The transition-edge microbolometer (TREMBOL)


Stuart M. Wentworth and Dean P. Neikirk

Department of Electrical and Computer Engineering

The University of Texas at Austin

Austin, TX 78712 USA


ABSTRACT

Operation of detectors at low temperatures has the advantages of lower noise and better thermal conditions. But detector responsivity for conventional bolometers suffers from a drop in the temperature coefficient of resistivity a as the temperature is lowered. The large a for a superconductor at its transition between normal conductivity and superconductivity makes it an ideal detector material for low temperature applications. However, it may be difficult to control temperature to the center of such a narrow transition.

In this paper, the transition-edge microbolometer (TREMBOL) is introduced. Following a brief background over superconducting bolometers, a TREMBOL structure will be presented which uses the low transition temperature (Tc) element lead as the superconductor. The paper concludes with directions for future research.

2. BACKGROUND OF SUPERCONDUCTING BOLOMETERS

Both conventional and composite superconducting bolometers have been fabricated. Conventional-type superconducting bolometers are large area structures where the detector film also acts as the radiation absorber. These have been built using tin[1] and niobium nitride[2]. The first bolometer to actually utilize the superconducting transition was a composite structure using a blackened aluminum foil absorber in conjunction with a tantalum temperature sensor[3]. Another such composite structure makes use of aluminum as the temperature sensor and bismuth as the absorber[4]. Recently, an infrared detector which utilizes a weak-link mechanism rather than a bolometric type mechanism has been fabricated using a bulk high transition temperature superconductor (HTcS)[5]. Other works have focused on the use of superconductors for optical and far-infrared detectors[6] 7-8.

The feasibility of a conventional-type bolometer using the transition of a HTcS film has been discussed by Richards, et al. 9. They obtained estimates of the sensitivity for such a bolometer by carefully measuring the resistance and the low frequency noise in the transition region of several HTcS samples. By assuming that sensitivity is not limited by thermal conductance out of the device, a noise equivalent power (NEP) as low as 10-12 W/was postulated.

Antenna-coupled microbolometers are faster and more sensitive than conventional bolometers, and find use in imaging array applications. The idea of coupling a superconducting microbolometer to a planar integrated antenna was first suggested by Neikirk[10]. Such a structure must perform three functions. It must absorb power from the antenna (i.e. it must provide a matched load to the antenna impedance). It must sense temperature changes due to the absorbed power (i.e. there must be a thermal detector). And finally, it must bias the superconductor to its transition temperature. If the ambient temperature is maintained below the transition temperature, a heater can be used to provide this bias. These three elements are not necessarily separate. In the simplest configuration of the TREMBOL, for instance, the superconducting film would act as both the load and the detector. In this case, the heater must be separate, and most likely would control the overall temperature to bias the entire array at the transition point. Such a TREMBOL is discussed in section 3 using lead as the superconducting material. The feasibility of a HTcS TREMBOL structure with an NEP possibly as low as 2.5 x 10-12 W/has recently been discussed by Hu and Richards[11].

A problem with this simplest configuration of the TREMBOL is that the superconducting thin film resistivity throughout the transition region may be too low to provide a matched load to the antenna. A 4 um x 4 um square of 1000 Å thick YBCuO type superconducting thin film would have a transition region resistance of no higher than about 20 [Omega]. Typical planar antennas, on the other hand, generally have impedances in the 100-300 [Omega] range. One structure which could possibly match to a HTcS resistor is a twin slot antenna[12] which drives the detector via a microstrip network. It is theoretically possible to achieve imbedding impedances as low as 5 [Omega] with these slot antennas. Another way to overcome the matched load problem is to build a composite TREMBOL structure, which is discussed in section 4.

3. THE TREMBOL USING A CONVENTIONAL SUPERCONDUCTOR

To investigate the TREMBOL, a lead detector element coupled to a bow-tie antenna was fabricated. Pb was chosen as the detector material because its transition temperature is above the boiling point of liquid He, and because it is easy to evaporate. Lead bismuthide devices were also attempted, but PbBi forms a discontinuous film unless it is evaporated on a cold substrate (on the order of 77 K). Pb is not without its own problems; our devices oxidized and became useless after about 48 hours of exposure to air.

The Pb devices fabricated were 4.5 um long, 5.0 um wide, and 1000 Å thick. Prior to device fabrication, gold contact pads were electroplated on the glass substrate. The antenna metal was gold, and chrome was used to promote adhesion. The Pb detector/Au antenna devices were fabricated using a bilayer photoresist process which produces a narrow photoresist bridge and a lip in the upper resist layer suitable for liftoff[13]. In this technique, metal for the antenna is evaporated at normal incidence to the substrate, and the narrow photoresist bridge casts a shadow in the detector region. The microbolometer material is then evaporated at an angle to the substrate such that deposition is accomplished under the photoresist bridge. The chip is soaked in acetone to lift off the unwanted metal. Following liftoff, the chip was stored in a desiccator to slow down oxidation of the Pb elements. Immediately prior to testing, the devices were wire-bonded to the printed circuit board used in the cryogenic testing apparatus. Aluminum wire was used, since it bonds quite easily to electroplated gold. Figure 1 shows the placement of the wire bonds to the tested TREMBOL. For maximum accuracy, a 4 point current-voltage test is desired as close as possible to the superconducting element.

Preliminary evaluation of the TREMBOL requires a resistance plot across the transition, and the resistance across the device as a function of power dissipated in the element A bolometer characteristic is that its resistance is a linear function of dissipated power. This property assumes that dR/dT is constant over the temperature range considered, and that the temperature change in the device is proportional to changes in its dissipated power. This leads to an expression for dc responsivity (the response of a detector to a step change in dissipated power)13,



rdc = Ib (dR/dP) .


(1)




Fig. 1: The 4-point measurement setup for the Pb TREMBOL. Aluminum wire is thermosonically bonded to electroplated gold contact pads.




Fig. 2: The resistive transition for the 4.5 um long, 5.0 um wide, 1000 Å thick Pb TREMBOL.


The dc curve of detector resistance R plotted as a function of dissipated power P therefore provides enough information to determine r. The responsivity can also be written



r = Ib R a |Zt|


(2)
where a is the temperature coefficient of resistance, and |Zt| is a thermal impedance term.

Figure 2 shows the Pb TREMBOL's resistive transition. The resistance was calculated based on the slope of the I-V curve from 0 to 1 mA at each temperature tested. Notice that there is a `foot' at the base of the transition. This is most likely `weak-link' type behavior, resulting from finite sized crystals in the small detector. The dR/dT slope in the region represented by the dashed line is 3.6 [Omega]/K.

The resistance of the TREMBOL as a function of power dissipated is shown in Fig. 3 at three temperatures near the transition. If the device is biased at too low a temperature (the curve for 7.46 K), the detector will remain superconducting, and there will be no Joule-heating of the element. Thus, no transition is seen for the 7.46 K plot. At 7.66 K, the device is biased at a high enough temperature to experience the transition. At 7.75 K, which lies just about in the center of the dashed transition region shown in Fig. 2, the dR/dP plot starts out quite steep. It levels off because at sufficient power dissipated, the device is driven beyond the transition region. The parameters of interest for this device are summarized in Table 1. Best dc responsivity of 274 V/W was calculated at 7.75 K. Results for a room temperature bismuth microbolometer are given for comparison13.

In actual operation, the Pb TREMBOL would have a severe problem coupling to the bow-tie antenna since its mid-transition resistance is only about 700 m[Omega]. The following two sections suggest methods for improving coupling of these kind of low resistance detectors to planar antennas.



Temperature

(K)
Bias current

(mA)
dR/dP

(k[Omega]/W)
Zth

(K/mW)
rdc

(V/W)
Pb: 7.46 - - - -
7.66 1.75 138 38 242
7.75 1.0 274 76 274
Bi: 300 1.0 -20 -67 -20
Table 1: Performance of the Pb TREMBOL operated near the middle of the transition regions shown in Fig. 3. dR/dP is the slope in the steepest region of each plot, centered at the bias current given. Also calculated is the thermal impedance Zth, and the responsivity rdc. For comparison, values for a room temperature Bi microbolometer are included13.




Fig. 3: Resistance-power plots at various temperatures near the transition for the Pb TREMBOL.

4. COMPOSITE TREMBOL

One solution to the mismatched load problem is to separate the load from the detector in a composite TREMBOL structure (Fig. 4). This type of device follows from research using tellurium as the detecting element in a room-temperature antenna-coupled composite microbolometer[14]. The load, which is matched to the antenna, may now also serve as the heater by applying a dc bias through the antenna leads. The superconducting detector element is in intimate thermal contact with, but is electrically isolated from, the load/heater element. Changes in heater temperature will be quickly followed by changes in detector temperature. As the detector changes temperature in the transition region, large changes in resistance will occur, which are measured with a current biasing circuit.

An advantage of this device is that the entire substrate can be cooled to below the transition temperature, and the detector can be "turned on" by applying a biasing current through the heater element. This can be quite useful for multiplexing as discussed in section 4.1. The superconducting film may also be used as a low-loss signal line away from the detector, since the entire structure is cooled to below the transition temperature. The device speed is limited by both the thermal mass of the heating element and by the thermal conductance between the heater and the detector. The detectable FIR frequency will be limited by the detector to load capacitance, and therefore the overlap area must be kept small. For a 4 um x 4 um detector separated from the load by 1000 Å of SiO2, capacitance limits operation to below 100 GHz.

Fabrication of the composite TREMBOL involves a choice of superconducting and resistive heater materials. The resistance of the heater material must be known over the temperature range of interest, and the proper dimensions must be chosen to achieve an impedance match with the antenna. Resistive heater candidates include nichrome and bismuth. Superconducting materials include the low Tc superconductors (such as tin, niobium, lead, and lead bismuthide), and the new high transition temperature superconductors. The HTcS materials, primarily YBaCuO compositions and related compounds, were discovered by Bednorz and Müller[15], and improved upon by a host of researchers (see references [16] and [17], and references contained therein). A variety of techniques can be used to deposit thin films of these materials[18]-19[20]. One way to fabricate a composite TREMBOL would be to use a photoresist lift-off technique for patterning the superconducting thin film. Following the deposition, patterning, and high temperature anneal of this film, a thin (1000-2000 Å) layer of silicon nitride or silicon dioxide would be deposited over the superconductor. This layer would serve as the electrical insulator, but is thin enough to be a good thermal conductor. Finally, the antenna/heater would be fabricated by the usual photoresist bridge technique used to make bismuth microbolometers13.

Consideration must also be given to annealing and substrate effects, since some of the more commonly cited substrate materials upon which superconducting thin films are deposited (such as strontium titanate) are not particularly good rf substrates. The more common substrates used in microwave circuits have thus far produced inferior YBaCuO films; for



Fig. 4: Cross section and top view of a composite TREMBOL. Radiation is coupled into the load/heater by the planar antenna structure, causing a rise in the temperature of the load. Due to the intimate thermal contact between the load and the superconducting detector, the resistance of the TREMBOL changes. This resistance change is sensed by the signal line.


instance, sapphire tends to dope the YBaCuO with Al during the high temperature anneal, which seriously degrades the film. Recently, however, a high Tc superconductor was successfully deposited on a LaAlO3 substrate[21], which has desirable rf properties.

4.1 Matrix addressing in a composite TREMBOL array

There are a variety of applications for microbolometer arrays in the FIR spectral region that require imaging, which is the mapping of the radiation intensity of a distributed source[22]. A single detector with mechanically scanned optics may be too slow to build up an image, especially in applications where the required integration time is long (astronomy), or where events occur quickly (plasma diagnostics). Thus, an array of detectors is highly desirable for imaging. A number of one dimensional line arrays have been fabricated using conventional bismuth microbolometers[23]. To image two dimensional objects, however, these line arrays still require mechanical scanning in the direction orthogonal to the array axis.

Two dimensional arrays using conventional microbolometers have proven to be very difficult to design, since it is not possible to matrix address resistors. A unique feature of the composite TREMBOL, however, would allow each individual detector in a 2-D array to be matrix addressed. Figure 5 shows a simple 3 x 4 array where each TREMBOL heater is connected to two address lines (row and column), and the detector element is connected to a signal line. If the entire array is chilled to below the superconducting transition temperature, then without an applied heater bias the detectors will act as electrical shorts. A specific detector may be turned on by setting the appropriate pair of heater address lines high and low. Only the signal generated by the activated detector will be transmitted out the signal line since all the other detectors on the line will be at zero resistance. In actual operation, a clocked cycle could be used to sample data out of one row at a time.

Proper operation of a TREMBOL array would require that each device be biased at its transition temperature. This may be difficult in an array where the heater resistance will vary from device to device. In this case, a look-up table could be used to set the bias across each device's heater to a predetermined optimum value.



Fig. 5: Multiplexing in a composite TREMBOL array. A 3 x 4 array is shown where each device consists of a heater connected to two address lines and a detector connected to a signal line. The device is turned on by applying enough heat to bring the detector to the center of the transition region.

5. DIRECTION OF FUTURE RESEARCH

The TREMBOL (transition-edge microbolometer) and the composite TREMBOL have been introduced as detectors for FIR imaging arrays. The TREMBOL utilizes a superconductor's sharp change in resistance at the normal conduction to superconduction transition. The structure of the composite TREMBOL enables heating of the individual detectors in an array up to their transition temperature, and can thus be used in multiplexing, which would be very advantageous for two-dimensional arrays.

Much work remains, however, to characterize the TREMBOL. A more thorough study of the Pb device is needed, perhaps with the inclusion of a passivating layer to protect the Pb from oxidizing. Other superconductors could be investigated for this application, including the high Tc superconductors.

A composite TREMBOL mask set is presently being designed which will provide for electroplated contacts on both the load/heater level and on the signal line level. It will also contain a number of test structures, including a signal line perpendicular to the bow-tie antenna.

6. ACKNOWLEDGEMENTS

The authors would like to thank Jason Lewis for his helpful suggestions. This project was supported by the National Science Foundation under grant #ECS-8552868.

7. REFERENCES

[1]. C.L. Bertin and K. Rose, "Radiant-Energy Detection by Superconducting Films," J. Appl. Phys., Vol. 39, No. 6, pp. 2561-2568, May 1968.

[2]. D.H. Andrews, R.M. Milton, and W. DeSorbo, "A Fast Superconducting Bolometer," J. Opt. Soc. Am., Vol. 36, No. 9, pp. 518-524, Sep. 1946.

[3]. D.H. Andrews, W.F. Brucksch, Jr., W.T. Ziegler, and E.R. Blanchard, "Attenuated Superconductors: I. For Measuring Infra-Red Radiation," Rev. Sci. Instrum., Vol. 13, pp. 281-292, July 1942.

[4]. J. Clarke, G.I. Hoffer, P.L. Richards, and N.H. Yeh, "Superconductive bolometers for submillimeter wavelengths," J. Appl. Phys., Vol. 48, No. 12, pp. 4865-4879, Dec. 1977.

[5]. M. Scott, R.G. Keefe, and R. Loehman, "High Transition Temperature Superconducting Infrared Detector," IEEE Trans. Electron Dev., Vol. 36, No. 1, pp. 62-65, Jan 1989.

[6]. K. Rose, "Superconductive FIR Detectors," IEEE Transactions on Electron Devices, Vol. ED-27, No. 1, pp. 118-125, Jan. 1980.

7 . J.M. Lewis, "Optical Detectors Based on Superconducting Bi-Sr-Ca-Cu-Oxide Thin Films," B.S. Thesis, Mass. Inst. of Tech., June 1989.

8. P.L. Richards, F. Auracher, and T. Van Duzer, "Millimeter and Submillimeter Wave Detection and Mixing with Superconducting Weak Links," Proc. of the IEEE, pp. 36-45, Jan. 1973.

9. P.L. Richards, J. Clarke, R. Leoni, Ph. Lerch, S. Verghese, M.R. Beasley, T.H. Geballe, R.H. Hammond, P. Rosenthal, and S.R. Spielman, "Feasibility of the high Tc superconducting bolometer," Appl. Phys. Lett., Vol. 54, No. 3, pp. 283-285, 16 Jan 1989.

[10]. D.P. Neikirk, "Integrated detector arrays for high resolution far-infrared imaging," Ph.D. thesis, California Inst. of Technol., pp. 148-149, 1984.

[11]. Q. Hu and P.L. Richards,"Design analysis of a high Tc superconducting microbolometer," Appl. Phys. Lett., Vol. 55, No. 23, pp. 2444-2446, 4 Dec. 1989.

[12]. S.M. Wentworth, R.L. Rogers, J.G. Heston, D.P. Neikirk, and T. Itoh, "A MM Wave/FIR Twin Slot Antenna on a Layered Substrate," to be published in Int. J. Infrared and Millimeter Waves, Vol. 11, No. 2, Feb. 1990.

[13]. D.P. Neikirk, W.W. Lam, and D.B. Rutledge, "Far-Infrared Microbolometer Detectors," Int. J. Infrared and Millimeter Waves, Vol. 5, No. 3, pp. 245-278, Mar 1984.

[14]. S.M. Wentworth and D.P. Neikirk, "Far-Infrared Composite Microbolometers," to be presented at the 1990 IEEE MTT-S International Microwave Symposium, Dallas, May 8-10, 1990.

[15]. J.G. Bednorz and K.A. Müller, "Possible High Tc Superconductivity in the Ba-La-Cu-O System," Zeitschrift für Physik B, Vol. 64, pp. 189-193, 1986.

[16]. M.K. Wu, J.R. Ashburn, C.J. Torng, P.H. Hor, R.L. Meng, L. Gao, A.J. Huang, Y.Q. Wang, and C.W. Chu, "Superconductivity at 93 K in a New Mixed-phase Y-Ba-Cu-O Compound System at Ambient Pressure," Phys. Rev. Lett., Vol. 58, No. 9, pp. 908-910, 2 Mar 1987.

[17]. H. Maeda, Y. Tanaka, M. Fukutomi, and T. Asano, "A New High-Tc Oxide Superconductor without Rare Earth Element," Jap. J. Appl. Phys., Vol. 27, p. L1209, 1988.

[18]. M.S. Osofsky, P. Lubitz, M.Z. Harford, A.K. Singh, S.B. Qadri, E.F. Skelton, W.T. Elam, R.J. Soulen, Jr., W.L. Lechter, and S.A. Wolf, "Thin-film high Tc superconductors prepared by a simple flash evaporation technique," Appl. Phys. Lett., Vol. 53, No. 17, pp. 1663-1664, 24 Oct. 1988.

19. M. Hong, S.H. Liou, J. Kwo, and B.A. Davidson, "Superconducting Y-Ba-Cu-O oxide films by sputtering," Appl. Phys. Lett., Vol. 51, No. 9, pp. 694-696, 31 Aug 1987.

[20]. Z.L. Bao, F.R. Wang, Q.D. Jiang, S.Z. Wang, Z.Y. Ye, K. Wu, C.Y. Li, and D.L. Yin, "YBaCuO superconducting thin films with zero resistance at 84 K by multilayer deposition," Appl. Phys. Lett., Vol. 51, No. 12, pp. 946-947, 21 Sep 1987.

[21]. R.W. Simon, C.E. Platt, A.E. Lee, G.S. Lee, K.P. Daly, M.S. Wire, J.A. Luine, and M. Urbanik, "Low-loss substrate for epitaxial growth of high-temperature superconductor thin films," Appl. Phys, Lett., Vol. 53, No. 26, pp. 2677-2679, 26 Dec. 1988.

[22]. D.P. Neikirk, D.B. Rutledge, and M.S. Muha, "Far-infrared imaging antenna arrays," Appl. Phys. Lett., Vol. 40, No. 3, pp. 203-205, 1 Feb. 1982.

[23]. P.P. Tong, D.P. Neikirk, D. Psaltis, D.B. Rutledge, K. Wagner, and P.E. Young, "Tracking Antenna Arrays for Near-Millimeter Waves," IEEE Trans. Antennas Propagat., Vol. AP-31, No. 3, pp. 512-515, May 1983.