8 ofcolor the a/Λ ratio. Figure 9 shows electron (SEM) images of the fabricated filters 9 shows scanning electron microscope (SEM) images of the fabricated color filters It can beFigure seen to that allsymbols the filters(a) are accurately. corresponding the tofabricated (y) of Figure 8. All filters are designed to be 0.8 of the a/Λ corresponding to the symbols (a) to (y) of Figure 8. All filters are designed to be 0.8 of the a/Λ ratio. ratio. It can be seen that all the filters are fabricated accurately. It can be seen that all the filters are fabricated accurately.

Figure 8. Optical microphotographsofof fabricated color filters on photodiode array. (a) Λ(a) =(a) 220 Figure Optical microphotographs of fabricated color filters on the photodiode array. =220 220nm, nm, Figure 8. 8.Optical microphotographs fabricated color filters onthe the photodiode array. ΛΛ =nm, (b) Λ = 230 nm, (c) Λ = 240 nm, (d) Λ = 250 nm, (e) Λ = 260 nm, (f) Λ = 270 nm, (g) Λ = 280 nm, (h) Λ = (b) Λ = 230 nm, (c) Λ = 240 nm, (d) Λ = 250 nm, (e) Λ = 260 nm, (f) Λ = 270 nm, (g) Λ = 280 nm, (h) Λ= (b) Λ = 230 nm, (c) Λ 240 nm, (d) Λ = 250 nm, (e) Λ = 260 nm, (f) Λ = 270 nm, (g) Λ = 280 nm, 290 nm, (i) Λ = 300 nm, (j) Λ = 310 nm, (k) Λ = 320 nm, (l) Λ = 330 nm, (m) Λ = 340 nm, (n) Λ = 350 nm, Λ = (i) 300Λnm, (j) Λ = 310 = 320 nm, (l)320 Λ = nm, 330 nm, Λ = 340 Λ 340 = 350nm, nm, (h)290 Λ nm, = 290(i)nm, = 300 nm, (j) nm, Λ = (k) 310Λnm, (k) Λ= (l) Λ(m) = 330 nm,nm, (m)(n) Λ= (o) Λ = 360 nm, (p) Λ = 370 nm, (q) Λ = 380 nm, (r) Λ = 390 nm, (s) Λ = 400 nm, (t) Λ = 410 nm, (u) Λ = 360nm, nm,(o) (p)ΛΛ== 360 370 nm, nm, (q) ΛΛ = 390 nm,nm, (s) Λ (t) Λ(s) = 410 Λ= (n)(o)ΛΛ= =350 (p) Λ Λ ==380 370nm, nm,(r)(q) = 380 (r)=Λ400 = nm, 390 nm, Λ =nm, 400(u) nm, 420 nm, (v) Λ = 430 nm, (w) Λ = 440 nm, (x) Λ = 450 nm, and (y) Λ = 460 nm. Λ (w) = 450 and (y)(x)ΛΛ == 460 nm. Λ (u) = 430 = 440 (t)420 Λ =nm, 410(v) nm, Λ =nm, 420(w) nm,Λ(v) Λ =nm, 430(x) nm, Λ nm, = 440 nm, 450 nm, and (y) Λ = 460 nm.

Figure 9. SEM images of fabricated color filters. (a) Λ = 220 nm, (b) Λ = 230 nm, (c) Λ = 240 nm, (d) Λ = 250 nm, (e) Λ = 260 nm, (f) Λ = 270 nm, (g) Λ = 280 nm, (h) Λ = 290 nm, (i) Λ = 300 nm, (j) Λ = 310 nm, (k) Λ = 320 nm, (l) Λ = 330 nm, (m) Λ = 340 nm, (n) Λ = 350 nm, (o) Λ = 360 nm, (p) Λ = 370 nm, (q) Λ = Figure SEM of (s) fabricated color (a)(u) ΛΛ = =220 (b)ΛΛ= = 240 nm, Λ =Λ 440=nm, 3809.nm, (r) Λimages = 390 nm, Λ = 400 nm, (t) Λfilters. = 410 nm, 420nm, nm, (v) 430230 nm,nm, (w) (c) Figure 9. SEM images of fabricated color filters. (a) Λ = 220 nm, (b) Λ = 230 nm, (c) Λ = 240 nm, nm, (d) Λ (d) Λ(x) = 250 nm,nm, (e)and Λ =(y)260 (f) Λ = 270 nm, (g) Λ = 280 nm, (h) Λ = 290 nm, (i) Λ = 300 Λ = 450 Λ =nm, 460 nm.

250 nm,nm, (e) Λ(k) = 260 Λ =(l) 270 (g)nm, Λ = 280 = 290(n) nm, Λ = 300 Λ 360 = 310nm, nm, (j)=Λ = 310 Λ =nm, 320(f) nm, Λnm, = 330 (m) nm, Λ =(h) 340Λnm, Λ (i) = 350 nm,nm, (o) (j) Λ= (k) Λ = 320 nm, (l) Λ = 330 nm, (m) Λ = 340 nm, (n) Λ = 350 nm, (o) Λ = 360 nm, (p) Λ = 370 nm, (q) Λ= (p) Λ = 370 nm, (q) Λ = 380 nm, (r) Λ = 390 nm, (s) Λ = 400 nm, (t) Λ = 410 nm, (u) Λ = 420 nm, Λ =(w) 390Λnm, (s)nm, Λ = 400 (t) Λ = 410 = 420 (v)380 Λ =nm, 430(r) nm, = 440 (x) Λnm, = 450 nm, andnm, (y) (u) Λ =Λ460 nm.nm, (v) Λ = 430 nm, (w) Λ = 440 nm, (x) Λ = 450 nm, and (y) Λ = 460 nm.

Materials 2018, 11, 1924

8 of 11

Materials 2018, 11, x FOR PEER REVIEW

8 of 11

5. Measured Results and Discussion

5. Measured Results and Discussion

Spectral sensitivity characteristics of color sensors were measured using a broadband spectral Spectral sensitivity characteristics of color sensors were measured using a broadband spectral response measurement system (CEP-25BXS, KeikiCo., Co.,Ltd., Ltd., Tokyo, Japan). response measurement system (CEP-25BXS, Bunkou Bunkou Keiki Tokyo, Japan). In this In this measurement, the wavelength resolution was 10 nm. nm.Different Different color sensors exhibited different measurement, the wavelength resolution wasset setto to 10 color sensors exhibited different sensitivity characteristics. Figure shows therelationship relationship between period andand the spectralspectral sensitivity characteristics. Figure 10 10 shows the betweenthe thegrating grating period the peak wavelength of the sensitivity. spectral sensitivity. The wavelength peak wavelength is linearly dependenton onthe the grating peak wavelength of the spectral The peak is linearly dependent grating period, and the slope of the straight line almost agrees with that of the calculated peak period, and the slope of the straight line almost agrees with that of the calculated peak wavelength wavelength shift of the color filters shown in Figure 5. The linear fitting equation for Figure 10 is shift of the color filters shown in Figure 5. The linear fitting equation for Figure 10 is given by the given by the following equation: following equation: λ = 1.28 × Λ + 105.4 (5) λpeakpeak= 1.28 × Λ + 105.4 (5)

Peak wavelength [nm]

700 650 600 550 500 450 400 200

300 400 Graging period [nm]

500

10. Dependence of the peak wavelength of of the sensitivity on the period.period. Figure Figure 10. Dependence of the peak wavelength thespectral spectral sensitivity ongrating the grating

When the monochromatic with wavelengthsofof450, 450, 500, 500, 550, nmnm is incident, When the monochromatic lightlight with wavelengths 550,600, 600,and and650 650 is incident, the the spectral characteristics of the incident light are reconstructed from the output signals obtained by spectral characteristics of the incident light are reconstructed from the output signals obtained by the the respective color sensors. Spectral characteristics of the reconstructed incident light are shown respective Spectral characteristics of thecharacteristics reconstructed incident shown with withcolor solid sensors. lines in Figure 11. For comparison, spectral of the incidentlight light are measured solid lines Figure 11. For comparison, spectral characteristicsOcean of theoptics, incident light measured withina commercially available spectrometer (HR4000CG-UV-NIR, Inc., Largo, FL, USA) with are shownavailable as originalspectrometer spectra with dotted lines. The peak wavelength the reconstructed spectrum a commercially (HR4000CG-UV-NIR, Oceanofoptics, Inc., Largo, FL, USA) are canoriginal be confirmed to bewith near dotted that of the original Therefore, we that the spectroscopic shown as spectra lines. Thespectrum. peak wavelength ofbelieve the reconstructed spectrum can measurement was successfully performed using the fabricated device. However, reconstructed be confirmed to be near that of the original spectrum. Therefore, we believe that the spectroscopic spectra become broader compared to the original spectra. This increment in width can be attributed measurement was successfully performed using the fabricated device. However, reconstructed spectra to the coarse resolution because only 25 filters are provided across a wavelength range of 300 nm. becomeThe broader compared tocan thebeoriginal This in width can be attributed to reconstructed spectra improvedspectra. by increasing theincrement number of filters. the coarse resolution because only 25 filters aretheprovided across aof wavelength Figure 12 shows the relationship between peak wavelengths reconstructed range spectra of and300 nm. the center wavelength of monochromatic lights extracted from the results of Figure 11. It is found The reconstructed spectra can be improved by increasing the number of filters. Materials 2018, 11, x FOR PEER REVIEW 9 of 11 that the peak wavelengths of the reconstructed spectra are in good agreement with the center wavelengths of the monochromatic lights.

Figure11. 11.Spectral Spectral reconstruction. Figure reconstruction.

700

h [nm]

650 600

Materials 2018, 11, 1924

9 of 11

Figure 12 shows the relationship between the peak wavelengths of reconstructed spectra and the center wavelength of monochromatic lights extracted from the results of Figure 11. It is found that the peak wavelengths of the reconstructed spectra are in good agreement with the center wavelengths of the monochromatic lights. Figure 11. Spectral reconstruction. 700

Peak wavelength [nm]

650 600 550 500 450

Original spectrum Reconstructed spectrum

400 400 450 500 550 600 650 700 Center wavelength of monochromatic light [nm] Figure 12. Dependence of the peak wavelengths of reconstructed spectra on the center wavelength of the incident monochromatic monochromatic lights. lights.

6. Conclusions 6. Conclusions We arrayarray spectroscope consisting of metamaterial color filterscolor and Sifilters photodiodes. We fabricated fabricateda filter a filter spectroscope consisting of metamaterial and Si The color filters based on guided-mode resonant metal gratings were newly designed. The peak photodiodes. The color filters based on guided-mode resonant metal gratings were newly designed. wavelength of the transmission spectrum could be controlled mainly by the grating period. The filters The peak wavelength of the transmission spectrum could be controlled mainly by the grating period. were fabricated in 25 patterns, with grating periodsperiods between 220 and at increments of The filters were fabricated in 25 patterns, with grating between 220 460 and nm 460 nm at increments 10 nm. SEM images of fabricated color filters confirmed that all the filters were fabricated accurately. of 10 nm. SEM images of fabricated color filters confirmed that all the filters were fabricated The peak wavelength of the spectral sensitivity found to befound linearly the grating accurately. The peak wavelength of the spectralwas sensitivity was to dependent be linearly on dependent on period. Several monochromatic lights were made incident on the fabricated device and the spectral the grating period. Several monochromatic lights were made incident on the fabricated device and characteristics of the incident light were reconstructed from the output signals obtained the the spectral characteristics of the incident light were reconstructed from the output signals from obtained respective color sensors. peak wavelengths of the reconstructed spectra were in good agreement from the respective colorThe sensors. The peak wavelengths of the reconstructed spectra were in good with the center wavelengths of the monochromatic lights. agreement with the center wavelengths of the monochromatic lights. Author Contributions: Conceptualization, Y.K.; methodology, Y.K.; software, D.E.; validation, Y.K., D.E. and K.H.; Author Contributions: Conceptualization, Y.K.; methodology, Y.K.; software, D.E.; validation, Y.K., D.E. and formal analysis, Y.K. and D.E.; investigation, Y.K., D.E. and K.H.; resources, Y.K., D.E. and K.H.; data curation, K.H.;and formal Y.K. anddraft D.E.;preparation, investigation, Y.K., D.E. and K.H.; Y.K., D.E. and K.H.; D.E. Y.K.; analysis, writing—original Y.K.; writing—review andresources, editing, Y.K.; visualization, D.E.data and curation, D.E. and Y.K.; writing—original draft funding preparation, Y.K.; Y.K. writing—review and editing, Y.K.; Y.K.; supervision, K.H.; project administration, Y.K.; acquisition, visualization, D.E. and Y.K.; supervision, K.H.; project administration, Y.K.; funding acquisition, Y.K. Funding: A part of this research was funded by the MEXT KAKENHI 16K13648 and 16H04342. Funding: A part of this research wasH. funded the MEXT KAKENHI 16K13648 and 16H04342. Acknowledgments: Authors thank Sai forby providing technical support for spectral response measurement. A part of this research was performed in Micro/Nano-Machining Research and Education Center (Tohoku University) and the Center for Integrated Nano Technology Support (Tohoku University) with assistance of MEXT Nanotechnology Platform. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

Ebbesen, T.W.; Lezec, H.J.; Ghaemi, H.F.; Thio, T.; Wolff, P.A. Extraordinary optical transmission through sub-wavelength hole arrays. Nature 1998, 391, 667–669. [CrossRef] Genet, C.; Ebbesen, T.W. Light in tiny holes. Nature 2007, 445, 39–46. [CrossRef] [PubMed] Kaplan, A.F.; Xu, T.; Guo, L.J. High efficiency resonance-based spectrum filters with tunable transmission bandwidth fabricated using nanoimprint lithography. Appl. Phys. Lett. 2011, 99, 143111. [CrossRef]

Materials 2018, 11, 1924

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

15.

16. 17.

18.

19. 20. 21. 22. 23. 24. 25. 26. 27.

10 of 11

Xu, T.; Wu, Y.K.; Luo, X.; Guo, L.J. Plasmonic nanoresonators for high-resolution colour filtering and spectral imaging. Nat. Commun. 2010, 1, 59. [CrossRef] [PubMed] Lee, H.S.; Yoon, Y.T.; Lee, S.S.; Kim, S.H.; Lee, K.D. Color filter based on a subwavelength patterned metal grating. Opt. Express 2007, 15, 15457–15463. [CrossRef] [PubMed] Chen, Q.; Cumming, D.R.S. High transmission and low color cross-talk plasmonic color filters using triangular-lattice hole arrays in aluminum films. Opt. Express 2010, 18, 14056–14062. [CrossRef] [PubMed] Chou, S.Y.; Krauss, P.R.; Renstrom, P.J. Nanoimprint lithography. J. Vac. Sci. Technol. B 1996, 14, 4129–4133. [CrossRef] Chou, S.Y.; Krauss, P.R.; Zhang, W.; Guo, L.; Zhuang, L. Sub-10 nm imprint lithography and applications. J. Vac. Sci. Technol. B 1997, 15, 2897–2904. [CrossRef] Bao, L.R.; Cheng, X.; Huang, X.D.; Guo, L.J.; Pang, S.W.; Yee, A.F. Nanoimprinting over topography and multilayer three-dimensional printing. J. Vac. Sci. Technol. B 2002, 20, 2881–2886. [CrossRef] Zhang, W.; Chou, S.Y. Fabrication of 60-nm transistors on 4-in. wafer using nanoimprint at all lithography levels. Appl. Phys. Lett. 2003, 83, 1632–1634. [CrossRef] Alkaisi, M.M.; Jayatissa, W.; Konijn, M. Multilevel nanoimprint lithography. Curr. Appl. Phys. 2004, 4, 111–114. [CrossRef] Roy, E.; Kanamori, Y.; Belotti, M.; Chen, Y. Enhanced UV imprint ability with a tri-layer stamp configuration. Microelectron. Eng. 2005, 78–79, 689–694. [CrossRef] Kanamori, Y.; Okochi, M.; Hane, K. Fabrication of antireflection subwavelength gratings at the tips of optical fibers using UV nanoimprint lithography. Opt. Express 2013, 21, 322–328. [CrossRef] [PubMed] Ito, S.; Kikuchi, E.; Watanabe, M.; Sugiyama, Y.; Kanamori, Y.; Nakagawa, M. Silica imprint templates with concave patterns from single-digit nanometers fabricated by electron beam lithography involving argon ion beam milling. Jpn. J. Appl. Phys. 2017, 56, 06GL01. [CrossRef] Lee, S.W.; Lee, K.S.; Ahn, J.; Lee, J.J.; Kim, M.G.; Shin, Y.B. Highly sensitive biosensing using arrays of plasmonic Au nanodisks realized by nanoimprint lithography. ACS Nano 2011, 5, 897–904. [CrossRef] [PubMed] Peer, A.; Biswas, R. Extraordinary optical transmission in nanopatterned ultrathin metal films without holes. Nanoscale 2016, 8, 4657–4666. [CrossRef] [PubMed] Nishiguchi, K.; Sueyoshi, K.; Hisamoto, H.; Endo, T. Fabrication of gold-deposited plasmonic crystal based on nanoimprint lithography for label-free biosensing application. Jpn. J. Appl. Phys. 2016, 55, 08RE02. [CrossRef] Peer, A.; Hu, Z.; Singh, A.; Hollingsworth, J.A.; Biswas, R.; Htoon, H. Photoluminescence enhancement of CuInS2 quantum dots in solution coupled to plasmonic gold nanocup array. Small 2017, 13, 1700660. [CrossRef] [PubMed] Hutley, M.C. Diffraction gratings. In Diffractive Optics for Industrial and Commercial Applications; Turunen, J., Wyrowski, F., Eds.; Akademie Verlag: Berlin, Germany, 1997; pp. 59–80. ISBN 3055017331. Ema, D.; Kanamori, Y.; Sai, H.; Hane, K. Plasmonic color filters integrated on a photodiode array. Electron. Commun. Jpn. 2018, 101, 95–104. [CrossRef] Chang, C.C.; Lin, N.T.; Kurokawa, U.; Choi, B.I. Spectrum reconstruction for filter-array spectrum sensor from sparse template selection. Opt. Eng. 2011, 50, 114402. [CrossRef] Kurokawa, U.; Choi, B.I.; Chang, C.C. Filter-Based Miniature Spectrometers: Spectrum Reconstruction Using Adaptive Regularization. IEEE Sens. J. 2011, 11, 1556–1563. [CrossRef] Chang, C.C.; Lin, H.Y. Spectrum Reconstruction for On-Chip Spectrum Sensor Array Using a Novel Blind Nonuniformity Correction Method. IEEE Sens. J. 2012, 12, 2586–2592. [CrossRef] Chang, C.C.; Su, Y.J.; Kurokawa, U.; Choi, B.I. Interference Rejection Using Filter-Based Sensor Array in VLC Systems. IEEE Sens. J. 2012, 12, 1025–1032. [CrossRef] Oliver, J.; Lee, W.B.; Lee, H.N. Filters with random transmittance for improving resolution in filter-array-based spectrometers. Opt. Express 2013, 21, 3969–3989. [CrossRef] [PubMed] Oliver, J.; Lee, W.; Park, S.; Lee, H.N. Improving resolution of miniature spectrometers by exploiting sparse nature of signals. Opt. Express 2012, 20, 2613–2625. [CrossRef] [PubMed] Norton, S.M.; Erdogan, T.; Morris, G.M. Coupled-mode theory of resonant-grating filters. J. Opt. Soc. Am. A 1997, 14, 629–639. [CrossRef]

Materials 2018, 11, 1924

28. 29. 30. 31. 32. 33. 34.

35. 36.

37.

38. 39.

40. 41. 42.

11 of 11

Jacob, D.K.; Dunn, S.C.; Moharam, M.G. Design considerations for narrow-band dielectric resonant grating reflection filters of finite length. J. Opt. Soc. Am. A 2000, 17, 1241–1249. [CrossRef] Brundrett, D.L.; Glytsis, E.N.; Gaylord, T.K.; Bendickson, J.M. Effects of modulation strength in guided-mode resonant subwavelength gratings at normal incidence. J. Opt. Soc. Am. A 2000, 17, 1221–1230. [CrossRef] Peters, D.W.; Kemme, S.A.; Hadley, G.R. Effect of finite grating, waveguide width, and end-facet geometry on resonant subwavelength grating reflectivity. J. Opt. Soc. Am. A 2004, 21, 981–987. [CrossRef] Kobayashi, T.; Kanamori, Y.; Hane, K. Surface laser emission from solid polymer dye in a guided mode resonant grating filter structure. Appl. Phys. Lett. 2005, 87, 151106. [CrossRef] Kanamori, Y.; Kitani, T.; Hane, K. Guided-Mode Resonant Grating Filter Fabricated on Silicon-on-Insulator Substrate. Jpn. J. Appl. Phys. 2006, 45, 1883–1885. [CrossRef] Kanamori, Y.; Kitani, T.; Hane, K. Control of guided resonance in a photonic crystal slab using microelectromechanical actuators. Appl. Phys. Lett. 2007, 90, 031911. [CrossRef] Kanamori, Y.; Katsube, H.; Furuta, T.; Hasegawa, S.; Hane, K. Design and Fabrication of Structural Color Filters with Polymer-Based Guided-Mode Resonant Gratings by Nanoimprint Lithography. Jpn. J. Appl. Phys. 2009, 48, 06FH04. [CrossRef] Kanamori, Y.; Ozaki, T.; Hane, K. Fabrication of Ultrathin Color Filters for Three Primary Colors Using Guided-Mode Resonance in Silicon Subwavelength Gratings. Opt. Rev. 2014, 21, 723–727. [CrossRef] Kanamori, Y.; Ozaki, T.; Hane, K. Reflection color filters of the three primary colors with wide viewing angles using common-thickness silicon subwavelength gratings. Opt. Express 2014, 22, 25663–25672. [CrossRef] [PubMed] Moharam, M.G.; Grann, E.B.; Pommet, D.A.; Gaylord, T.K. Formulation for stable and efficient implementation of the rigorous coupled-wave analysis of binary gratings. J. Opt. Soc. Am. A 1995, 12, 1068–1076. [CrossRef] Lalanne, P.; Morris, G.M. Highly improved convergence of the coupled-wave method for TM polarization. J. Opt. Soc. Am. A 1996, 13, 779–784. [CrossRef] Smith, D.Y.; Shiles, E.; Inokuti, M. The Optical Properties of Metallic Aluminum. In Handbook of Optical Constants of Solids; Palik, E.D., Ed.; Academic Press: Orlando, FL, USA, 1985; Volume 1, pp. 369–406. ISBN 0125444206. Edwards, D.F. Silicon (Si). In Handbook of Optical Constants of Solids; Palik, E.D., Ed.; Academic Press: Orlando, FL, USA, 1985; Volume 1, pp. 547–569. ISBN 0125444206. Philipp, H.R. Silicon Dioxide (SiO2) (Glass). In Handbook of Optical Constants of Solids; Palik, E.D., Ed.; Academic Press: Orlando, FL, USA, 1985; Volume 1, pp. 749–763. ISBN 0125444206. Aster, R.C.; Borchers, B.; Thurber, C.H. Parameter Estimation and Inverse Problems, 2nd ed.; Academic Press: Waltham, MA, USA, 2012; pp. 93–127. ISBN 9780128100929. © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).