DOI 10.1007/s00340-008-3332-8
Dependence of terahertz radiation on gap sizes of biased
multi-energy arsenic-ion-implanted and semi-insulating GaAs
antennas
T.-A. Liu· R.-H. Chou · C.-L. Pan
Received: 30 October 2008 / Published online: 11 December 2008 © Springer-Verlag 2008
Abstract We investigate the characteristics of terahertz
radiation pulses using biased multi-energy arsenic-ion-implanted and semi-insulating GaAs photoconductive an-tennas with different gap sizes in terahertz time-domain spectroscopy. At a specific fluence excitation, with increas-ing antenna gap size, the absolute values of the (peak) normalized terahertz waveform minimum (valley), as well as the bandwidth, reveal an increasing trend for multi-energy arsenic-ion-implanted GaAs antennas and a decreas-ing trend for semi-insulatdecreas-ing GaAs antennas. We find that the largest reachable bias fields applied to arsenic-ion-implanted GaAs antennas are higher than those applied to semi-insulating GaAs antennas. On the basis of pump fluence dependences of peak terahertz amplitude, we de-duce that multi-energy arsenic-ion-implanted GaAs anten-nas have the ability to acquire higher THz power at even higher pump fluence in comparison with semi-insulating GaAs antennas.
PACS 85.60.Bt· 85.60.Gz · 73.50.Pz · 61.72.Uj ·
85.30.-z· 77.22.Jp · 78.47.Cd · 07.57.Hm
T.-A. Liu (
)Center for Measurement Standards, Industrial Technology Research Institute, 321, Sec. 2, Kuang Fu Road, Hsinchu 300, Taiwan
e-mail:[email protected] R.-H. Chou· C.-L. Pan
Institute of Electro-Optical Engineering, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu 300, Taiwan R.-H. Chou
e-mail:[email protected] C.-L. Pan
e-mail:[email protected]
1 Introduction
Terahertz (THz) radiation pulses from biased photoconduc-tive (PC) antennas have been widely used in various disci-plines including chemical, biological, imaging, and telecom-munications [1–4]. The emission characteristics of THz ra-diation are associated with the pump conditions or device structures that we adopt. Hence, the aim of most researchers is to discover an optimum operation for efficient THz radia-tion generaradia-tion.
Under fixed pump conditions, the PC antenna factors in-volving geometries, gap sizes G, and photoconductive ma-terials, will significantly affect the THz emission features of radiation power, peak width, and spectrum. In terms of G, PC antennas with a G larger than 5 mm are able to pro-vide giant THz radiation power [5]. In practical employ-ment, however, such larger-aperture antennas require heavy equipment leading to a confined movement range for the an-tennas. To overcome this disadvantage, one can utilize mid-aperture antennas with a G ranging from 0.1 to 1 mm, since it provides impetus to generate THz radiation in a compact environment.
Another antenna factor that determines the emission per-formance of an antenna is its PC materials. From the point of view of electrical properties, PC material with high resis-tivity is preferred as a substrate, since it can have a large bias applied to it and thus obtain high THz power. To reach this goal, liquid-encapsulated Czochralski-grown or low temper-ature molecular-beam-epitaxy grown materials have been used during the production of GaAs-based antennas. The former and latter materials are known as semi-insulating GaAs (SI-GaAs) [6] and LT-GaAs. In the optical pump– probe differential reflection measurement, the carrier life-time of a SI-GaAs based antenna usually has the value of over several picoseconds, while a LT-GaAs based antenna
possesses a carrier lifetime shorter than one picosecond [7], so it can provide a broader emission spectrum than a SI-GaAs antenna [8]. Nevertheless, the drawback of a LT-GaAs antenna is that it is difficult to be reproduced even under the same experimental conditions, except for a specific ap-proach [9]. On the other hand, multi-energy arsenic-ion im-plantation in GaAs (multi-GaAs:As+) will result in a highly nonstoichiometric disturbance and lead to material charac-teristics similar to LT-GaAs. During the fabrication of a multi-GaAs:As+antenna, the ion dose can be precisely con-trolled, implying that such a device has the reproducible property [10]. Moreover, a multi-GaAs:As+ antenna has a high breakdown voltage and a good carrier mobility, and therefore has replaced the LT-GaAs antenna as a promising candidate for a THz radiation source [11–13].
If the antenna factors are fixed at specific conditions, the variations of pump fluence F and nominal bias field Eb ap-plied to the antennas will have a profound influence on THz emission features as well. Experimentally, one usually ob-serves from pump fluence dependence that the peak THz amplitude grows first at low fluence, while it saturates at a high F [14,15]. From bias-field dependence, one can ob-serve that the peak THz amplitude, from a large-aperture antenna, scales linearly with the bias field [16]. Theoreti-cally, both types of dependences can be simply described by the scaling rule [17]. Besides the influence of pump flu-ence and bias field, Ralph and Grischkowsky found that the THz field magnitude will be strongly enhanced if the illu-mination position of the laser spot approaches the anode of a 80-µm-gap SI-GaAs antenna [18]. This enhanced THz ra-diation, obtained by asymmetrical illumination, results from the extremely large bias field near the anode, due to material defects.
In the present contribution we report on the characteris-tics of THz radiation pulses from multi-GaAs:As+antennas with varying G, and analyze several essential emission char-acteristics including the absolute value of the THz waveform minimum (valley), bandwidth, bias field, and pump fluence dependence of peak THz amplitude. The performances of multi-GaAs:As+ antennas are compared to those obtained from similar devices made of SI-GaAs.
2 Experiments
Our experiments were carried out by employing THz time-domain spectroscopy. We made use of two types of materi-als to construct THz radiation emitters. One type was fabri-cated on multi-GaAs:As+, and the other type was fabricated on highly resistive SI-GaAs. For each type of antenna, we chose five samples with gap sizes G of 0.02, 0.1, 0.2, 0.5, and 1 mm. The antenna consisted of two coplanar strip lines with Au metallic coating layers of 300 nm as illustrated in Fig.1. The carrier lifetimes for both materials were obtained from the optical pump-probe differential reflection measure-ment.
The experimental setup for the generation and measure-ment of THz radiation was similar to the common free-space electro-optical (EO) detection from the emitter of biased PC antennas [19]. The pump or probe pulses, with an 800-nm wavelength and 130-fs pulse width, were also provided by the mode-locked Ti:sapphire laser, operating at the repeti-tion rate of 85 MHz and with the pump power of 500 mW. The laser excitation positions are depicted in Fig.1. The pump spot size incident on the sample was fixed to ap-proximately 0.1-mm diameter. In the cases of G= 0.02 and 0.1 mm, the spot positions were central, with respect to the gap, so that both illuminations were nearly uniform. In the cases of G= 0.2, 0.5, and 1 mm, the position of the laser spot was located near the anode in order to obtain an op-timum THz radiation. The generated THz radiation beam was thus collimated and focused onto the EO sensor of ZnTe with a thickness of 1.5 mm, by a pair of off-axis paraboloidal mirrors. Since the THz radiation was collected by parabolic mirrors in this experiment, the temporal resolution was lim-ited by group-velocity mismatch between the optical probe beam and THz radiation. The other coherent polarized probe beam was collimated onto the EO crystal and the polariza-tion was modified because of the modified refractive index from the THz field. The probe beam was then transmitted through the analyzer and we sampled the THz field from the time delay between the pump and probe beams.
Fig. 1 Schematic representation of DC-biased PC antennas excited by femtosecond laser pulses. L and W are equal to 10 mm and 0.1 mm, re-spectively. The illuminated regions are indicated by red circular spots.
The left- and right-hand diagrams illustrate the uniform illumination for gap spacing G= 0.02 or 0.1 mm, and the asymmetric illumination for G= 0.2, 0.5, or 1 mm, respectively
3 Results and discussion
3.1 Waveforms and spectra
Figure 2 shows the normalized photoreflectance changes
R/Rof multi-GaAs:As+and SI-GaAs based antennas as
Fig. 2 Transient photoreflectance changes R/R for multi-GaAs:As+(open blue circles) and SI-GaAs (solid red circles)
a function of the time delay t . Fitting these data to exponen-tial decays yields the results that the carrier lifetimes of the multi-GaAs:As+and SI-GaAs antennas are 0.8 and 2.5 ps, respectively. Obviously, the carrier lifetime of each antenna is much longer than the pump pulse duration (=0.05 ps).
Figure3shows the normalized THz radiation waveforms
Er and their respective Fourier-transformed amplitude ˜Er spectra from multi-GaAs:As+ and SI-GaAs antennas with four different values of G at a pump fluence of 70 µJ/cm2. To control nominal bias field strengths of 3.5 kV/cm for pre-venting the risk of antenna damage, the bias voltages applied to the anode were 7, 35, 70, and 175 V for G= 0.02, 0.1, 0.2, and 0.5 mm, respectively. From Figs.3a and c, it is seen that each waveform consists of a dominant peak and a minor valley followed by a slowly varying tail. In the time domain, the analysis for these waveforms can be characterized by the absolute value|Eminr | of the THz waveform minimum, the peak and valley full width, and the negative tails. In the fre-quency domain, we focus on the spectrum bandwidth f .
From each waveform, we can clearly identify that the peak width is larger than the valley width, under different values of G. Quantitatively, the value of the peak width is about 0.33 ps less than that reported by Katzenellenbogen
Fig. 3 THz radiation waveforms Er(t )as a function of time delays t for (a) multi-GaAs:As+and (c) SI-GaAs antennas with gap size G of 0.02
Fig. 4 Bandwidth f (square marks) and absolute value |Emin
r | (point marks) of THz
waveform minimum as a function of gap size G for multi-GaAs:As+(blue) and SI-GaAs (red) antennas
and Grischkowsky (∼0.38 ps) [20]. Furthermore, for dif-ferent values of G, the values of the valley width are also different and range irregularly between 0.30 and 0.27 ps without any explicit gap-size dependence. We also observe from each waveform that a relatively long negative tail with the duration of about 3.7 ps follows the valley and is quite obvious in contrast to those found in other similar experi-ments [8].
From Fig.3, we obtain the dependences of |Emin r | and
f on G and illustrate them in Fig.4. As can be seen in the case of the multi-GaAs:As+ antenna, as G increases from 0.02 mm to 0.5 mm, f increases from 1.2 to 1.6 THz, and
|Emin
r | also increases from 0.44 to 0.62 in a strongly non-linear way, satisfying a saturation function with the form
|Emin
r | = (0.45G0.198+ 0.24)/(1 + 0.14G), which is also plotted in Fig.4. On the contrary, in the case of the SI-GaAs antenna, |Ermin| decreases linearly from 0.64 to 0.47, and
f also decreases monotonically from 1.77 to 1.41 THz. Obviously, the curves of f versus G vary in a manner similar to those of |Ermin| versus G for both kinds of PC materials. What is especially interesting is that the quantity
f of multi-GaAs:As+antennas is smaller for G= 0.02 or 0.1 mm, but becomes larger and the spectrum thereby ex-tends to beyond 4.2 THz provided that G= 0.2 or 0.5 mm. Accordingly, we infer that under the condition of a large enough gap size the THz radiation from a multi-GaAs:As+ antenna will probably have a larger f than that from a SI-GaAs antenna. Indeed, this deduction has been assured to be valid for our previously used large-aperture GaAs:As+ an-tennas where the numerical simulations confirmed that the larger refractive index and larger absorption coefficient were responsible for the larger f [21].
The results of Fig.4also point out that the values of f are much larger than those produced with a large-aperture SI-GaAs antenna (f ∼ 0.27 THz) [5], as well as our previ-ously used small-aperture SI-GaAs or multi-GaAs:As+ an-tenna (f ∼ 1 THz) [12]. This larger f is a reflection of a
narrow negative peak and a large|Eminr |, and it is just favor-able for applications such as THz medical or security imag-ing requirimag-ing broadband THz radiation.
In the introduction, we mentioned that the bias field Eb, near the anode of a SI-GaAs or multi-GaAs:As+ antenna is extremely large, so we can make use of this property to acquire a maximum THz power output by focusing a fem-tosecond laser pulse on that area. For the sake of reaching the same nominal Eb, an antenna with a larger G is applied with a higher bias voltage, leading to a higher Ebnear the anode. As a result, this higher Ebsignificantly increases the space-charge-field screening effect accompanying the larger
f and|Ermin| [22]. This explains why f and|Ermin| of a multi-GaAs:As+antenna increase with the increase of G, but the mechanism leading to the decreasing trend in|Eminr | of SI-GaAs antennas is unknown to us.
3.2 Bias field and pump fluence dependence
Next, we study how the peak THz amplitude Ermax varies with the bias field Eband pump fluence F using different PC materials and varying the G factor. Figure5a plots the Eb dependences of Ermax from SI-GaAs and multi-GaAs:As+ antennas with three different values for G. As seen in this figure, the data reveals linear variations in the case of G= 0.02 mm. In the cases of G= 0.5 and 1 mm, it is worth not-ing that nonlinear behaviors occur at both ends of the curves together with linear relationships appearing in the middle parts of the curves. Such nonlinear behaviors become more obvious in the case of multi-GaAs:As+ antennas between 6 and 9 kV/cm, where the amplitude reduction is probably owing to the heating effect within the antennas. This im-plies that an even higher THz radiation power may be ac-quired under good cooling apparatus in the pumping area of the sample. Besides, the multi-GaAs:As+ antenna has the advantage of enduring a nominal bias field of as high as
Fig. 5 (a) Bias field Eb and (b) pump fluence F dependence on
peak THz amplitude Emax
r for multi-GaAs:As+(blue) and SI-GaAs
(red) antennas with gap size G of 0.02 (square marks), 0.5
(trian-gle marks), and 1 mm (point marks). The pump fluence F is fixed at 70 µJ/cm2in (a), and the nominal bias field Ebis kept at 3.5 kV/cm
in (b)
8 kV/cm without causing any electrical damage to itself un-der photoexcitation. By contrast, the bias field applied to the SI-GaAs antenna has to be constrained below the threshold of 4 kV/cm to prevent electrical damage to the device.
The F dependences of Emaxr are plotted in Fig.5b where
Ermax increases monotonically with the increase of F for each material and G. We also found that the antenna with a larger G possesses a higher Ermaxfor each material, and
Ermaxof a multi-GaAs:As+ antenna is lower than that of a SI-GaAs antenna in the measurable range (i.e. 0–70 µJ/cm2) with the same G. Nevertheless, we perceive that Ermax from SI-GaAs antennas reveals slightly saturated trends at a high F , leading to a reverse consequence that if F is over 70 µJ/cm2Ermaxfrom multi-GaAs:As+antennas will be ex-pected to be higher than the SI-GaAs cases, according to the experimental data using the scaling rule [17].
In the following, we employ the scaling rule to interpret the observations in Fig.5. It is known that the scaling rule relates Ermaxto both Eband F by the formula
Ermax≈ D F Fs+ F , (1) where D= A √ nLEb 4π ε0c2η 0τdz , (2) Fs= (1+ nL)hν q(1− R)μeη0 . (3)
In this notation, nL is the refractive index, μe is the elec-tron mobility, A refers to the pump spot area, and hν is the photon energy.
As described in the aforementioned experimental con-ditions, the diameter of our pump spot area stays at about
0.1 mm, corresponding to an area of 7.8× 10−3mm2. For the cases of G= 0.02, 0.5, and 1 mm, the values of A are equal to the gap area 3.1× 10−4mm2 in the former case, but equal to the pump spot size 7.8× 10−3mm2in the lat-ter two cases as indicated in Fig.1. According to (1) and (2), Ermax of a 0.02-mm-gap antenna is proportional to A, and thus has a lower value either in the bias or pump fluence dependence since its value of A is smaller than the other two cases. As a matter of fact, although the latter two cases have the same value of A, the value of Ermaxof a 1-mm-gap antenna is higher than that of a 0.5-mm-gap antenna rather than having the same value of Emaxr . The reason is because the strength of the trap-enhanced bias field Ebnear the an-ode depends on G. As we have mentioned in subsection A, if one applies the same nominal bias field to each antenna, the larger G, the higher Ebnear the anode. Thus, Ermaxof a 1-mm-gap antenna with a higher Ebis higher than that of a 0.5-mm-gap antenna with a lower Ebaccording to (2).
In (1) and (2), nLand μeshow involvement with the op-tical and electrical properties of PC materials. Due to ion implantation, the values of nL for a multi-GaAs:As+ an-tenna are larger than those of a SI-GaAs anan-tenna, and the effective μe of a multi-GaAs:As+ antenna is estimated to be about 1500 cm2/V s less than that of a SI-GaAs antenna (μe∼ 5000 cm2/V s). According to (3), a larger nL and a smaller μecorrespond to larger D and Fs, leading to Emax
r of a multi-GaAs:As+ antenna being lower than a specific
F but becoming higher than a specific F at the same G. For G= 0.5 mm, the fit to the experimental data of Fig.5b yields D= 15.6, Fs= 870.5 µJ/cm2for a multi-GaAs:As+ antenna and D= 10.2, Fs= 550.5 µJ/cm2for a SI-GaAs an-tenna. According to these two fits, it can be seen that Ermax of a 0.5-mm-gap multi-GaAs:As+antenna is lower than that
of a SI-GaAs antenna if F < 70 µJ/cm2, but becomes higher if F > 70 µJ/cm2. Similar phenomena have also been found in our large-aperture GaAs:As+ antennas [21], and hence we believe that the mid-aperture multi-GaAs:As+antennas will benefit from high pump fluence as well.
4 Conclusions
We have presented a study of gap-size-dependent phenom-ena on THz radiation. We used several multi-GaAs:As+and SI-GaAs mid-aperture antennas with different bias voltages and gap sizes to observe their bipolar THz waveforms in an optical pump–probe experiment. At fixed pump fluence and nominal bias field, with the increase of gap size of the anten-nas, we observe that both the bandwidth of the THz radiation and the absolute value of the THz waveform minimum de-crease linearly for SI-GaAs antennas, whereas they inde-crease monotonically for multi-GaAs:As+antennas, resulting in a consequence that the bandwidth from a multi-GaAs:As+ an-tenna is larger than that from a SI-GaAs anan-tenna at a large gap size. The bandwidths of the THz radiations from our an-tennas are considerably larger compared to other small- or large-aperture antennas. In the dependences of bias field and pump fluence on the peak THz amplitude, the measured data and associated theoretical prediction curves indicate that the multi-GaAs:As+antenna benefits from its higher reachable bias fields, and can generate even higher THz power, prob-ably at high pump fluence, in comparison with a SI-GaAs antenna.
Acknowledgements This work was partially supported by the Aca-demic Top Universities Program of the Ministry of Education and var-ious grants of the National Science Council of Taiwan, Republic of China.
References
1. H.M. Pickett, J. Mol. Spectrosc. 148, 371 (1991)
2. T.R. Globus, D.L. Woolard, T. Khromova, T.W. Crowe, M. Bykhovskaia, B.L. Gelmont, Hesler, A.C. Samuels, J. Biol. Phys. 29, 89 (2003)
3. D.M. Mittleman, M. Gupta, R. Neelamani, R. Baraniuk, J.V. Rudd, M. Koch, Appl. Phys. B: Photophys. Laser Chem. 68, 1085 (1999)
4. T.-A. Liu, G.-R. Lin, Y.-C. Chang, C.-L. Pan, Opt. Express 13, 10416 (2005)
5. T. Hattori, K. Tukamoto, H. Nakatsuka, Jpn. J. Appl. Phys. 40, 4907 (2001)
6. M.S. Markram-Ebied, in Semi-insulating III–V Materials, ed. by D.C. Look, J.S. Blakemore (Shiva, Nantwich, 1984)
7. F.W. Smith, H.Q. Le, V. Diaduk, M.A. Hollis, A.R. Calawa, S. Gupta, M. Frankel, D.R. Dykaar, G.A. Mourou, T.Y. Hsiang, Appl. Phys. Lett. 54, 890 (1989)
8. M. Tani, S. Matsuura, K. Sakai, S. Nakashima, Appl. Opt. 36, 7853 (1997)
9. K.G. Eyink, M.A. Capano, S.D. Walck, T.W. Haas, B.G. Street-man, J. Electron. Mater. 26, 391 (1997)
10. A. Hussain, S.R. Andrews, Appl. Phys. Lett. 88, 143514 (2006) 11. J. Lloyd-Hughes, E. Castro-Camus, M.D. Fraser, C. Jagadish,
M.B. Johnston, Phys. Rev. B 70, 235330 (2004)
12. T.-A. Liu, M. Tani, C.-L. Pan, J. Appl. Phys. 93, 2996 (2003) 13. T.-A. Liu, G.-R. Lin, Y.-C. Lee, S.-C. Wang, M. Tani, H.-H. Wu,
C.-L. Pan, J. Appl. Phys. 98, 013711 (2005)
14. J.T. Darrow, X.-C. Zhang, D.H. Auston, Appl. Phys. Lett. 58, 25 (1990)
15. J.T. Darrow, X.-C. Zhang, D.H. Auston, J.D. Morse, IEEE J. Quantum Electron. 28, 1607 (1992)
16. P.K. Benicewicz, A.J. Taylor, Opt. Lett. 18, 1332 (1993) 17. P.K. Benicewicz, J.P. Roberts, A.J. Taylor, J. Opt. Soc. Am. B 11,
2533 (1994)
18. S.E. Ralph, D. Grischkowsky, Appl. Phys. Lett. 59, 1972 (1991) 19. Q. Wu, X.-C. Zhang, Appl. Phys. Lett. 67, 3523 (1995)
20. N. Katzenellenbogen, D. Grischkowsky, Appl. Phys. Lett. 58, 222 (1991)
21. R.-H. Chou, T.-A. Liu, C.-L. Pan, J. Appl. Phys. 104, 053121 (2008)
22. R.-H. Chou, T.-A. Liu, C.-L. Pan, Jpn. J. Appl. Phys. 47, 8419 (2008)
