Transparent Fourier transform spectrometerVladislav Jovanov,1 Eerke Bunte,2 Helmut Stiebig,2 and Dietmar Knipp1,*

1Jacobs University Bremen, School of Science and Engineering, Electronic Devices andNanophotonics Laboratory, 28759 Bremen, Germany

2Institute of Photovoltaics, Research Center Jülich, 52425 Jülich, Germany*Corresponding author: d.knipp@jacobs‑university.de

Received November 3, 2010; accepted December 15, 2010;posted December 22, 2010 (Doc. ID 137616); published January 13, 2011

We realized a transparent Fourier transform spectrometer and investigated its operating principle. The spectrometerconsists of a low-reflectivity Fabry–Perot interferometer and a partially transparent photodetector, which allows forthe operation of the spectrometer in transmission. By changing the distance between the low-reflectivity mirrors ofthe Fabry–Perot interferometer, the light intensity that reaches the partially transparent photodetector is modulatedand the spectral information of the incident light is determined by the Fourier transform of the generated photo-current. This transparent Fourier transform spectrometer allows easy miniaturization and integration into any kindof optical system. © 2011 Optical Society of AmericaOCIS codes: 300.6300, 120.6200, 230.5170.

Various types of Fourier transform microspectrom-eters based on the Michelson interferometer [1,2], thestanding-wave approach [3–8], and the Fabry–Perot in-terferometer [9–12] have been developed in recent years.Research on microspectrometers is aimed at developinga compact and portable spectrometer with low manufac-turing costs and good performance in order to enablepractical applications of spectrometry.The working principle of a Fourier transform spectro-

meter is based on obtaining the spectral information ofthe incident light from the recorded interference pattern(interferogram) by applying the Fourier transform. In thecase of a Michelson interferometer or a standing-wavespectrometer, the interferogram is created owing tothe two-beam interference of optical waves propagatingin the same or the opposite direction. On the other hand,in the case of a Fabry–Perot interferometer, the interfer-ogram is created by varying its transmission, whichdepends on the distance between the mirrors of theFabry–Perot interferometer.The operation principles of the Fourier transform spec-

trometer based on a Fabry–Perot interferometer can bedescribed by multiple-beam interference. Assuming thatthe Fabry–Perot interferometer is formed between twopartially transparent mirrors with the same reflectivity(R), the transmission of the Fabry–Perot interferometerfor normal light incidence is given by the Airy function[13]:

Tðd; λÞ ¼�1þ 4R

1 − R2 sin2

�2πdλ

��−1; ð1Þ

where d is the distance between the mirrors and λ is thewavelength of the incoming light. Assuming a low reflec-tivity of the mirrors, Eq. (1) becomes

Tðd; λÞ ≈�1 −

2R

1 − R2

�þ 2R

1 − R2 cos

�4πdλ

�: ð2Þ

From Eq. (2) we can observe that by changing the dis-tance between the mirrors, we are changing the transmis-sion and, consequently, the intensity of the transmittedlight. The transmitted light intensity can be separated

into dc and ac components, where the ac component car-ries the spectral information of the incoming light. Thephotocurrent of the photodetector is proportional tothe transmitted light intensity, and the spectral informa-tion of the incoming light can be determined by calculat-ing the Fourier transform of the ac component of thegenerated photocurrent.

The low-reflectivity Fabry–Perot interferometer is typi-cally formed by two parallel low-reflectivity mirrors. Theinterferogram is created by modulation of the distancebetween the mirrors, and it is recorded with a photode-tector placed behind the Fabry–Perot interferometer. Inour transparent spectrometer design, we used a partiallytransparent photodetector to record the created inter-ferogram. The partially transparent photodetector isdeposited on the second low-reflectivity mirror of theFabry–Perot interferometer (Fig. 1). Using this approach,we are able to obtain the spectral information of the in-coming light, while, at the same time, the spectrometeroperates in transmission.

The key component of the transparent Fourier trans-form spectrometer is an ultrathin and partially transpar-ent photodetector (Fig. 2), which consists of a p-i-nphotodiode sandwiched between two transparent con-ductive oxide (TCO) layers. The TCO layers act as

Fig. 1. (Color online) Schematic sketch of the transparentFourier transform spectrometer. A low-reflectivity Fabry–Perot interferometer is combined with a partially transparentdetector in order to realize spectrometers that can operate intransmission.

274 OPTICS LETTERS / Vol. 36, No. 2 / January 15, 2011

0146-9592/11/020274-03$15.00/0 © 2011 Optical Society of America

electrical contacts of the device and antireflection coat-ings. High transparency of the photodetector is requiredto allow for the spectrometer to operate in transmission.On the other hand, a certain fraction of the incident lighthas to be absorbed, so that a photocurrent can be gen-erated. Amorphous silicon was used as the materialfor the p-i-n photodiode, while zinc oxide was used asthe material for the TCO layers. Amorphous silicon exhi-bits excellent optical properties in the visible part of theoptical spectrum [14]. The overall thickness of the opticaldetector is less than 600 nm, while the thickness of thep-i-n photodiode itself is less than 100 nm. The thicknessof the active i-layer is less than 50 nm. More details on thematerials used and the fabrication process of partiallytransparent photodetector are given in [14–16].Figure 3 exhibits the experimentally measured quan-

tum efficiency and transmission of the partially trans-parent photodetector. The quantum efficiency of thephotodetector mainly depends on the generation of car-riers in the intrinsic layer of the p-i-n photodiode. Thephotogenerated carriers in the p- and n-layers recombinedue to the short carrier lifetime in the doped regions, andtherefore only the carriers created in the intrinsic layercontribute to the overall photocurrent. For short wave-lengths, the quantum efficiency is low owing to theabsorption in the zinc oxide layers. The partially trans-parent detector exhibits maximum quantum efficiencyof ∼35% for a wavelength of 470 nm. For longer wave-lengths, the quantum efficiency drops owing to the lowabsorption coefficient of amorphous silicon [14]. Thepartially transparent photodetector covers the visual partof the optical spectrum ranging from 380 to 680 nm. Thetotal transmission of the transparent Fourier transform

spectrometer depends mainly on the transmission prop-erties of the partially transparent photodetector. Thesensor exhibits a transmission higher than 10% for wave-lengths longer than 480 nm.

The low-reflectivity mirrors of the Fabry–Perot interfe-rometer are realized by two sheets of glass, where the sec-ond glass sheet acts also as a substrate for the partiallytransparent photodetector. The first glass sheet wasmounted on a piezomicroactuator, and the distance be-tween mirrors was modulated up to 6:25 μm. A He–Ne la-ser emitting at a wavelength of 633 nm was used as a testinput for the experimentally realized transparent Fouriertransform spectrometer. The “photocurrent-versus-time”curves obtained with the oscilloscope are mapped tothe “photocurrent-versus-mirror-displacement” curve, ac-counting for nonlinear motion of the piezomicroactuator.To determine the spectrum of the test input, a discreteFourier transform is applied to the ac component of themeasured photocurrent. The ac component of the mea-sured photocurrent is shown in Fig. 4(a). The photocur-rent is modulated by the displacement between themirrors of the Fabry–Perot interferometer, owing to thechange of its transmission. Figure 4(b) shows the opticalspectrum obtained from the ac component of the mea-sured photocurrent, and the test input wavelength of633 nm is properly distinguished.

The spectral range that can be measured correctly withthis spectrometer is limited by the formation of higherorder harmonics, which define the free spectral range ofthe Fabry–Perot interferometer. The free spectral rangeof the Fabry–Perot interferometer and the quantum

Fig. 2. (Color online) Partially transparent photodetectors.The photodetectors (p-i-n diodes) were deposited on a glasssubstrate.

Fig. 3. (Color online) Quantum efficiency and transmission ofthe partially transparent photodetector.

Fig. 4. (Color online) Output of partially transparent spec-trometer: (a) measured photocurrent (ac component) as a func-tion of displacement between low-reflectivity mirrors of theFabry–Perot interferometer and (b) optical spectrumof a He–Ne laser obtained with a transparent Fabry–Perotspectrometer.

January 15, 2011 / Vol. 36, No. 2 / OPTICS LETTERS 275

efficiency of the photodetector determine the spectro-meter’s operating range. The experimentally realizedtransparent Fourier transform spectrometer exhibitsan operation range from 380 to 680 nm, covering exactlythe visible part of optical spectrum. The spectral resolu-tion of the demonstrated spectrometer is common for allFourier transform spectrometers, and it is defined withRayleigh criterion:

Δk ¼ πΔLC

; Δλ ¼ λ22 ·ΔLC

; ð3Þ

where ΔLC is the total change of distance between themirrors. The resolution of the spectrometer in the wave-number domain depends only on the change of the dis-tance, whereas in the wavelength domain, the resolutiondepends on the cavity length change and the wavelength.From Eq. (3) it is clear that the ability of the spectrometerto distinguish nearby wavelengths increases for largermirror displacements. The spectral resolution deter-mined by Eq. (3) for a wavelength of 633 nm and mirrordisplacement of 6:25 μm is equal to 32 nm and showsgood agreement with experimentally obtained spectralresolution (the FWHM criteria), which is around 40 nm.In summary, a transparent Fourier transform spectro-

meter based on a low-reflectivity Fabry–Perot interfe-rometer was demonstrated. The operational range ofthe spectrometer covers the visible part of the opticalspectrum ranging from 380 to 680 nm. The spectrometeroperates in transmission for longer wavelengths, forwhich the partially transparent sensor exhibits suffi-ciently high transmission. The spectral resolution of thespectrometer depends on the displacement between thelow-reflectivity mirrors, and the experimentally achievedspectral resolution was 40 nm. The properties of thetransparent spectrometer can be tailored by selectingthe appropriate absorbing materials. For example, byusing crystalline silicon, the operating range can be ex-tended to the near-IR part of the optical spectra. Also, therealization of two-dimensional spectrometer arrays isfeasible by integration of an array of partially transparent

photodetectors on a transparent substrate, such assapphire, in a process compatible with classicalmicroelectronics [17].

References

1. O. Manzardo, H. P. Herzig, C. R. Marxer, and N. F. de Rooij,Opt. Lett. 24, 1705 (1999).

2. K. Yu, D. Lee, U. Krishnamoorthy, N. Park, and O. Solgaard,Sens. Actuators A 130–131, 523 (2006).

3. D. A. B. Miller, IEEE J. Quantum Electron. 30, 732 (1994).4. H. L. Kung, S. R. Bhalotra, J. D. Mansell, D. A. B. Miller, and

J. S. Harris, IEEE Sel. Top. Quantum Electron. 8, 98 (2002).5. D. Knipp, H. Stiebig, S. R. Bhalotra, E. Bunte, H. L. Kung,

and D. A. B. Miller, IEEE Trans. Electron. Dev. 52, 419(2005).

6. E. le Coarer, S. Blaize, P. Benech, I. Stefanon, A. Morand,G. Lérondel, G. Leblond, P. Kern, J. M. Fedeli, and P. Royer,Nat. Photon. 1, 473 (2007).

7. M. Renault, Y. Hadjar, S. Blaize, A. Bruyant, L. Arnaud,G. Lerondel, and P. Royer, Opt. Lett. 35, 3303 (2010).

8. M. Sasaki, X. Y. Mi, and K. Hane, Appl. Phys. Lett. 75,2008 (1999).

9. P. M. Zavracky, K. L. Denis, H. K. Xie, T. Wester, andP. Kelley, Proc. SPIE 3514, 179 (1998).

10. L. J. Chen, T. F. Kao, J. Y. Lu, and C. K. Sun, Opt. Express14, 3840 (2006).

11. M. Pisani and M. E. Zucco, Opt. Express 17, 8319 (2009).12. Y. Q. Deng, R. Kersting, V. Roytburd, J. Z. Xu, R. Ascazubi,

K. Liu, X. C. Zhang, and M. S. Shur, Int. J. Infrared Milli.Waves 25, 215 (2004).

13. E. Hecht, in Optics (Addison-Wesley, 2002), pp. 385–442.14. W. Luft and Y. Tuso, Hydrogenated Amorphous Silicon

Alloy Deposition Processes (Marcel Dekker, 1993).15. O. Kluth, A. Löffl, S. Wieder, C. Beneking, W. Appenzeller,

L. Houben, B. Rech, H. Wagner, S. Hoffmann, R. Waser,J. A. Selvan, and H. Keppnet, in Conference Record of the26th IEEE Photovoltaic Specialists Conference (IEEE,1997), pp. 715–718.

16. H. Stiebig, H.-J. Büchner, E. Bunte, V. Mandryka, D. Knipp,and G. Jäger, Thin Solid Films 427, 152 (2003).

17. G. Andreou, Z. K. Kalayjian, A. Apsel, P. O. Pouliquen,R. A. Athale, G. Simonis, and R. Reedy, IEEE Circuits Syst.Mag. 1, 22 (2001).

276 OPTICS LETTERS / Vol. 36, No. 2 / January 15, 2011