The operation principle of the well in quantum dot stack infrared photodetectorJheng-Han Lee, Zong-Ming Wu, Yu-Min Liao, Yuh-Renn Wu, Shih-Yen Lin, and Si-Chen Lee

Citation: Journal of Applied Physics 114, 244504 (2013); doi: 10.1063/1.4849875 View online: http://dx.doi.org/10.1063/1.4849875 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/114/24?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Tuning the dynamic properties of electrons between a quantum well and quantum dots J. Appl. Phys. 112, 043702 (2012); 10.1063/1.4746789 Multiple-photon peak generation near the ˜10 m range in quantum dot infrared photodetectors J. Appl. Phys. 109, 064510 (2011); 10.1063/1.3556432 Two photon absorption in quantum dot-in-a-well infrared photodetectors Appl. Phys. Lett. 92, 023501 (2008); 10.1063/1.2833691 Origin of photocurrent in lateral quantum dots-in-a-well infrared photodetectors Appl. Phys. Lett. 88, 213510 (2006); 10.1063/1.2207493 High-performance 30-period quantum-dot infrared photodetector J. Vac. Sci. Technol. B 23, 1129 (2005); 10.1116/1.1900730

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

The operation principle of the well in quantum dot stack infraredphotodetector

Jheng-Han Lee,1 Zong-Ming Wu,1 Yu-Min Liao,2 Yuh-Renn Wu,2 Shih-Yen Lin,1,3

and Si-Chen Lee1,2,a)

1Graduate Institute of Electronics Engineering, National Taiwan University, Taipei 10617, Taiwan2Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei 10617, Taiwan3Research Center for Applied Sciences, Academia Sinica, Taipei 115299, Taiwan

(Received 5 July 2013; accepted 3 December 2013; published online 30 December 2013)

The well in the quantum dot stack infrared photodetector (WD-QDIP) is proposed which can be

operated at high temperature �230 K. The operation principle of this device is investigated,

including the carrier transport and the enhancement in the photocurrent. The WD-QDIPs with

different well numbers are fabricated to study the mechanisms. It is realized that the carrier

transport from the emitter to the collector in traditional quantum dot infrared photodetectors

consists of two channels deduced from current-voltage characteristics and dark current activation

energy at different temperatures. At temperatures below 77 K, the current transports through the

InAs quantum dot channel, whereas at temperatures higher than 77 K, the current is dominated by

the GaAs leakage channel. In addition, the non-equilibrium situation at low temperatures is also

observed owing to the presence of photovoltaic phenomenon. The carrier distribution inside the

QDs is simulated to investigate the reasons for the increase of photocurrent. Based on the

simulation and the photocurrent response, the hot carrier (electron) scattering effect by the

insertion of a quantum well layer is inferred as the most probable reason that lead to the

enhancement of the response and regarded as the key factor to achieve high- temperature

operation. VC 2013 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4849875]

I. INTRODUCTION

Quantum dot infrared photodetectors (QDIPs) have been

studied for many years owing to their outstanding perform-

ance including high sensitivity of normally incident light,

long life time, and high-temperature operation.1–6 The tradi-

tional view regarded the dark current being contributed from

carriers (electrons) escaping from QDs by thermionic emis-

sion, field-assisted tunneling, and defect tunneling when

QDs are n-type doped or the carrier transport is conducted at

low temperature.7 Therefore, several structures to block the

dark current6,8,9 or enhance the quantum efficiency3,4,10 or

use the multiband detection with the tunneling barrier (T-

QDIP) and superlattice combined with the AlGaAs graded

barrier to block the dark current and select the photo sig-

nals11 were proposed to lower the dark current and enhance

the operation temperature. They all depend on a complicated

and precise epitaxial process to lower the dark current and

ensure high responsivity and detectivity at high tempera-

tures. However, in our study, it is found that after only the

adding insertion of a quantum well (QW) in a quantum dot

stack infrared photodetector (WD-QDIP), the device could

be operated at a high temperature (230 K) even though it has

larger dark current than a standard InAs/GaAs QDIP without

the QW insertion layer.

In reality, the GaAs leakage channel should be taken

into account12 especially for the QDIPs with undoped QD

layers. In this case, the carriers inside the QDs which can be

photo-excited out of the QDs and become photo-carriers are

provided by the dark leakage current that is driven by ther-

mal and biases. Because of small QD height (several mono-

layers), the high-energy hot electrons in the dark current

cannot be captured by QDs easily. Hence, the dark current

transport in undoped-QD infrared photodetectors is different

from the QDIPs with n-doped QD layers. In this study, the

two-channel carrier transport in the WD-QDIPs is demon-

strated. One channel is the QD channel and the other one is

the GaAs leakage channel. To confirm the two-channel

mechanism and understand the enhancement of the photo

current, the WD-QDIPs adopting one to three QW layers in

the QD stacks are fabricated to study the transport mecha-

nism in details.

II. EXPERIMENTS

Figures 1(a)–1(d) show the device structure of samples A

to E. Samples A and B were grown by VG gas-source molecu-

lar beam epitaxy (MBE) system. Samples C, D, and E were

grown by Riber Compact 21 solid-source MBE system.

Sample A is a standard ten-stack InAs/GaAs QDIP (No-

Well). Sample B is the WD-QDIP with a 10 nm-thick

In0.15Ga0.85As insertion layer in the middle of ten-stack QD

layers. Sample C has the same structure as sample B but with

different QD densities due to different growth conditions in

different MBE systems. Samples D and E are also the

WD-QDIPs with two and three 10 nm-thick In0.15Ga0.85As

insertion layers in the devices, respectively. Three InAs QD

layers were sandwiched between two neighboring insertion

a)Author to whom correspondence should be addressed. Electronic mail:

sclee@cc.ee.ntu.edu.tw. Tel.: þ886-2-23635251 ext. 440. Fax: þ886-2-

23675509.

0021-8979/2013/114(24)/244504/7/$30.00 VC 2013 AIP Publishing LLC114, 244504-1

JOURNAL OF APPLIED PHYSICS 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

layers and between the insertion layer and the contact layer in

samples D and E. Hence, the total QD stacks QDs are nine

and twelve in samples D and E, respectively. The 10 nm-thick

In0.15Ga0.85As insertion layers were doped with Si to a con-

centration of 1� 1017 cm�3 and the 2.4 ML undoped InAs

QDs were grown at 490 �C. The 50 nm-thick undoped GaAs

spacers, the 200 nm-thick top contact and 600 nm-thick bot-

tom contact doped with Si to a concentration of 2� 1018 cm�3

were grown at 580 �C. The device fabrication follows a

standard mesa etching and lift off process. The mesa and the

incident window areas are 380� 280 and 200� 200 lm2,

respectively. Afterwards, all the samples are analyzed by

temperature-dependent I-V and responsivity measurements.

The background photocurrent is measured at 10 K under the

illumination of 300 K blackbody surroundings. Owing to the

lower absorption ability of x-y polarized wave for the GaAs/

In0.15Ga0.85As/ GaAs QW structure, the carriers in QWs are

not excited by normally incident light easily, almost all the

photo-carriers come from QDs. Therefore, all the samples are

measured with normally incident light by the Bruker Vertex

70 Fourier transform infrared spectroscopy (FTIR) system

with MIR light global source to obtain the response spectra.

III. RESULTS AND DISCUSSION

The temperature-dependent I-V characteristics are

measured to study the carrier transport mechanism as shown

in Figs. 2(a)–2(e) for samples A to E, respectively. The back-

ground photocurrent measured at 10 K indicates the back-

ground limited performance (BLIP) temperature is about

50 K for all samples. The analyses of dark current reveal two

transport channels for carriers, i.e., the QD channel and the

GaAs leakage channel as shown in Fig. 3. When the temper-

ature is lower than 77 K, all the samples exhibit almost the

same turn-on voltage of �61 V. This turn-on voltage results

from the existence of built-in potential (�0.65 eV) estab-

lished by GaAs/InAs heterojunction in order to achieve ther-

mal equilibrium. At low temperatures, free carriers are

frozen out to silicon donors and the dense QDs pinch off the

GaAs leakage channel, so the carriers can only transport via

the QD channel. In this case, an external bias larger than 1 V

is required to suppress the built-in potential, and then drives

carriers to transport through the QD channel. At high temper-

atures (>77 K), carriers are thermally excited from donors to

conduction band, the GaAs leakage channel is opened. The

dark current begins to increase exponentially at small biases

and linearly at high biases. The exponential increase indi-

cates that the dark current is dominated by the thermionic

emission. The linear increase indicates the ohmic conduc-

tion. According to thermionic emission current formula:

Id¼A* T2 exp (�Ea/kT), where A* is the Richardson con-

stant, T is temperature in Kelvin, and Ea is the activation

energy, the activation energy at a fixed bias can be obtained

from the slope of the Arrhenius plot as shown in Figs. 4(a)

and 4(b) for samples A and B, respectively. The activation

energy at zero bias can be estimated by linear extrapolation

as shown in the inset of Figs. 4(a) and 4(b). It is about

0.05� 0.07 eV, which is much smaller than the built-in

potential of 0.65 eV. The activation energy for samples C to

E is similar to those of samples A and B. It signifies that the

carriers transport via the GaAs leakage channel instead of

QD channel at higher temperatures. A n-i-n homojunction

channel from the nþ GaAs emitter through the undoped

GaAs to the nþ GaAs collector could result in such a small

activation energy.

At low temperatures, a non-equilibrium phenomenon,

i.e., the photovoltaic effect is also observed in all samples

because of the presence of the constant dark current at a bias

range between �1� 1 V as shown in Fig. 2. The photovol-

taic effect is induced by two unequal processes, the photo-

excitation of electrons out of quantum dots (route 1) and the

thermal-assisted ionization of carriers from donors (route 2)

being injected into QDs as shown in Fig. 5. Basically, elec-

trons are thermalized from the donor level of the nþ GaAs

contact layer and injected into the QDs driven by the Fermi

level difference. It provides the QDs close to the nþ GaAs

layer to have higher carrier concentration. All of the Fermi

levels are equal eventually to achieve thermal equilibrium in

the device. However, at low temperatures, electrons have not

enough thermal energy to be ionized from the nþ contact. On

the contrary, electrons in the QDs are easily excited out of

QDs by weak background blackbody radiation to flow back

to nþ GaAs layer and then captured by silicon impurities.

Therefore, the Fermi levels on the two sides of the

GaAs/InAs heterojunction could not be equalized which

results in this thermal non-equilibrium and photovoltaic

FIG. 1. The schematic structures of quantum dot infrared photodetectors for

samples (a) A, (b) B and C, (c) D, and (d) E.

244504-2 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

FIG. 2. The temperature-dependent

I-V curves for sample (a) A, (b) B, (c)

C, (d) D, and (e) E.

FIG. 3. The schematic figure to describe the two-channel system. One chan-

nel is the QD channel, the other is the GaAs leakage channel.

FIG. 4. The Arrhenius plots of normalized dark current for sample (a) A and

(b) B. The inset in each figure indicates the activation at zero bias.

244504-3 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

effect. When the temperature is higher than 77 K, the photo-

voltaic effect disappears owing to the sufficient supply of

electron from donors which can be injected from nþ contact

into the QDs to achieve thermal equilibrium.

To investigate the effect of the doping in the QW, the

carrier distribution in high density QDs is simulated

with Gummel’s iteration method combined with Poisson-

Schrodinger equation, drift-diffusion equation, and self-

consistently at zero bias and high temperatures in the z

direction, as shown in Figs. 6(a) and 6(b) for samples A and

B. Owing to smaller thickness of QDs, the carrier distribu-

tion in QDs looks like many peaks. In Fig. 6, it is obvious

that QDs near the In0.15Ga0.85As QW have larger carrier con-

centrations owing to the contribution from doping carriers.

The average concentration distribution in every QD stack is

plotted in Figs. 7(a) and 7(b). For ten-stack QDs, an apparent

increase in carrier is found within two to three stacks QDs

near the QW as shown in Fig. 7(a). Even though increasing

the number of QWs from two to three, the carrier concentra-

tion from the first to sixth QD layer is almost same as shown

in Fig. 7(b), indicating the increase of carriers indeed occur

within two to three QD stacks near the QW. In fact, the

enhancement in the photo current arises not only from the

supplement of extra carriers provided by the original doping

carriers in the QW but from the scattered electrons when

high-energy hot electrons pass through the QW.13–16 The

source of high energy hot electrons is due to the electric

field. The electrons injected from nþ contact are drifted by

electric field and gain energies, they eventually form a distri-

bution with higher temperature than that of the local thermal

equilibrium. The scattered electrons are transferred to low

energy distribution and easily captured by adjacent QDs

when they flow out of the QW or are injected with the aid of

thermal field-assisted as shown in Figs. 8(a) and 8(b).

Therefore, the background photocurrent in sample B

(WD-QDIP) can be enhanced and kept larger than the dark

current when the temperature increases as shown in Fig.

9(a). Like the photocurrent in sample B, it is also larger than

the dark current from 10 to 230 K in samples C to E. The

insufficient carrier supplement in sample A (No-Well) leads

to smaller photo current than the dark current at high temper-

atures as shown in Fig. 9(b).

Figures 10(a)–10(e) show the response spectra of sam-

ples A to E measured at 10 K, respectively. In comparison

with the responsivity of sample A, samples B to E which

adopt the structure of WD-QDIP have higher responsivity

for larger photo current responses as shown in Fig 11(a). The

photocurrent responses are enhanced by two to ten times

FIG. 5. The non-equilibrium conduction band diagram to show the photo-

voltaic effect. The route 1 describes that electrons excited out of quantum

dots and trapped by the Si impurities. The route 2 describes the thermal-

assisted ionization process from the donor level into the quantum dots.

FIG. 6. The carrier distribution based

on simulation considering Poisson-

Schrodinger equation, drift-diffusion

equation, and self-consistently at zero

bias for sample (a) A and (b) B.

FIG. 7. The average carrier concentra-

tion in every QD stack for sample (a)

A, B, and C, and (b) D and E.

244504-4 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

depending on the biases. It results from the extra photo-

excited carriers after the adding of the In0.15GaAs0.85/GaAs

QW. However, the total carriers contributed by the doping

effect in samples B to E are just increased by 5% to 20%. It

implies that the scattered electrons dominate the photocur-

rent. With increasing the temperature, the hot electrons

become harder to be captured by QDs. Therefore, the adopt-

ing of QW to transform high-energy hot carriers to

low-energy scattered electrons becomes important. As men-

tioned above, the BLIP temperature of all the samples is

about 50 K. Actually, the real highest operation temperatures

for samples A to E are 50, 230, 100, 150, and 90 K, respec-

tively. It results from the photocurrent enhancement with the

increase of temperature. The different highest operation tem-

peratures for samples B to E is probably related to the den-

sity of QDs. A higher density and larger size of QDs shrink

the GaAs leakage channel and makes the scattered electrons

be captured by QDs easily, so the photocurrent response can

be increased more. Figure 11(b) shows a higher photocurrent

response measured at 77 K for samples B than others. The

photocurrent response and spectrum as shown in Fig. 11(c)

also indicate that sample B could be operated at 200 K with a

high responsivity (�0.14 A/W). As mentioned above, the

probably reason leading to sample B to have higher photo-

current is its higher density of QDs (5.2� 1011 cm�2) than

FIG. 8. The conduction band diagram showing the scattering and thermal-

field assisted tunneling effect that increase the photo current (a) through the

QD channel at low temperatures and (b) through the GaAs leakage channel

at high temperatures.

FIG. 9. The photo current and dark current measured from 10 to 230 K at

the bias of 1.2 V for sample (a) A and (b) B.

FIG. 10. The responsivity spectra measured at 10 K for sample (a) A, (b) B, (c) C, (d) D, and (e) E.

244504-5 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

the others as shown by the images of transmission electron

microscopy (TEM) in Fig. 12. The density of QDs for sam-

ples A, C, D, and, E is 3.5� 1011, 2.5� 1011, 4.3� 1011, and

2.2� 1011 cm�2. Therefore, the operation temperature of the

WD-QDIPs is also dependent on the density of QDs. For

sample A, the standard QDIP without the In0.15Ga0.85As

QW, has the lowest operation temperature even though it has

larger density of QDs than samples C and E. It means the hot

electrons are hard to be captured. In a word, for the

WD-QDIPs, the most important factor to affect the operation

temperature is not only the carrier supplement of low-energy

scattered electrons but the high density of QDs to pinch off

the leakage channels.

IV. CONCLUSIONS

In summary, the WD-QDIPs with different numbers of

QW are investigated. The underline principle for a high-

temperature operation of WD-QDIP is realized. The two-

channel conduction for current is found and the photovoltaic

effects are observed. They all are explained by temperature-

dependent I-V characteristics and the activation energy

measurements. The enhancement in the photo current is ana-

lyzed by the simulation of the carrier distribution and is

found to be dominated by low-energy scattered electrons.

The TEM images reveal that the density of QDs increases

the captured possibility of the scattered electrons to enhance

FIG. 11. The photo current response varied with biases for samples A to E at (a) 10 K and (b) 77 K.

FIG. 12. The TEM cross-section images for sample (a) A, (b) B, (c) C, (d) D, and (e) E.

244504-6 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31

the operation temperature. Therefore, a WD-QDIP with high

density of QDs could possibly improve the operation temper-

ature of QDIPs to room temperature.

ACKNOWLEDGMENTS

The authors would like to thank the National Science

Council of the Republic of China, Taiwan, for financially

supporting this research under Contract No. NSC 101-2120-

M-002-010.

1C. T. Huang, Y. C. Chen, and S. C. Lee, Appl. Phys. Lett. 100, 043512

(2012).2J. H. Lee, Y. T. Chang, C. J. Huang, S. Y. Lin, and S. C. Lee, IEEE

Photon. Technol. Lett. 22, 577–579 (2010).3H. S. Liang, S. Y. Wang, C. P. Lee, and M. C. Lo, Appl. Phys. Lett. 92,

193506 (2008).4H. Lim, S. Tsao, and M. Razeghi, Appl. Phys. Lett. 90, 131112 (2007).5S. Y. Lin, Y. R. Tsai, and S. C. Lee, Appl. Phys. Lett. 78, 2784 (2001).

6S. F. Tang, S. Y. Lin, and S. C. Lee, Appl. Phys. Lett. 78, 2428

(2001).7A. D. Stiff-Roberts, X. H. Su, S. Chakrabarti, and P. Bhattacharya, IEEE

Photon. Technol. Lett. 16, 867–869 (2004).8S. Y. Wang, S. D. Lin, H. W. Wu, and C. P. Lee, Appl. Phys. Lett. 78,

1023 (2001).9A. D. Stiff, S. Krishna, P. Bhattacharya, and S. W. Kennerly, IEEE J.

Quantum Electron. 37, 1412 (2001).10S. Chakrabarti, A. D. Stiff-Roberts, P. Bhattacharya, S. Gunapala, S.

Bandara, S. B. Rafol, and S. W. Kennerly, IEEE Photon. Technol. Lett.

16, 1361–1363 (2004).11A. G. U. Perera, G. Ariyawansa, G. Huang, and P. Bhattacharya, Infrared

Phys. Technol. 52, 252–256 (2009).12V. Ryzhii, I. Khmyrova, V. Pipa, V. Mitin, and M. Willander, Semicond.

Sci. Technol. 16, 331–338 (2001).13T. Yajima, Y. Hikita, and H. Y. Hwang, Nature Mater. 10, 198–201

(2011).14M. Heiblum, D. C. Thomas, C. M. Knodler, and M. I. Nathan, Appl. Phys.

Lett. 47, 1105 (1985).15M. Heiblum and M. V. Fischetti, IBM J. Res. Dev. 34, 530–549 (1990).16Y. C. Chao, M. H. Xie, M. Z. Dai, H. F. Meng, S. F. Horng, and C. S. Hsu,

Appl. Phys. Lett. 92, 093310 (2008).

244504-7 Lee et al. J. Appl. Phys. 114, 244504 (2013)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

131.104.62.10 On: Sun, 17 Aug 2014 11:38:31