Discussion

In the Figure 18, the response of TM light should be zero theoretically, because the TM light which is diffracted by gratings has the electric field always paralleled with the surface. However, no matter how we change the direction of polarization, the response of TM light is not zero.

Figure 19 shows the spectral response of L36 under different polarization. When polarization is 45-degree, the responsivity reaches maximum. Therefore, the light with 45-degree is called TE mode. On the other hand, the light with 135-degree is called TM mode. Although TM response is minimum, it is not zero. This is because edge light coupling produces response of TM mode, and we will discuss this phenomenon in the next section.

(2) Edge Light Coupling

The TM light absorption comes from edge light coupling. The device without any surface structure is measured in two different ways. One is the 45-degree facet light coupling and the other is normal light incident. The experiment of edge light coupling is shown in Figure 20. In the Figure 21, the response of edge light coupling for normal incident is about half of 45-degree facet coupling. By experimental data in Figure 21, we realize that the edge light coupling has great influence on our sample.

Fig. 18 The refraction of TE and TM light.

(a) (b)

Fig. 19 Experiment of edge light coupling (a) edge light coupling for normal light incident (b) 45-degree facet coupling

hν

Ceramic Plate

Gold wire

Gold Pad

hν

Ceramic Plate

Gold wire

Gold Pad

0 20 40 60 80 100 120 140 160 180 0.018

0.020 0.022 0.024 0.026 0.028 0.030 0.032 0.034

Degree (

⁰

)

TE

TM L36

Re s ponsi v it y( A/ W )

Fig. 20 The spectral response of L36 under different polarization.

5 6 7 8 9 10 11 12

0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035

edge light coupling 45-degree facet coupling

Re sp onsi v it y( A/ W )

Wavelength(

μ

m)

P1.2V

Fig. 21 The response of 45-degree facet coupling and edge light coupling.

Substrate

1⁰

75⁰

Fig. 22 The edge light coupling scheme.

5 6 7 8 9 10 11 12

0.000 0.001 0.002 0.003 0.004 0.005 0.006

TE

TM

Re sp onsi v it y( A/ W )

Wavelength(

μ

m)

P1.2V

Fig. 23 The spectral response of edge light coupling under different polarization.

Figure 22 shows the edge light coupling scheme. When the light is incident with 1-degree, the refraction angle can be calculated by Snell’s Law.

n 1 sinθ1= n 2 sinθ2 (1)

The refraction angle is 15-degree. Therefore, the refraction light has large fraction of electrical field which is perpendicular to the superlattice plane. This is why the TM light can be absorbed by our device. In Figure 23, the response does not change too much no matter what the polarization light is incident. This is because the photoresponse is from edge coupling light of the mesa edge. Consequently, the light coupling efficiency depend on mesa height and mesa shape.

As the mesa height is deeper, the response can be observed more obviously.

(3) Response with different period grating

The introduction of surface plasmon is discussed previously. Figure 24 shows the transmission of our sample. Because the wavelength is longer than trench width, the transmission decays as the wavelength increase theoretically. However, there is a 9

μ m peak in L33 device. It is attributed to

the surface plasmon between the GaAs and metal. The peak in L36 device is at 13.5

μ m, and it is

not in the detection region. The plamon peak is not as sharp as other papers. This is because the mesa size is only 300

μ m× 300 μ m. There are only twenty five slips on the mesa of L33.

Therefore, the transmission enhancement by surface plasmon is not obvious.

The TM response is due to edge light coupling and the TE response is caused by edge light coupling and grating structure. In above section, we have described that TE response is equal to TM response in the edge coupling part. Therefore, the total TE response minus the total TM response leaves response of grating structure. Figure 25 shows the response of grating structure.

The shape of L39 in Figure 25 is similar to the P1.2V response in Figure 7. This result is expected by transmission in Figure 24, because the transmission spectrum of L39 in long wavelength is flat.

In L36 case, the transmission rises when wavelength is longer than 8

μ m. Therefore, the response

of L36 is sharper than L39 at long wavelength. There is a 9

μ m peak in L33 case. We can observe

the phenomenon in Figure 25. The response of L33 is larger than L36 in 8.5

μ m ~ 9.5 μ m region

in Figure 25. Figure 26 shows the diagram of responsivity vs. open air fraction at different wavelength. The short wavelength response does not match with the transmission spectrum, because of the low group velocity and relaxation. By dispersion relation, the group velocity of short wavelength is slow, and the relaxation effect will be serious. Therefore, the responsivity saturates under high open air fraction.

(4) Quantum efficiency and diffraction angle

The internal quantum efficiency for unpolarized light is given by

4 2

1{1 exp( 1.62910 )}

2

i

N

d

W s N Sin

η

= − − ⁻ × × × × ×

θ

(2) where N is doping density, W is well width, S is number passes,

θ

is incident angle.

Therefore, the internal quantum efficiency calculated from Eq. (2) is 0.03.

The external quantum efficiency is given by

^0.738 _{i e}

Aeff η η= ^xt (3)

The external quantum efficiency of 45-degree facet coupling is 0.01566. Because the response is proportional to external quantum efficiency, we can obtain the external quantum efficiency of L39. The external quantum efficiency of L39 is 0.02. The internal quantum efficiency of L39 calculated from Eq. (3) is 0.027. Therefore, the diffraction angle is derived by using Eq. (2). The diffraction angle of L39 is 67-degree. By the same method, the diffraction angle of L33 and L36 are 68-degree and 67-degree.

5 6 7 8 9 10 11 12 13 14 15 16 17 18

0.00 0.05 0.10 0.15 0.20 0.25

L39 L36 L33

In ten s ity (a.u .)

Wavelength(

μ

m)

Transmission

Fig. 24 The transmission of L33, L36, and L39

5 6 7 8 9 10 11 12

Fig. 25 The response come from grating structure.

25 30 35 40 45 50

Fig. 26 The responsivity vs. open air fraction

D. 結論與計畫自評

In this project, superlattice infrared photodetectors with gratings for coupling the normal incident light are fabricated and studied. Base on this structure, response shape is tunable due to surface plasmon. We focus on the response of different open air fraction and compare the response with transmission.

The experiment results show that grating structure can effectively couple the normal incident light for topside-illumination and the response is affected by surface plasmon in L33 structure.

By means of the experiment demonstration, the TM response is due to edge light coupling and the TE response is caused by edge light coupling and grating structure. The detectors with grating have good IR absorption spectrum of TE polarization. The transmission is enhanced by surface plasmon. There is a plasmon peak in the detection region. This is why the response shape is tunable. Thus, the applicability of the grating structure coupling is verified.

There are some options to improve the present work:

(1) In order to achieve the plasmon peak in our detection region, we can try to use different trench width or period of grating structure.

(2) The metal thickness can be optimized to enhance the plasmon peak.

(3) We can change the mesa size to increase the amount of trenches on the mesa.

(4) The diffraction effect caused by the grating should be studied further to enhance the coupling efficiency.

E. 參考文獻

[1] 1 J.A. Porto, F.T. Garcia-Vidal, and J.B. Pendry, Phys. Rev. Lett. 83, 2845 (1999).

[2] L. Martin-Moreno, F.J. Garcia-Vidal, H.J. Lezec, K.M. Pellerin, T. Thio, J.B. Pendry, and T.W.

Ebbesen, Phys. Rev. Lett. 86, 1114 (2001).

[3] Q. Cao and P. Lalanne, Phys. Rev. Lett. 88, 057403 (2002).

[4] U. Schröter and D. Heitmann, Phys. Rev. B 58, 15 419 (1998).

[5] E. Popov, M. Nevière, S. Enoch, and R. Reinisch, Phys. Rev. B 62, 16100 (2000).

[6] H. Räther, Surface Plasmons on Smooth and Rough Surfaces and on Gratings (Springer-Verlag, Berlin, 1988).

二、參考文獻

詳見各部分後面所附

三、計畫自評

由於計畫之中包含數個子計畫，計畫自評的段落安排在各個子題目的討論及自評之 中，以利於作各個子題目之自評與討論，在此不再撰寫。

附件一：出席國際會議心得

會議名稱：2005 雷射與光電研討會

Conference on Lasers and Electro-Optics 2005 (CLEO 2005)

一、 參加會議經過

CLEO 2005 是每一年都會舉行的光電領域的研討會，今年是在巴爾的摩舉辦，會議的主題 包含所有與光電相關的議題，包含：雷射應用，固態雷射，半導體雷射，光儲存技術，非 線性光學應用，光學材料與薄膜，超快光學，高場物理，光學元件，生醫應用，先進光學 感測，光纖與導波雷射，發光二極體以及量子電子的應用。與會的人士除了包含論文發表 者，相關領域研究人員還有全美所有重要的光電廠商。會議進行除了一般的邀請演講，口 頭及壁報論文的發表，還有就是以廠商展示新商品為主的類似光電展的盛會。

會議一共進行六天的時間，參加的人數有四，五千人之多，參與的廠商也有三百多家，發 表的論文超過兩千篇，所以每天都有不同主題的論文發表，和大會安排的邀請演講。中間 三天有大會安排的廠商光電商品展，是我覺得整個會議最有看頭的部分，偌大的會場擠滿 了各式各樣的光電產品，處處都有令人驚奇的發現。

這次大會特別為學生安排了一系列的活動，我覺得也是相當值得參加的。包括了學生論壇，

請一些學者參與學生的早餐或午餐會分享經驗，還有教學生一些發表科學論文的方法，以 及提供了許多工作或實習的機會讓大家參與等等。

我的壁報論文被安排在第四天的中午發表，我的研究主題是以超晶格紅外線偵測器為主，

並向大家介紹我們在低電壓操作下的努力成果，在發表的過程中也結識了一些與自己領域 相關的學者專家，互相討論的結果深深讓我覺得受用無窮。其他的時間我則是選擇一些有 興趣的主題論文發表參與或參加上述的學生專屬活動。

二、 與會心得

由於是國際性的光電盛會，所以在參加會議的過程中，可以不斷的吸收有關光電領域的最 新發展與最新知識。我這次有幸以壁報論文被接受而參與此次盛會，得以站在這個國際性 的殿堂與全世界的光電工作者一起參與其中，感到十分榮幸。除了一般會議都有的邀請演 講，請到的都是各領域大師級的人物外，其餘的論文發表也讓我對於光電領域的深度與廣 度有了新的一番認識。

雖然大多數的主題和我所做的題目並不相關，但可以藉此瞭解各領域的發展，並得知科技 進步的程度，也是令人難忘的經驗。尤其是來自世界各地的廠商無不卯足全勁讓大家看到 自己公司的最新發展，各式各樣意想不到的雷射產品，光電元件，讓人看的目不暇給，對 於科技的發展的進步又有了新的體認。

這次參與會議，對於許多研究團隊都是跨校或甚至跨國際的研究計畫感到驚訝，尤其是看 到全世界許多的頂尖學者的合作研究得到了驚人的成果，讓我深深覺得台灣應該要努力走 向國際化，才能將視野提升到國際級的水準，雖然台灣現在的研究成果已經在世界上佔有

了一席之地，不過相對來說，我們的國際合作機會並不普遍，也許這是我們以後應該努力 的目標，做出更好的研究成果以吸引國際上一流的研究團隊與我們合作。

三、 建 議

相對於我們一般在國內參與的所謂國際性的研討會大多都是台灣研究團隊參加，國外的研 討會所提供的深度與廣度確實不是國內研討會可以相提並論的。在這樣國際性的舞台之 上，雖然也看到了不少台灣人參與其中，可是也希望有朝一日台灣可以自行舉辦類似規模 的研討會，一方面可以讓我們參與會議者可以免除舟車勞頓之苦，也可以提升台灣的國際 能見度。

現今國家政府財政並不寬裕，所以補助學生的款項並不充足，雖然我此次有幸承蒙教育部 補助，但金額連機票都不夠，對我們這些參與會議的學生而言，出過開會不啻為沈重的經 濟負擔，也許以後學校可以以類似助學貸款的方式先行補助學生的出國費用，待以後再慢 慢償還。出國開會對學生與國際學者接觸是很重要的，除了可以加強自己的外語能力，也 可以證明自己的研究成果足以站在世界的舞台，對自己也是一種肯定。如果可能的話，希 望能將補助金額提高，或尋求相關的解決方式，以減輕學生的經濟壓力。

四、 攜回資料名稱及內容

大會會議相關資訊，所有發表論文的光碟，光電廠商具參考價值之廣告文宣，與相關研究 者之通訊資料等。

五、 其 他

無

附件二：出席國際發表論文

Low-Voltage Operation Photodetector Made by Coupling Superlattice and Quantum Wells

J. H. Lu¹, Y. C. Wang, C. L. Wang, C. H. Kuan²

Graduate Institute of Electronics Engineering and Department of electrical engineering, National Taiwan University No.1, Sec. 4, Roosevelt Road, Taipei, Taiwan, Republic of China 10617

1[email protected] ²[email protected]

C. W. Yang, S. L. Tu

Opto Tech Corporation, 1, Li-hsin Rd. V, Hsinchu Science-based industrial Park, Hsinchu, Taiwan, Republic of China

Abstract : We have investigated a novel photodetector structure of coupling superlattice and quantum wells. This device can be operated at low bias range and even the photovoltaic mode. The broadband response is achieved by this structure.

© 2005 Optical Society of America

OCIS codes: (040.3060) Infrared ; (230.5590) Quantum-well devices

1. Introduction

In recent years, the applications such as the focal plane array (FPA) and remote temperature sensing drive the research of novel intersubband transition devices. Some useful characteristics of quantum well infrared photodetector (QWIP) including multicolor, photovoltaic are also investigated [1-3]. In addition to QWIP, the superlattice infrared photodetector (SLIP) is another promising structure to achieve those applications. Compared with the quantum well, superlattice has the properties including low power consumption, broadband photoresponse and voltage-tunable. The performance of SLIP has been proved by our group [4-6]. The perspective of our detector is to operate at low bias, therefore, low power consumption and low dark current are expected. We also hope the operation temperature can be higher and the broadband and flatband spectrum can be observed. Based on our perspective, a new structure of superlattice coupled to quantum wells is designed.

Figures 1(a) and (b) show the band diagrams of our sample under photovoltaic and photoconductive mode, respectively. Under zero bias, the photovoltaic mode, electrons in superlattice excited to the second miniband will go through the barrier by group velocity and captured by some quantum wells. Because of those captured electrons, the potential on superlattice side is relatively positive and the quantum well is negative. Therefore, a built-in potential will exist and the energy band will be bended as shown in Fig. 1(a). On the other hand, due to the dopant migration during the growth process, a built-in electric field is created at the barrier in quantum well structure. Then, electrons in quantum well excited to the bound state or the continuum band can tunnel through the barrier and become the photocurrent. Hence, under zero bias, the photocurrent from superlattice and quantum wells can be measured simultaneously.

For the photoconductive mode, Fig. 1(b) shows our sample operated under negative bias. When operated under

negative bias, this structure works just as the general photodetector. Electrons in superlattice excited by infrared radiation will go through the barrier and be accelerated by the electrical field on the quantum well structure. At the same time, electrons in quantum well can also overcome the barrier and then become the photocurrent. On the other hand, under positive bias, electrons will be attracted to top contact. In the superlattice, the electrons transport to top contact and leave the superlattice a positive electric field. Therefore, electrons will be attracted back to superlattice

negative bias, this structure works just as the general photodetector. Electrons in superlattice excited by infrared radiation will go through the barrier and be accelerated by the electrical field on the quantum well structure. At the same time, electrons in quantum well can also overcome the barrier and then become the photocurrent. On the other hand, under positive bias, electrons will be attracted to top contact. In the superlattice, the electrons transport to top contact and leave the superlattice a positive electric field. Therefore, electrons will be attracted back to superlattice

在文檔中 兼具偵測多波段紅外線及可見光的偵測器陣列模組之研發 (3/3) (頁 64-78)