Ultrabroadband photoconductive detection: Comparison with free-space electro-opticsamplingShunsuke Kono, Masahiko Tani, and Kiyomi Sakai

Citation: Applied Physics Letters 79, 898 (2001); doi: 10.1063/1.1394719 View online: http://dx.doi.org/10.1063/1.1394719 View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/79/7?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Generation and detection of ultrabroadband terahertz radiation using photoconductive emitters and receivers Appl. Phys. Lett. 85, 164 (2004); 10.1063/1.1768313 Detection of up to 20 THz with a low-temperature-grown GaAs photoconductive antenna gated with 15 fs lightpulses Appl. Phys. Lett. 77, 4104 (2000); 10.1063/1.1333403 Detectors and sources for ultrabroadband electro-optic sampling: Experiment and theory Appl. Phys. Lett. 74, 1516 (1999); 10.1063/1.123601 Comparison of terahertz waveforms measured by electro-optic and photoconductive sampling Appl. Phys. Lett. 73, 3184 (1998); 10.1063/1.122712 Coherent terahertz radiation detection: Direct comparison between free-space electro-optic sampling andantenna detection Appl. Phys. Lett. 73, 444 (1998); 10.1063/1.121894

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:

129.119.67.237 On: Sun, 17 Aug 2014 18:21:00

APPLIED PHYSICS LETTERS VOLUME 79, NUMBER 7 13 AUGUST 2001

This a

Ultrabroadband photoconductive detection: Comparison with free-spaceelectro-optic sampling

Shunsuke Kono,a) Masahiko Tani, and Kiyomi SakaiKansai Advanced Research Center, Communications Research Laboratory, 588-2, Iwaoka,Nishi-ku, Kobe 651-2492, Japan

~Received 13 February 2001; accepted for publication 18 June 2001!

The spectral response of a low-temperature-grown GaAs photoconductive~PC! antenna is comparedwith that of the electro-optic sampling technique for the frequency region up to 40 THz. The spectralresponse of the PC antenna for broadband detection is found to be determined by the temporalprofile of the number of photocarriers injected by ultrashort optical gate pulses. ©2001 AmericanInstitute of Physics.@DOI: 10.1063/1.1394719#

p-hellene

r

nr

fs-

en

thththol

oOherh

l r

tras

himnrabe

ighdon

gthgeandnd-

asss-

fo-

toionoff-edhethe-theeretter

ole

nts.

es

te..4nnationiontheablesng-ionpli-

edintem

satiai

Photoconductive~PC! antennas gated with ultrashort otical pulses have been widely used and studied for the coent detection of far-infrared electromagnetic transients caterahertz~THz! radiation. The response of the PC antenwas originally considered to be governed by the increasphoto-carrier number, however, their bandwidth has beenported to be up to several THz.1,2

We recently demonstrated that the PC antenna bawidth can be extended about three times wider than theported value by gating with optical pulses as short as 153

The achieved bandwidth of 20 THz3 means that the bandwidth of the low-temperature-grown GaAs~LT–GaAs! PCantenna can be as broad as that of the electro-optic~EO!sampling technique using thin EO crystals for the coherdetection of THz radiation higher than 10 THz.4,5

This fast response of the PC antenna is explained byrapid increase in the number of photocarriers injected byultrashort gating pulses. For ultrabroadband detection,PC antenna works as an integrating detector, so the phcurrent from the antenna is supposed to be proportionathe time integration of the incident THz radiation.6,7 In Ref.3, we pointed out that the calculated time differentiationthe PC-detected THz wave form was similar to the Edetected THz wave form. However, the differentiation of twave forms assumes the temporal behavior of the numbephotocarriers as a step function. Thus, further analysisbeen necessary to quantitatively estimate the spectrasponse of the PC antenna for ultrabroadband detection.

In this letter, we experimentally compared the specresponse of a LT–GaAs PC antenna gated with 15 fs lapulses with that of the EO sampling technique using a tEO crystal, by measuring the THz radiation from the saZnTe emitter. As a result, we found that the spectral respoof the PC antenna is mainly determined by the tempoprofile of the photocarrier number injected by the probeam.

We used a mode-locked Ti:sapphire laser as the lsource for the pump and probe measurements of THz ration. We obtained a 12 fs pulse width at a 75 MHz repetiti

a!Present address: Fundamental Research Laboratory, NEC Corpor34 Miyukigaoka, Tsukuba-shi, 305-8501, Japan; electronic ms-kouno@cq.jp.nec.com

8980003-6951/2001/79(7)/898/3/$18.00rticle is copyrighted as indicated in the article. Reuse of AIP content is sub

129.119.67.237 On: Sun,

r-d

aine-

d-e-.

t

eee

to-to

f-

ofase-

lernesel

tia-

rate with an average power of 350 mW, a center wavelenof 800 nm, and a spectral width of 90 nm. The averapump-beam power was 120 mW after an optical chopperthat of the probe beam varied between 1 and 10 mW depeing on the detectors.

The pulse width at the emitter and at the detectors westimated to be about 15 fs by using a different crocorrelation measurement system. The pump beam wascused onto a 100-mm-thick ~110! ZnTe crystal with a silver-coated off-axis parabolic mirror of 50 mm focal lengthgenerate a broadband THz radiation by optical rectificateffect. The generated THz beam was collimated by anaxis parabolic mirror of 76 mm focal length and then focusonto the detectors by another identical parabolic mirror. Tprobe beam after the optical delay line was merged intoTHz-radiation path by a high resistivity silicon wafer inserted between the two off-axis parabolic mirrors. Thenprobe beam and THz radiation were made collinear and wfocused at the same position on the detectors by the laoff-axis parabolic mirror. The detectors used were a dipPC antenna or a 14-mm-thick ~110! ZnTe crystal. They werecarefully adjusted to the focal point of the THz radiatiowithout changing the positions of other optical componen

The PC antenna was a 30mm dipole photoconductiveantenna with a 5mm gap in the center. The metal electrodwere fabricated on 1.5-mm-thick LT–GaAs layer grown at250 °C on a 0.4-mm-thick semi-insulating GaAs substraThe carrier lifetime of the LT–GaAs was estimated to be 1ps by measuring the transient photoreflectance. The antewas reversed so that both the probe beam and THz radiawere incident on the side of the electrodes to avoid distortof the wave form due to dispersion and absorption inGaAs antenna substrate. This antenna geometry also enreplacing the PC antenna with the EO crystal without chaing of the probe beam alignment and the THz radiatpaths. The photocurrent from the PC antenna was preamfied with a low-noise current amplifier and then detectwith a lock-in amplifier referenced to an optical chopperthe pump beam path. The total gain of the detection syswas 1011 V/A.

For the EO detection, we used a 14-mm-thick ~110! ZnTecrystal on a glass substrate. Thep-polarized probe beam wadirected into a quarter wave plate for ap/2-phase shift

on,l:

© 2001 American Institute of Physicsject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

17 Aug 2014 18:21:00

emaaehee

th

-Crti

PCzlinin

io

th

tndhaz

cti-ndion

tion

i--

theetheup-

e

he-

retrumbe-

ue

for

na

899Appl. Phys. Lett., Vol. 79, No. 7, 13 August 2001 Kono, Tani, and Sakai

This a ub to IP:

n,

after passing through the thin EO crystal and then dividinto s- andp-polarization components by a polarizing beasplitter. Both components were detected with an autobanced photoreceiver, and the difference current was preplified with a built-in preamplifier. The power of the probbeam was limited up to 1 mW to avoid saturation of tbalanced receiver. The noise level of the measurem(DJ/J0) was about 3.6331028 which was 44% above theshot noise limit~J050.257 mA!. After the lock-in amplifica-tion of the output voltage of the balanced photoreceiver,total gain was 0.3331010 V/A.

Figure 1~a! shows the signal wave form of the THz radiation from the~110! ZnTe emitter detected with the Pantenna. Very fast oscillations were observed and the shooscillation cycle was estimated to be 0.053 ps, whichabout 25 times shorter than the carrier lifetime of theantenna. Figure 1~b! shows the signal wave form of the THradiation from the same emitter detected by the EO samptechnique. The first four fast oscillation cycles are similarthe respective figures, but the subsequent slow oscillatdiffers from each other.

Figure 2 shows the Fourier transformed spectrum ofPC-detected signal THz wave form shown in Fig. 1~a!. Thefrequency distribution of the detected radiation extendsabout 50 THz. With the ZnTe emitter, the detectable bawidth of the PC antenna was extended two times wider tthe previous report.3 The peaks at 14.3, 19.9, and 25.2 TH

FIG. 1. ~a! Time resolved wave form of THz radiation from~110! ZnTeemitter detected with PC antenna.~b! Time resolved wave form of THzradiation from ~110! ZnTe emitter detected with EO sampling techniqusing 14-mm-thick ~110! ZnTe crystal.

FIG. 2. Fourier transformed spectrum of the PC-detected signal waveshown in Fig. 1~a!.rticle is copyrighted as indicated in the article. Reuse of AIP content is s

129.119.67.237 On: Su

d

l-m-

nt

e

ests

g

ns

e

o-n

are due to the phase-matching condition of the optical refication effect in the ZnTe emitter. The absorption baaround 5 THz is attributed to the TO phonon absorpt~vTO55.3 THz! in the ZnTe emitter.8 The sharp absorptionpeak at 10.56 THz can be due to the two-phonon absorpof the TO phonon in the ZnTe emitter.

The photocurrent from the PC antenna,J(t), is given bythe following equation:9

J~t!5emE2`

`

E~ t !N~ t2t!dt,

wheree is the electron charge andm is the electron mobility.Here,E(t) is the field amplitude of the THz radiation incdent to the PC antenna andN(t) is the number of photocarriers created by the gating probe beam. We suppose thatN(t)is determined by the increase in the number of carriers intime of the full width at the half maximum of the gate pulsand by the decay of the carrier number determined bycarrier lifetime of the antenna substrate material. In this sposition,N(t) is approximated by the following equation:

N~ t !5N0 exp~2t/tc!@ tanh~1.76t/Dt !11#/2,

where the carrier lifetime,tc , was 1.4 ps and the pulswidth, Dt, was 15 fs. The solid curve in Fig. 3~a! shows theFourier transformation ofN(t), represented byN(v), withDt515 fs.

The frequency distribution of the photocurrent from tPC antenna,j (v), is the frequency distribution of the incident THz radiation,e(v), convoluted withN(v):

u j ~v!u}TGaAs~v!ue~v!uuN~v!u.

Here, the Fresnel-transmission coefficientTGaAs(v)52/@11nGaAs(v)# of the GaAs is included, andnGaAs(v) is therefractive index of GaAs in the frequency range of THz.

Figure 3~b! shows the signal spectrum,j (v), shown inFig. 2 divided byTGaAs(v)uN(v)u. We denote this spectrumasePC(v). The spectral components lower than 10 THz areduced while the higher ones are enhanced. The speccomponents higher than 40 THz are enhanced too much

m

FIG. 3. ~a! Frequency distribution of the photocarrier number in PC anteninjected by gating probe pulses. Solid curve representsN(v) when the pulsewidth of the gating probe beam was 15 fs.~b! Fourier transformed spectrumshown in Fig. 2 divided byN(v) whenDt515 fs.

ject to the terms at: http://scitation.aip.org/termsconditions. Downloaded

17 Aug 2014 18:21:00

npr

nw.orthethEc

mar

thsla

th

a-li

eha

nlec

n.

ennatheherthe

fectsfi-t ofspotlso

waslbe

ndam

sizeectHzna

f aEOec-ar-

on,an-ech-e ofinedar-

tsu

y, J.

pl.

Fe

900 Appl. Phys. Lett., Vol. 79, No. 7, 13 August 2001 Kono, Tani, and Sakai

This a

cause the noise components ofj (v) are divided byN(v),which rapidly decreases at high frequencies. The frequecomponents higher than 40 THz cannot be accurately reduced in the present measurement conditions.

The dashed curve in Fig. 4 represents the Fourier traformed spectra of the EO-detected signal wave form shoin Fig. 1~b!. We observed the structures as shown in Figdue to the group velocity mismatch and the phonon abstion lines in the ZnTe emitter. The spectral response ofEO sampling technique is explained by the frequency depdence of the mismatch between the group velocity ofgating probe beam and the THz phase velocity in thecrystal, and by the reflection loss at the EO crystal surfagiven by the following equation:5

G~v!52

nZnTe~v!11

3c~exp$2 i2pvd@ng~l0!2nZnTe~v!#/c%21!

2 i2pvd@ng~l0!2nZnTe~v!#,

whereng(l0) is the group refractive index of the probe beacalculated from the refractive index of ZnTe in the neinfrared region10 andnZnTe(v) is the refractive index of ZnTefrom the far to midinfrared region.11 The crystal thickness,d,was 14mm. The dashed-dotted curve in Fig. 4 showsestimation ofG(v). The cutoff observed at 39 THz waconsistent with the estimation using the above formuwhereG(v)50.

As ePC(v) is considered to be the actual spectrum ofTHz radiation incident on the detectors, we expectedePC(v)filtered by G(v) to reproduce the spectrum of the THz rdiation detected with the EO sampling technique. The socurve in Fig. 4 shows the spectrumePC(v) multiplied byG(v), which well reproduces up to 40 THz the EO-detectFourier transformed spectrum represented by the dascurve. The signal higher than 40 THz in the solid curve isartifact due to the division byN(v). This comparison showsthat the spectral distribution of the PC antenna is maidetermined by the temporal profile of the number of the eltrons injected by the ultrashort-gate optical pulses.

Our comparison shows that the PC-antenna frequeresponse seems independent on its geometrical structure

FIG. 4. Solid curve represents the Fourier transformed spectra shown in3~b! multiplied by G(v). Dashed curve represents the Fourier transformspectrum of the EO-detected signal wave form shown in Fig. 1~b!. Dotted-dashed curve representsG(v).

rticle is copyrighted as indicated in the article. Reuse of AIP content is sub

129.119.67.237 On: Sun,

cyo-

s-n

2p-en-eOe,

-

e

,

e

d

dedn

y-

cyWe

did not clearly observe the resonant peaks due to the antstructure. This may be due to the large radiation loss onantenna structure, especially in the frequency range higthan several THz. Thus, the characteristic impedance ofantenna does not show a strong frequency dependence.12 Be-sides the radiation impedance, the antenna structure afthe collection efficiency of the incident radiation. This efciency is determined by the overlap between the focal spothe THz beam and PC antenna dipole. As the THz beamsize depends on the frequency, the collection efficiency adepends on the frequency of the incident radiation.13 Thefocal size of the gating probe beam at the detectorsestimated to be about 30mm. The area on the EO crystasensitive to the incident radiation was defined by the probeam size and was comparable to the PC-antenna size~30mm!. In our detection geometry, the relative position aoverlapping between the THz beam and gating probe bewere the same for both the detectors. Therefore, theeffect for both the detectors was the same and will not affthe comparison of the spectral response in terms of the Tcollection efficiency. More detailed analysis on the anteneffect will be discussed in another paper.

In conclusion, we compared the spectral response oPC antenna gated with 15 fs laser pulses with that of thesampling technique using a thin EO crystal. After the corrtion considering the temporal profile of the photocreated criers and the group velocity mismatch in the EO detectithe Fourier transformed spectrum obtained by the PCtenna reproduced the one obtained by the EO sampling tnique. This comparison shows that the spectral responsthe PC antenna for broadband detection is mainly determby the temporal behavior of the number of photocreated criers.

The authors thank Hironori Takahashi at HamamaPhotonics K. K. for providing the thin ZnTe crystals.

1S. E. Ralph, S. Perkowitz, N. Katzennellenbogen, and D. GrischkowskOpt. Soc. Am. B11, 2528~1994!.

2P. Gu, M. Tani, K. Sakai, and T.-R. Yang, Appl. Phys. Lett.77, 1798~2000!.

3S. Kono, M. Tani, P. Gu, and K. Sakai, Appl. Phys. Lett.77, 4104~2000!.4Q. Wu and X.-C Zhang, Appl. Phys. Lett.71, 1285~1997!.5A. Leitenstorfer, S. Hunsche, J. Shah, M. C. Nuss, and W. H. Knox, ApPhys. Lett.74, 1516~1999!.

6F. G. Sun, G. A. Wagoner, and X.-C. Zhang, Appl. Phys. Lett.67, 1656~1995!.

7M. Tani, K. Sakai, and H. Mimura, Jpn. J. Appl. Phys., Part 236, L1175~1997!.

8O. Madelung,Semiconductor—Basic Data, 2nd edition~Springer, Berling,1996!.

9S.-G. Park, M. R. Melloch, and A. M. Weiner, Appl. Phys. Lett.73, 3184~1998!.

10D. Palik,Handbook of Optical Constants of Solids~Academic, San Diego,1998!.

11A. Manabe, A. Mitsuishi, and H. Yoshinaga, Jpn. J. Appl. Phys.6, 593~1967!.

12P. R. Smith, D. H. Auston, and M. C. Nuss, IEEE J. Quantum Electron.24,255 ~1988!.

13P. Uhd Jepsen, R. H. Jacobsen, and S. R. Keiding, J. Opt. Soc. Am. B13,2424 ~1996!.

ig.d

ject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

17 Aug 2014 18:21:00