Ultrahigh-Q Photonic Crystal Nanocavity Realized with Locally Modulated Line Defect

 

Eiichi Kuramochi, Takasumi Tanabe, Akihiko Shinya and Masaya Notomi
Optical Science Laboratory

 An advantage of photonic crystal (PC) nanocavities is compatibility of a high quality factor (Q) with an ultra-small mode volume (V) [1].. A higher Q has been demanded in order to enhance device performance [2,3] and decrease the device insertion loss resulting from the reduction in transmittance caused by cavity-waveguide coupling. Although intensive studies have raised the theoretical Q value to 106 [1], a further increase will be difficult to achieve using conventional approaches.
 A new scheme for realizing a higher Q PC nanocavity involves a Fabry-Perot cavity composed of a loss-less line-defect mode and two mode-gap mirrors [4]. The use of a mode-gap mirror instead of a photonic bandgap (PC) mirror greatly reduces the out-of-plane loss from the cavity because of the smooth and gradual decay of the electromagnetic field in the mode-gap barrier region. We have found that such a mode-gap can be generated by modulating the line-defect width [1]. In this study, we designed a nanocavity in which holes around the cavity center were shifted away from a line defect. In addition, we employed a tapered shift structure (this 3-stage structure is shown in Fig. 1) in which a gradual change of line-defect width enhanced the advantage of the mode-gap mirror. Numerical studies with the 3D finite difference time domain method revealed that the 3-stage structure can achieve an ultrahigh-Q (108) with an ultrasmall V (1.7 (λ/n)3 ). Next, we fabricated this structure in a Si PC slab by using electron beam lithography [5]. The intensity spectrum of the cavity resonant mode is shown in Fig. 2. The ultranarrow linewidth (1.5 pm) corresponded to a Q of 1.0×106.
 In summary, our newly designed cavity has increased the theoretical Q value by two orders of magnitude and we achieved an experimental Q of one million for the first time as a PC-based nanocavity [6]. The Q achieved here is sufficiently high for many device applications and this report is a first step towards adding new functions (optical buffering and optical delay) to PC-based devices.

[1] M. Notomi et al., Opt. Express 12 (2004) 1551.
[2] M. Notomi et al., Opt. Express 13 (2005) 2678.
[3] T. Tanabe et al., Opt. Lett. 30 (2005) 2575.
[4] E. Kuramochi et al., Appl. Phys. Lett. 88 (2006) 041112.
[5] E. Kuramochi et al., Phys. Rev. B72 (2005) 161318(R).
[6] E. Kuramochi et al., LEOS2005, PD1.1 (2005).

Fig. 1. Schematic and SEM image.
Fig. 2. Theoretical Q and resonant spectrum.

[back] [Top] [Next]