Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.cplire.ru/html/lab234/pubs/ASC98/eac-02.pdf Дата изменения: Mon Feb 28 21:19:28 2005 Дата индексирования: Tue Oct 2 14:57:36 2012 Кодировка: Поисковые слова: m 11

ASC'98, Palm Desert. Report EAC-02

"This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible"

Two-Stage S-band DC SQUID Amplifier
Georgy V. Prokopenko, Dmitry V. Balashov, Sergey V. Shitov, Valery P. Koshelets
Institute of Radio Engineering and Electronics, Russian Academy of Sciences, Mokhovaya 11, 103907 Moscow, Russia

Jesper Mygind
Department of Physics, Technical University of Denmark, B309, DK-2800 Lyngby, Denmark Abstract A 6 mm x 6 mm chip comprising two identical dc SQUID based amplifiers (SQAs) has been designed, fabricated and tested as a two-stage RF amplifier in a frequency range 3.5-4.0 GHz. Each SQA consists of a double washer type dc SQUID with novel integrated input resonant circuit. The reflection coefficient of both input and output of the SQA has been measured in the two-stage configuration. To avoid SQA saturation at the wide band noise tests a tunable 40 MHz bandpass YIG-filter has been used. The following parameters of the two-stage SQA have been measured at the 3.65 GHz: gain of (17.5±1) dB, 3 dB bandwidth of about 250 MHz, and noise temperature (4.0±0.5) K what corresponds to intrinsic flux noise S1/2 0.6 µ0Hz-1/2 and energy sensitivity i 75 : (7.8 10-33 J/Hz). If the measured RF mismatch between the input/output of SQA and source signal/post-amplifier is taken into account a gain of (20.5±1.5) dB and noise temperature as low as (2.5±1.0) K (intrinsic flux noise S1/2 0.4µ0Hz-1/2, energy sensitivity i 47 :) can be estimated.

I

NTRODUCTION

The dc SQUID based RF amplifier (SQA) is advantageous for integration with a SIS mixer and a flux-flow oscillator (FFO) to complete a fully superconducting submm wave receiver [1]. This is due to its low noise, low power consumption and excellent compatibility with both the SIS mixer and the FFO. It has been shown that RF amplifiers based on dc SQUID can realize power gain up to 20 dB at frequencies of about 100 MHz [2]. Recently the RF amplifiers based on niobium dc SQUIDs with a noise temperature of (3.0±0.7) K at 500 MHz (at an operating temperature of 4.2 K) have been demonstrated; the gain was typically 18 dB [3]. A noise temperature as low as (0.25±0.15) K has been achieved at 1.8 K. This is optimistic result for developing of RF dc SQUID as intermediate frequency amplifier taking into account the linear decreasing of the noise temperature with the bath temperature up to about 0.1 K [4]. We have recently developed an advanced design of SQA for considerably higher frequency of about 4 GHz [5], [6]. This paper presents the recent results for two-stage 4 GHz SQA, which demonstrate its feasibility as an intermediate frequency amplifier for a submm wave integrated receiver [7]. The SQA seems especially attractive for an Imaging Array Receiver. II. SQA DESIGN
AND

attain low noise temperature of the system and efficiently reduce the noise influence of a post-stage HEMT-amplifier. To overcome these problem we have used the two-stage configuration. Two identical single-stage SQAs are connected by the quarter-wave 50 coplanar line which works as an rf impedance transformer between two stages within the frequency band 3.5-4.0 GHz. The design of the interconnecting line allows us to measure SQAs either separately or in the two-stage connection. The equivalent diagram of the two-stage balanced SQA with the novel input resonant circuit is shown in the Fig. 1. The main parameters of the SQA are listed in the Table 1. The single-stage of SQA is consists of the double washer SQUID which has two square holes of the same size formed by a 30 µm wide strip of the top electrode film crossing the washer (60 µm в 150 µm) patterned in the bottom electrode, as shown in Fig. 2. The two parts of the washer are connected in parallel to form a simple gradiometer in order to minimize an external interference. Two shunted micron-size Nb-AlOxNb SIS tunnel junctions are placed at the ends of the 30 µm wide strip; this strip is also used as an integrated control line (± CL) for the magnetic bias of the SQUID. The input coil

T W O - STAGE
C1 FIRST STA G E OF S Q A SIS L R Lowpass f ilt er Contro l l ine + R
sh

SQ A
Input coplanar l ine 50 L
SQA

SIS
CO IL

R C2 R BFQ R

sh

Lowpass f ilt er Contro l l ine -

Int erconnect ion copl an ar lin e 50 S E COND S T AGE OF S Q A O u tput coplanar li n e 50 DC Bia s S QUI D 2 DC Bia s S QUI D 1

MAIN PAR

AMETERS.

The design of the integrated S-ban on the lumped element approach elsewhere [5], [6]. It was shown [5], SQA has relatively low gain 10 dB;

d SQA which is based has been published [6] that a single-stage that is not sufficient to

Manuscript received September 14, 1998 The work was supported in parts by the Russian Program for Basic Research, the Russian SSP "Superconductivity", the Danish Research Academy, and the Danish Natural Science Foundation.

Fig. 1. Equivalent diagram of the two-stage S-band balanced SQA integrated in one chip. It consists of two identical single-stage SQAs connected by quarter-wave length 50 coplanar line.


Inpu t lin e

TABLE 1 MAIN PARAMETERS OF TWO-STAGE SQA #2_4(2) PARAMETRS: 1 2 3 4 5 . . . . . SQUID inductance LSQA Coupling coefficient k2 Area of SIS junction Capacitance of SIS junction Critical current IC MEASURED VALUES L2SQA 70 pH k2 0.6 0.8 - 1.2 µ m2 CSIS 0.1 pF I1C 24 µ A, I2C 26 µ A C 0.1 L 1 Rsh 8 = LSQA/Rsh 10-11 sec L1COIL L2COIL 3 nH C1=C2=C3=C4 1 pF 6 mm в 6 mm fc 3.7 GHz R1d 23 , R2d 21 V1B 37 µ V,V2B 35 µ V G1(max) 10.5±1.0 dB G2(max) 17.5±1.0 dB TN(min) 12±0.5 K TN(min) 4.0±0.5 K S1/2(min) 0.6µ0Hz-1/2 i(min) 75 : (7.810-33 J/Hz) f|3dB 250 MHz L1
SQA

Ou tp u t lin e

Ou tp u t lin e

Fig. 2. Photograph of the central part of the single-stage balanced SQA. The resonant input coil consists of two parts connected in series. The tuning capacitors C1 and C2 (partly seen at the top and the bottom) are connected to the centers of the input coils. The wide wiring film (in the center of the photograph) is connecting the two SQUID loops via the two SIS junctions shunted by Ti strips resistors Rsh. The SQA output lines are connected via coplanar lines (low-pass filters, see Fig. 1).

6. Mc Cumber parameter C 7. Inductance parameter L 8. Shunt resistance Rsh per SIS junction 9. Time constant of SQUID 10. Inductance of input coil LCOIL 11. Input capacitance's C1, ..., C4 12. Size of two-stage SQA chip 13. Central operating frequency fc 14. Dynamic resistance Rd (op. point) 15. Bias voltage at the operational point 16. Power Gain of single-stage SQA 17. Power Gain of two-stage SQA 18. Noise Temperature of single-stage SQA 19. Noise Temperature of two-stage SQA 20. Intrinsic flux noise of two-stage SQA 21. Intrinsic energy sensitivity of two-stage SQA 22. Frequency bandwidth of two-stage SQA

consists of two identical series connected four-turn sections that are positioned inside the corresponding holes in the washer and placed in-plane with the washer. This makes the parasitic capacitance much smaller than the junction capacitance. The planar resonant flux transformer provides efficient coupling over a wide bandwidth (> 5%) with low mutual inductance between the coils. The capacitors C1, C2 are chosen to tune the resonance of the input coil LCOIL at the signal frequency fs 3.7 GHz. The inductance of the SQUID washer is presented by LSQA. The Nb-AlOx-Nb SIS tunnel junctions of the SQUID are shunted by the low inductive resistors, Rsh. The resistors, R, (about 500 each) are used to prevent the leak of the rf signal. Two coplanar low-pass filters with a cut-off frequency of about 50 GHz are used to transmit the dc bias and the signal (fs), but prevent the Josephson current, fJ>>fs, from leaking out of the SQUID. These filters are also transforming the relatively low output impedance of the SQUID (see Table 1) into about 50 . The resistor RBFQ = 0.1-1 is used to prevent the flux trapping in the loop of the output circuit. This specific configuration of the output circuit is designed to cancel the possible signal leakage from the input of the SQA. We call this Balanced Output SQUID Amplifier. The fabrication procedure of the SQA is based on our standard Nb-AlOx-Nb process currently used for production of integrated circuits with micron-size SIS junctions [7], [8]. III. E
XPERIMENTAL

The coolable 3 - 6 GHz HEMT amplifier was situated in the liquid He (30 dB, Tn is about 50 K referred to the SQA output, see Fig. 4). The additional room temperature FET amplifier (25 dB) is used in front of the spectrum analyzer HP-8563A. The combination of a solid state noise source (TNS 2.0105 K at 4.0 GHz [9]) and the precise step attenuator was used to supply the well calibrated signal. It has been found in [5] that effects of saturation and direct detection can change the I-V curve of the SQA. To avoid these effects the tunable 40 MHz band-pass YIG-filter was used. To reduce the influence of the room temperature noise at the input of the SQA, the stainless steel cable was followed by 20 dB attenuator placed at 4.2 K. The losses in the input/output cables were carefully measured by replacing the SQA with an rf connection.

0 40
I VC 's SQ A1

1 00

2 00 40

n
C urre nt ( µA) 20

0

20 O pe ra t io n Po int

n 0/2
0 0 0 1 00 V olt ag e ( µV) 2 00

SET-UP.

The experiments were performed with the sample placed at a holder inside a liquid 4He cryostat shielded by two external µ-metal cans. The typical dc I-V curves of SQA are presented in Fig. 3 at the different dc magnetic bias.

Fig. 3. Typical SQA IVCs at the different dc magnetic bias.


Nois e Te mp eratu re o f SQ A+ HEMT ( K), Nois e Te mp eratu re o f HEMT, (K)

60 50 40 30 20 10 0 3.5
Tin = 6.0 K Tin = 6.75 K Tin = 7.5 K Tin = 11.0 K Tn of HEMT

60 50 40 30 20 10 0 4. 0

S Q A N o is e T e m p e r a t u r e , T n ( K )

25 20 15 10 5 0 3.5
S Q U ID A m p lif ie r (S Q A ) G a in SQA N o is e T e mp eratu re

25 20 15 10 5 0 4 .0

3. 6

3.7

3. 8

3.9

3 .6

3.7

3 .8

3.9

Freq uen cy, f (G Hz )

Fr e q ue nc y , f ( GH z ) Fig.5. Noise temperature and gain of the two-stage SQA measured for the input signal temperature of 6.75 K. The error bar for the data are 0.5 K and 1.0 dB, correspondingly.

Fig.4. Noise Temperature of the system: two-stage SQA + HEMT amplifier, measured at four different values of the input signal. The noise temperature of the system without the SQA is shown by the solid line.

IV. E

XPERIMENTAL

RESULT

S AND

DISCUSSI

ON.

The noise temperature and gain of the system has been evaluated via hot/cold response read by the spectrum analyzer. The value of the background noise applied to the input of SQA (the "cold" signal) was estimated as (5.75± 0.25) K for cold attenuator of 20 dB and the step attenuator set at 110 dB. Several different settings of the step attenuator have been used to check the linearity of the two-stage SQA. The set of curves presenting the formally evaluated noise figure of the system is shown in Fig. 4. These curves were measured at four different settings of the step attenuator 30, 26, 24 and 19 dB that corresponds to the "hot" signal level of 6.0, 6.75, 7.5 and 11.0 K (±0.25 K). The measured noise temperature of the HEMT-amplifier referred to the output of the SQA is also shown in Fig. 4 by a solid line. To extract the gain of the SQA, we used wellknown equation for cascaded amplifiers: GSQA = [(N2-N1)/(Th-Tc)]/[(N2-N1)/(Th-Tc)], (1)

where N1, N2 are the power levels at the output of the system SQA+HEMT that correspond to the input signals of Tc, Th. The second set of values N1, N2 are measured at Tc, Th with the HEMT amplifier only. The gain and the noise temperature of the experimental SQA for Tin= 6.75 K are shown in Fig. 5 as a function of the frequency. The gain up to 17.5 dB and a noise temperature as low as 4.0 K were realized for a two-stage SQA. The 3 dB gain bandwidth of the SQA was estimated as 250 MHz. To realize the ultimate (best possible) performance of the SQA, the reflection at input/output ports of the real device is of great interest. The test signal was applied from synthesizer HP 8673A via directional coupler (-16 dB) [10] which was connected directly to the appropriate port of the SQA. The HEMT amplifier was connected to the "reflection" port of the cold directional coupler. This measuring configuration prevents the possible influence of input noise of the HEMT amplifier (50 K) to the SQUID. The input and output reflection coefficients have been measured for two-stage SQA at three different values of dc bias of the first-stage SQUID and for the same dc bias of the second stage SQUID. The experimental data are shown in Fig. 6.

The experimentally measured input reflection of SQA at the best operation point occurs more than one. In the reference [11] it has been shown that the current J() around the SQUID loop induces a voltage -jMi J() into the input circuit, where Mi - mutual inductance between input coil and SQUID loop. The part of amplified signal is coupling back into the input circuit, and we may measure the back reaction larger than the signal applied to the input of the SQA. In order to realize whether our test procedure is correct, the measurements of the input reflection at the different bias voltages, but with the same Rd have been performed. Two zero-gain bias voltages of 50 µV and 80 µV (closer to the RN region) have been selected. The input reflection coefficient of such "inactive" point with the same Rd was used for the further estimation of the "real" input reflection of the SQA biased at its "active" point. The correction to the SQA input reflection gives a maximum available gain of 20.5 dB and the noise temperature as low as 2.5 K for the two-stage SQA as it is shown in Fig. 7. It should be mentioned that the nonlinearity effect is clearly seen even for rather small input signals (Fig. 4). To investigate the nature of the nonlinearity we applied to the input of SQA the signals from two synthesizers with closely spaced frequencies, both within the frequency band of the

1.0 0.8 0.6 0.4 0.2 0.0 3.5
Reflect ion at t he sam e R d Reflect ion at t he R N Reflect ion at t he O u tp ut

1.0 0.8 0.6 0.4 0.2 0.0 4.0

Re flec tion co eff icient

3.6

3.7

3.8

3.9

Freq uenc y, f ( GH z )

Fig. 6. The reflection coefficients of two-stage of the SQA: reflection at the input in the point with the same Rd as in the operational point, at the input in the region close to RN; reflection at the output.

S Q A Ga in , G ( d B )


S Q A N o i se T e m p er at u r e, T n ( K )

25 20 15 10 5 0 3. 5
S Q UI D Am p l i f i e r ( S Q A ) G a i n S Q A N o i se T e m p er at u r e

25 S Q A G a in , G ( d B ) 20 15 10 5 0 4 .0

16.1 dB and 10.3 dB - for two-stage SQA amplifier. The evaluation of the dynamic range for single- and two-stage SQA is in good quantitative agreement with experimental data. To increase the dynamic range of rf SQA, it seems reasonable to work-out a technique analogous to the FluxLocked Loop (FLL) as it has been demonstrated for dc SQUID-magnetometer in [12]. V. CONCLUSI
ONS.

3 .6

3. 7

3 .8

3. 9

Fr e que nc y , f ( G H z )

Fig. 7. Noise temperature and gain of the two-stage SQA evaluated for the input signal temperature of 6.75 K with taking into account the mismatch between the input/output of SQA and source signal/post-amplifier. The error bars for the data are 1.0 K and 1.5 dB, correspondingly

SQA. We have observed the linear transfer of the spectrum by SQA within its frequency range as shown in Fig. 8. One can see the absence of harmonics at difference frequencies (nfs1 ± mfs2) that point out to linear conversion of spectrum by SQA. The nature of the nonlinearity can be described in terms of the nonlinear flux-to-voltage transfer function similarly to the usual low frequency SQUID used in a magnetometer. The second stage of the SQA (SQA2) seems decreasing the dynamic range of the whole SQA with factor proportional to its gain G2. As it has been shown in reference [2], the dynamic range D of tuned dc SQUID-based rf amplifier can be evaluated as: D Q 0 /LkB(T+TN),
2 2

The experimental study of a two-stage S-band amplifier based on dc SQUID have been performed in order to estimate its feasibility for use as an IF amplifier in a low noise SIS receiver. The study has shown encouraging results with the good consistency of design and measured data. The following parameters of the real two-stage SQA have been measured: noise temperature about 4.0 K, gain 17.5 dB and 3 dB bandwidth of about 250 MHz at center frequency of 3.65 GHz. The results of this preliminary study seem quite encouraging for future investigations of the SQA for integration with a SIS mixer and FFO in a fully superconducting sub-mm wave receiver. REFEREN
CES.

[1] V.P. Koshelets, S.V. Shitov, L.V. Filippenko, A.M. Baryshev, A.V. Shchukin, G.V. Prokopenko et al, "Integrated Submillimeter Heterodyne Receivers", Proc. 30th ESLAB Symp., ESTEC, Noordwijk, The Netherlands, ESA SP-388, pp. 193-202, 1996. [2] C. Hilbert and J. Clarke, "DC SQUIDs as radio frequency amplifiers", J. Low Temp. Phys.., vol. 61, pp. 263-280, 1985. [3] M. Muck, J. Gail, C. Heiden, M.-O. Andre and J. Clarke, "Low Noise Radio Frequency Amplifiers Based On Niobium dc SQUIDs with Microstrip Input Coupling", Abstracts of 4th Twente HTS Workshop, p.11, 6-7 May 1998. [4] F.C. Wellstood, C. Urbina and J. Clarke, "Hot-electrons effects in metals", Phys. Rev., B49, pp.5942-5955, 1994. [5] G.V. Prokopenko, S.V. Shitov, V.P. Koshelets, D.V. Balashov, J. Mygind, "A dc SQUID based low noise 4 GHz amplifier", IEEE Trans. on Appl. Supercond., vol. 7, pp. 3496-3499, 1997. [6] G.V. Prokopenko, V.P. Koshelets, S.V. Shitov, D.V. Balashov, J. Mygind, "Low noise and wide-band 4 GHz amplifier based on dc SQUID", Inst. Phys. Conf., Ser.No 158, pp. 731-735, 1997. [7] V.P. Koshelets, S.V. Shitov, A.M. Baryshev, I.L. Lapitskaya, L.V. Filippenko, H. van de Stadt, J. Mess, H. Schaeffer, T. de Graauw, "Integrated Sub-MM Wave Receivers", IEEE Trans. on Appl. Supercond., vol. 5, pp. 3057-3060, 1995. V.P. Koshelets, S.V. Shitov, L.V. Filippenko, A.M. Baryshev, W. Luinge, H. Golstein, H. van de Stadt, J.-R. Gao, T. de Graauw, "An Integrated 500 GHz Receiver with Superconducting Local Oscillator", IEEE Trans. on Appl. Supercond., vol.7, pp. 3589-3592, 1997. [8] V.P. Koshelets, S.A. Kovtonyuk, I.L. Serpuchenko, L.V. Filippenko, and A.V. Shchukin, "High Quality Nb-AlOx-Nb Tunnel Junctions for Microwave and SFQ Logic Devices", IEEE Transactions on Magnetics, vol. 27, pp. 3141-3144, 1991. [9 ] Noise Com, NC 3208-A, SN1409 8732. [10] NARDA, 2-18 GHz MOD 4203-16. [11] J.M.Martinis and J. Clarke, "Signal and Noise Theory for a DC SQUID Amplifier", J. Low Temp. Phys.., vol. 61, pp. 227-236, 1985. [12] D. Drung, R. Cantor, M. Peters, H.J. Scheer, and H. Koch, "Low-noise high speed dc superconducting quantum interference device magnetometer with simplified feedback electronics", Appl. Phys. Lett. vol. 57, pp. 406408, 1990.

(2)

where 0 is the maximum rms. flux that can be applied to the SQUID without introducing noticeable nonlinearity, is typically of the order of 0.1 for a single-stage amplifier and 10-1/(G1)1/2 for two-stage amplifier; Q is the quality factor, L is inductance of the SQUID, T is the physical temperature, TN is the noise temperature. We estimated the dynamic range from this equation for a single-stage SQA is equal about

-72 -76 -80 -84 -88 -92 -50 -40 -30 -20 -10

-72

Spectrum transfer of SQA

Output Power, (dBm)

-76 -80 -84 -88 -92 50

0

10

20

30

40

Central frequency 3.65 GHz, SPAN 100 kHz

Fig.8. Transfer of the Spectrum by SQA.