EXHIBIT 1026
`
`I. M. BAKER et al., “Photovoltaic CdHgTe – silicon hybrid focal plane,”
`
`Infrared Technology X, Proceedings of SPIE v. 510, 1984
`
`
`
`
`
`TRW Automotive U.S. LLC: EXHIBIT 1026
`PETITION FOR INTER PARTES REVIEW
`OF U.S. PATENT NUMBER 8,599,001
`IPR2015-00436
`
`

`
`Volume 510
`
`or SPIE-1he International Society for Opt1'cal'Eng1'neen'ng
`
`
`
`
`Infrared Technology X
`/\
`\
`
`Irving J. Spiro, Richard A. Mollicone /
`Chairmenl Editors
`''
`
`Cooperating Organizations
`Optical Sciences Center/ University of Arizona
`Institute of Optics! University of Rochester
`
`
`
`
`August 23-24, 1984
`San Diego. California
`
` A
`
`‘ V 5
`
`1026-001
`
`1026-001
`
`

`
`The papers appearing in this book comprise the proceedings of the meeting mentioned on the cover and title page. They reflect the authors‘
`opinions and are published as presented and without change, in the interests of timely dissemination. Their inclusion In this publication does
`not necessarily constitute endorsement by the editors or by SPIE.
`'
`
`'Please use the following format to cite material from this book:
`Authorls). ‘Title of Paper," /nlrared fechnalogy X, Irving J. Splro. Richard A. Mollicone. Editors. Proc. SPIE 510, page numbers (1985).
`
`ISSN 8755-0180
`ISBN 009252-545—2
`
`Copyright® 1985, The Society of Photo-Optical Instrumentation Engineers. individual readers of this book and nonprofit libraries acting for them
`are freely permitted to make lair use of the material in it, such as to copy an article for use in teaching or research. Permission is granted
`to quote excerpts from articles in this book in scientific or technical works with acknowledgment of the source, including the author‘s name.
`the book name. SPlE volume number. 9998, and year. Reproduction of figures and tables is likewise permitted in other articles and books, pro-
`vided that the some acknowledgment-of-the-source information is printed with them and notification given to SPIE. Republication or system-
`atic or multiple reproduction of any material in this book (including abstracts) is prohibited except with the permission of SPIE and one of the
`authors. In the case of authors who are employees of the United States government. its contractors or grantees, SPIE recognizes the right of the
`United States government to retain a nonexclusive, royalty-free license to use the author's copyrighted article for United States government
`purposes. Address inquiries and notices to Director of Publications, SPIE, P.0. Box 10, Bellingham. WA 98227-0010 USA.
`
`Printed in the United States of America.
`
`.
`
`ii / SP/E Vol. 510 Infrared Technology X (7984)
`
`1 026-002
`
`1026-002
`
`

`
`_This material may be protected by Copyright law (Title 17 US. Code)
`
`Photovoltaic CdHgTe
`
`silicon hybrid focal planes
`-
`I. M. Baker
`
`Mullard Limited, Southampton, Hampshire, U.K.
`
`R. A. Ballingall
`
`R.S.R.E., Malvern, Worcestershire, U.K.
`
`Vabstract
`
`number of array
`Photovoltaic C.M.T.-Si-hybrids have been demonstrated in a
`configurations from 32 x 1 to 64 x 64.
`The photodiode and hybrid interconnection‘
`technology is based on the loophole technique,which has the main advantages of high
`density and therefore good material utilisation in two dimensional arrays.
`For linear
`arrays,
`the CCD multiplexer is suitable for most applications.
`For two—dimensional
`arrays,however,the co—ordinate addressed array may offer significant benefit over the ccq
`particularly at long wavelength.
`32 x 32 arrays of both types have been fabricated.
`A co-ordinate addressed array of ll.4um cut-off produced an NETD of < 0.1K in an
`imaging demonstration.
`
`Introduction
`
`Many future infrared systems will use multiplexed linear or two-dimensional arrays.
`The incorporation of multiplexingcircuitry on the focal plane results in reduced
`encapsulation size and cost.
`It also provides a capability to fabricate and scan large,
`high density arrays for staring applications.
`
`The specialised nature of many applications often calls for a custom design approach,
`particularly with regard to the array format and infrared cut-off wavelength.
`The best
`design philosophy,then,is the hybrid approach where the infrared sensitive material and the
`readout electronics are optimised independently and combined using an appropriate
`interconnect technology.
`The CdHgTe diode array technology described in Section 2 meets
`this requirement,in that it is applicable to a wide range of cut-off wavelengths, array
`formats and silicon multiplexers.
`'
`
`For linear arrays, CCDs are suitable for the whole range of cut-off wavelengths.
`However,
`the situation is less straightforward for two-dimensional arrays,and at longer
`wavelengths the co-ordinate addressed array has many advantages. Section 3 describes the
`issues relating to the choice of two-dimensional multiplexer,and research on two devices,
`a 32 x 32 CCD and RALSA,a random access co-ordinate addressed array,is described subsequenty.
`
`2.
`
`Hybrid technology
`
`‘The
`There are two main approaches to the fabrication of silicon-C.M.T. hybrids.
`most commonly reported (References 1 and 2)
`is the indium bump interconnect. This is a
`useful technique for fairly widely spaced arrays,but there are mechanical problems at
`-interelement pitches of 40pm or less. Also,
`there is some evidence that the mechanical
`stresses involved in the bonding process can impair the performance of long wavelength
`detectors,and in general the extra C.M.T. handling processes are not conducive to long
`material lifetimes (Reference 3).
`32 x 32 hybrids have been fabricated successfully
`using indium bumping as a comparative exercise,but the manufacturability does not compare
`favourably with alternative techniques.
`For certain technologies, however,
`this approach
`is the only practical one.
`»
`
`The diodes
`The second technology is often known as the metal strap interconnect.
`are joined using a metal track which connects the top of each diode to a pad on the
`silicon integrated circuit.
`The general approach is to use the silicon chip as a substrate
`for the manufacture of the diode array and interconnection.
`The C.M.T. handling is
`minimised,and the processes are well established.
`The main technical problem is that the
`interconnection metallisation and the gaps in the C.M.T. monolith next to each diode
`consume area that would otherwise be photosensitive. Therefore the focal plane
`obscuration is much higher than with indium bumped arrays.
`
`SPIE Vol. 5 10 Infrared Technology X I7984) / 127
`
`1026-003
`
`1026-003
`
`

`
`T
`
`n-type junction
`
`passwauon
`
`
`
`cdHgk
`.. .
`'
`9"“?
`'%V
`///// %/-—-Silicon
`metal pad
`
`Figure la
`Formation of planar junction
`
`
`
`F1 9”” lb loophole interconnect
`Hybrid interconnect
`
`for planar junction
`
`/4,
`
`‘ Stages in manufacture ol loophole-interconnected diode array
`
`
`
`Simple loophole diode for close-packed arrays
`
`Figure 1c
`
`Figure 1.
`
`Schematic of hybrid technology‘
`
`A process which has all the advantages of the metal strap interconnect but an
`obscuration of only 5% on a 40um pitched array has been devised.
`The diodes have a unique
`structure and have been named loophole diodes.
`The loophole technology is illustrated
`in Figure l.
`The silicon integrated circuit is prepared with alignment features for the
`CdHgTe monolith and subsequent masking stages.
`The metallisation is a titanium-platinum-
`gold multilayer which gives better resistance to the chemical and mechanical side effects
`of the hybrid process. With reference to Figure la,the CdHgTe monolith is polished to a
`flatness and parallelism of <1um, passivated on both sides and bonded to the silicon chip.
`
`Using the chip as a substrate, a matrix of junctions is ion implanted or diffused into
`the 'p‘-type monolith.
`The junctions are then individually connected to the underlying
`pads on the silicon by cutting loum diameter holes through the junctions, doping the exposed
`CdHgTe in each hole and backfilling with metallisation,
`(Figure lb).
`In high density,
`two—dimensiona1 arrays the first diffusion stage may be omitted and the loophole diffusion
`driven deeper so that photosensitivity is achieved by lateral collection to a cylindrical
`junction (Figure lc). Finally an ohmic contact is made to the 'p' monolith to complete
`the structure.
`
`The main advantage of the loophole technique is that it requires just one masking
`stage to achieve a junction and interconnection to the silicon. Also the structure is
`thermally and mechanically stable, and the connection yield is high, currently 99.8%.
`Research has resulted in a good phenomenological understanding of diffusion in 'p'-type
`CdHgTe for array fabrication.
`
`The Si"91e diff“5i°“ te°h“°1°9Y: such as that illustrated in Figure 1c, is usually
`adopted and in this case,
`the sensitive area is an annulus centred on the circular junction.
`The dimensions of the annulus depend upon the diffusion lengths of minority carriers in
`both the 'p' and 'n' regions.
`In practice these diffusion lengths fall into the range
`5-l0Um for 10-l2um cut—off wavelength material.
`A natural choice of pitch for two-dimension’
`a1 arrays of loopholes is about 40um. At this value both the crosstalk and the dead space
`between elements are minimised.
`Figure 2 shows a
`scanning electron microscope
`photograph of a loum diameter loophole interconnect.
`.
`'
`
`isej———-f——\,——%
`
`
`
`‘ .:€I ; m§. ’ 4n
`
`I22 / SPIE Vol. 510 Infrared Technology X (7984)
`
`
`
`1 026-004
`
`1026-004
`
`

`
`
`
`Figure 2.S.E.M. photograph of loum dia.loophole.
`
`Fig.3 E.B.I.C. mode photograph of p—n
`junction and sensitive area
`
`
`
`Signalarb.
`
`units 100
`
`Distance um
`
`150
`
`Figure 4. HeNe laser spot scan of four adjacent
`
`loophole diodes.
`
`The
`In the 3-5pm band the optimum pitching is somewhat higher, at around Soum.
`spatial sensitivity profile has been studied using_a scanning electron microscope in EBIC
`mode and also a scanned he1ium—neon laser.
`An EBIC photograph of a two-dimensional
`loophole diode is shown in Figure 3. This technique discloses the p—n junction and
`sensitive area.
`The sensitivity profiles of four adjacent diodes as revealed by a scanned
`laser are shown in Figure 4.
`
`A main factor controlling yield in two-dimensional arrays has been the presence of
`sub-grain structure in the CdHgTe material.
`The structure is electrically active and in
`3-Sum arrays gives rise to anomalous crosstalk between adjacent diodes intercepting
`sub-grain boundaries.
`In the 8—l4um arrays the RoA is significantly reduced in diodes
`lying on the structure. Recent progress in the material preparation and diode
`technology has considerably improved the material structure and resulted in improved
`uniformity of the diode properties.
`'
`
`The loophole technology can be applied to both 3-Sum and B—l4um devices:and
`the performance capability is illustrated in Figure 5 in terms of RoA.
`It is seen that
`ROA is a strong function of cut—off wave1ength,and the accurate measurement of wavelength
`is critical.
`
`Values for responsivity, detectivity and quantum efficiency need careful definition
`in view of lateral collection,which can enlarge the actual optical area.
`The definition
`we have adopted is that the area is defined by the 1/e point response to a HeNe laser
`spot scan.‘ A performance summary of a linear array with a cut-off wavelength of l0.0um
`and sensitive area of 55 x 55um is shown in Figure 6.
`‘
`
`SPIE Vol. 610 Infrared Technoloyyx {I984} / 723
`
`1026-005
`
`1026-005
`
`

`
`‘
`
`55 1 55pm
`
`umpv _77K
`12
`
`phogodgogg Q,-my
`2251/a1/no
`
`
`
`
`5
`10
`I5
`20
`Elernentnumbcr
`
`E If”
`
`-
`
`17
`EU
`
`31> ‘I
`
`_ .
`
`
`
`
`
`
`
`25
`
`30
`
`I 3
`—
`-:_I 5
`5 ‘°‘" jjjuu °‘
`$4I
`5 ml
`,0 ‘J-.lIA.L_.__
`
`
`QTVIOBG 7
`
`5
`
`8
`Cut-ofl uavoiendh ‘(pun ‘
`
`/ An. ot 32 x 32 any
`0
`
`‘
`
`
`
`'
`‘we 2509” my’ .350
`
`'50
`
`40°
`
`Figure 5.
`
`RCA v cut-off for CdHgTe
`arrays at 82K and 192K
`
`Figure 6
`
`Performance summary of a 32
`element planar array of l0.0um
`
`3,
`
`.
`
`.
`
`
`Two-dimensional multiplexer design
`a
`CCD or co—ordinate addressed
`
`The multiplexer design tends to fall into two categories:
`device and the co-ordinate addressed array.
`'
`
`the charged coupled
`
`The normal operational
`Most reported work to date has been on the CCD approach.
`procedure is to inject photocurrent
`from the CdHgTe diode into the CCD during a period
`called the stars time T
`. During the rest of the frame time Tf,
`the charge packets are
`sequentially clocked out into a single video output.
`'
`
`The effect that the CCD has on the performance of the system can be simply
`.
`expressed as:—
`’
`
`5
`
`ET
`
`N
`
`D
`
`or
`
`-T—f-
`,5
`
`where NETD is the noise equivalent
`temperature difference.
`
`Tf and TS are defined above
`
`by the photon flux density
`T
`is determined by the output data rate of the CCD,
`1
`and thg charge capacity of each CCD pixel.
`In the 3-Sum bagd the photon flux is low,
`resulting in a low Tf/T ratio, however at longer wavelength the stare time is limited;
`for instance in a typicgl imaging situation at 11pm the stare time is of the order of
`lOus.
`If adequate sensitivity is to be achieved at the longer wavelengths there are two
`requirements. Firstly,
`the CCD clocking rate must be high, say greater than 5MHz,to
`decrease T
`. This in effect restricts the choice of CCD technology to a buried channel
`type and pats severe demands on the external signal processing.
`Secondly, it is desirable
`to incorporate a means to subtract part of the background pedestal charge using circuitry
`within each pixel area to improve Ts. This demands a sophisticated design capability.
`
`724 / S_P/E Va‘/. 510 Infrared TechnoIagyX (1984)
`
`
`
`1 026-006
`
`1026-006
`
`

`
`For
`Another important issue with CCDs is the injection of charge into the CCD.
`efficient and low noise injection the photodiode slope impedance must be greater than the
`input impedance of the CCD. Whilst this is easy to achieve at wavelengths less than
`about 9.5um, it is not so at longer wavelengths.
`
`two dimensional CCD applications tend to be restricted
`For these reasons, at present,
`to wavelengths below 9.Sum. However as the CdHgTe photodiode development and CCD
`V
`development progresses many of these objections to the use of CCDs
`in the 8-l4um band
`will disappear.
`
`The general
`An alternative to CCD readout is co-ordinate addressing (Reference 4).
`approach is to connect
`the diode array to a matrix of MOS switches.
`The diodes are read
`out a row at a time, using a decoder,
`into an off focal plane bank of integrating
`amplifiers.
`
`Lower RoA diodes can be used because the input
`This approach has many advantages.
`impedance of the amplifier is much.lower-than the CCD input impedance. There is no well
`saturation problem and consequently long wavelength operation is practical.
`The uniformity
`correction is easier;
`it is a simple linear correction.
`The technique offers much better
`blooming protection than CCDs,and the technology is much simpler and cheaper.
`
`The principal disadvantage is that T /T »cannot be less than 32 for a 32 x 32 array
`and it is uncompetitive with CCDs in the
`553m band.
`
`The co-ordinate addressed array is therefore more suited for applications in the long
`wavelength band, particularly those involving high flux densities.
`On the other hand the
`main advantages of the CCD are high charge sensitivity and high output data rate,and the
`main application areas than are at short wavelength or for systems requiring fast frame
`rate. There is considerable overlap between the two approaches, however, and other
`features such as the snapshot capability of CCDs or the random access capability of
`co—ordinate addressed arrays may be of overriding importance.
`
`There is a strong argument,then,for researching both technologies as these advantages
`are fairly complementary.
`The co-ordinate addressed array in this programme is called
`RALSA,for random access line scanned array.
`It is currently being developed
`in two
`formats:
`32 x 32 and 64 x 64.
`
`4,
`
`Random access line scanned arrays (RALSA).
`
`The integrated circuit uses standard
`A schematic of RALSA is shown in Figure 7.
`3um NMOS technology.
`The use of decoders enables complete control of the scanning so that,
`for instance,the information from patches of detectors are read more frequently to
`produce more sensitivity. This gives the array an autozoom capability.
`
`The random access function can be used to access individual diodes so that in
`automatic pretesting,
`the array can be fully characterised.
`An example of the computer
`printout for one device is shown in Figure 8. Here the data is presented in three
`formats:
`a spatial distribution, a carpet plot and a probability plot. Automatic cut—off
`wavelength measurements are also made routinely. Measurements such as:
`responsivity,
`noise spectra and detailed I.V. characteristics are performed on a representative number
`of elements across the array.
`
`The off—focal plane current integrators have JFET inputs with an equivalent input
`noise voltage of Bnv Bz'%_ 3 diode resistance of greater than14k9 will ensure that the
`detectivity is degraded by less than 3db.
`
`Imaging experiments have shown that a 32 x 32 RALSA with a cut off wavelength of
`1l.4um and an average R0 of 55kfi, produced a MRTD of less than 0.1K after signal
`conditioning.
`
`SPIE Vol. 510 Infrared Technology X (I .954) / I25
`
`1026-007
`
`1026-007
`
`

`
`I
`
`I-.—————————————— '"'
`Column shit! nzgisler
`
`
`The‘LSA'concepl
`
`ioccl plum
`01!
`charge amplifiers
`and multipicxer
`,:_._.._,..____._..
`
`
`
`
`
`Lincshifl
`
`register
`
`DIAC omput
`
`L.
`
`LSA Output
`
`Figure 7.
`
`Circuit diagram illustrating principle of co—ordinate addressed array
`
`D LKSI TM“
`0
`
`1 ‘WK URI
`iii
`
`Summary of zero bias
`photocurrentin 2" FOV
`
`Summary of zero bias
`resistance
`-
`
`Summary of series
`resistance
`
`
`
`
`2%“
`D LKBC HR!
`0 F$I'Yl5N 3%
`
`lnluuv-flfidhno
`
`B 'fi*.""" ‘"9 08:90.82! l!.H.I::"!'.'.'!
`
`Figure 8.
`
`Computer output of 32 x 32 RALSA showing Io, R0 and 'Rs
`
`726 / SP/E Vol. 5 70 Infrared Technology X (1984)
`
`—1 0
`
`1026-008
`
`1026-008
`
`

`
`5.
`
`‘
`
`CCD hybrid results
`
`The CCD
`Current research is aimed towards both linear and two dimensional CCDs.
`‘technology uses a twin-layer buried channel
`(Reference 5) which results in a high speed -
`charge capacity product,particularly at reduced temperatures. At 77K the measured charge
`capacity is 8 x 10" electrons cm
`for a gate oxide thickness of 8002 and a phase amplitude
`of 12V.
`The transfer efficiency is better than 0.99995 for a loum gate length and a four
`phase clock rate of lMHz.
`
`Photomicrographs of
`The 32 x 32 IRCCD is organised in the parallel serial format.
`the silicon and hybridised device are shown in Figure 9.
`The pixel size is 48 x 4Bum.
`The measured charge capacity of each pixel is 6 x 10‘ electrons at 77K.
`The output
`transfer function is 2 electrons/uV and the circuit is designed to operate at high slew
`rates in order to allow sufficiently long settling time for digitisation of the output.
`Thisli: enhanced by bursting the four phases in loonsecs. and pausing while the output is
`samp e .
`-
`
`f:H
`
`3;5%:
`z:
`.6;
`
`r
`
`Figure 9.
`
`Photomicrographs of 32 x 32 CCD chip and hybrid
`
`RD/(RD + l/gm),where RD is the diode
`The low frequency injection efficiency is given D
`slope resistance. Very high efficiency depends strong y on cut-oi: wavelength, rield of
`View, and diode bias-voltage} however a 32 element linear CCD with l0.lum cut—off diodes
`has been demonstrated with 80% injection efficiency in an f/lfield of view.
`'
`
`For 3-Sum detectors the upper frequency response is limited by the impedance of the
`CCD input at low current levels. At 4um cut Off,f/1 and 77K, a typical photocurrent is
`1.lnA giving an input MOSFET l/gm of 9 x 10°.
`The capacitance of the CdHgTe diode plus
`CCD input was calculated to be 0.43pF for a loophole diode and l.2pF for a 55 x Ssum
`planar diode,giving an upper 3dB frequency response
`(gm/2nC) of 40kHz and 25kHz
`respectively. At Sum cut off,
`the corresponding response is 200kHz and llOkHz.
`Radiometric measurements have been performed on hybridised intermediate wavelength linear
`CCDs.
`The object of the exercise was to determine a figure of merit for the sensitivity
`of each detector element. Rate dependent parameters such as detectivity are not easily
`obtainable in multiplexed detectors since the measured output is an
`integrated and 5amP1ed q“a“t1tY' Appropriate figures of merit are still under
`discussion.
`'
`
`SFIE Vol. 510 Infrared Technology X I I 984/ / I 2 7
`
`1026-009
`
`1026-009
`
`

`
`CCDs_are sensitive devices and there are a large number of phase and bias
`variables that can affect the assessment of signal and noise-
`K
`programmable driving
`and data acquisition system-has been built to give better reproducibility. This approach
`is very flexible and cost effective,in that the reprogramming required for different CCD
`configurations is simple.
`The CCD waveform information and most of the bias voltages
`are input via a microcomputer keyboard,and stored.
`The micro drives an interface that
`is at present capable of speeds of up to 4MHz and phase amplitude of 25.5V.
`The CCD
`output waveform is acquired by a fast sample and hold, and an A-D converter, and is
`transferred to the micro for data processing and display.
`A typical output is shown
`in Figure 10.
`An estimate of the noise is obtained by repetitively digitising a given
`output pulse using a 16 bit A-D converter and calculating the standard deviation of a
`specific number of consecutive samples. At present,
`the results on 3-Sum linear devices
`are consistent with the output being diode noise limited.
`
`The
`A 32 x 32 CCD with 4.5um diodes has been used in an imaging demonstration.
`C.M.T. material was standard bulk grown Brid man
`The array had one defective pixel
`but no other defects.
`‘The average R0 was 10
`:2 in 211 FOV and therefore had near 100%:
`injection efficiency.‘ The array was operated in near reverse bias to ensure a linear
`response. Using simple single point non—uniformity correction, a MRTD of less than
`»0.075K per frame was measured using a thermal target.
`‘
`
`5
`
`E
`9
`"
`5
`
`o 0
`
`I
`
`Average 5
`
`level
`
`D
`
`-
`
`i
`‘
`'5
`9
`
`.
`
`mass
`
`.......w
`0 6...
`.
`Ringo I 0mV
`Diool position
`Voltage on CCD
`1-!
`-
`VDD
`— 22-8
`Davies subs
`o
`-
`-2ra
`awn
`human
`o
`-
`CMTwmmm-103
`‘Sun
`- 22-4
`a ring
`. 25-5
`ssdm.
`.1r2
`saunums- 94
`mwpmu
`- 12-5
`Ru-cam
`- 22-5
`PTGwlao
`hmnwm_
`- o
`Iuucumw -aw:
`mo
`- as
`sunaum - no
`
`.n
`
`§
`'5
`2
`
`.
`
`
`
`.........w
`.36 06,...
`‘‘ R''wr°w""
`
`_
`
`E
`”
`i
`’
`
`°
`
`wamwm
`
`” °mmnm ”
`
`Figure 10
`
`Computer printout of radiometric data from a 32 x 1 intermediate
`wavelength CCD
`
`6.
`
`Off focal plane electronics
`
`In the long wavelength band the electronics required to correct for fixed pattern
`noise in the CCD type of focal plane is more complex than for the co-ordinate addressed
`focal plane. This is due to the non~linearity of the CCD input circuits,which is a
`complex function of exponential
`form.
`The simple linear load line of the co-ordinate
`addressed array,combined with the slower read-out speed,considerab1y simplifies the
`‘signal conditioning.
`
`128 / SPIE Vol. 570 Infrared fochnology X (1984)
`
`__
`
`1026-010
`
`1026-010
`
`

`
`
`
`Using a two point linear correction,high quality imaging has been demonstrated.
`RALSA with a cut-off wavelength of ll.4um was shown to have a NETD of less than 0.1K
`after non—uniformity correction,and has remained stable for three years of trials.
`
`A
`
`The present philosophy in the long wavelength band,then,is to put as much of the
`electronics off the focal plane as possible. This enables commercially available devices
`to be used to a greater extentythus »resulting in a cheaper system that can be updated
`as more advanced components become available.
`
`In the 3—5um band the injection efficiency into the CCD input is so high under normal
`conditions that the CCD response is linear. This, combined with a highly linear CCD
`output circuit, enables non-uniformity correction to be effective,with a simple two point
`linear correction,
`in most
`thermal imaging situations.
`
`Conclusion
`
`A hybrid technology has been described which offers the advantages of good
`manufacturability, simplicity and high performance,particularly for high density
`two-dimensional arrays.
`
`The multiplexer design issues for two—dimensional arrays have been described and a
`lease has been presented that at least two multiplexer types should be considered,
`(the
`CCD and
`the co-ordinate addressed array)
`in View or their complementary advantages;summarised
`in Table 1.
`
`Advantages:-
`
`
`IC)(3 C7
`
`®@@@@@G
`
`3—5pmand 8-12pm
`Lower NETD in 3-Sum band
`Fully multipiexed on chip
`Faster frame rate
`Fem, pad connecuons
`Suitable rorrou applications
`Snapshotcapabflfly
`
`RALSA
`
`(D
`
`LowervR| products at long
`W3VB|°fi9lh3
`® R°"d°'" “C935
`@ LOWBI’ "970 3* '0"9 W"°fl9th
`© Less complex technology
`® tess drlvlne waveforms
`
`Table 1
`
`Comparison between CCD and co-ordinate addressed arrays
`
`Acknowledgments
`
`The work described in this paper was supported by the Ministry of Defence,
`Procurement Executive, DCVD.
`'
`
`The authors wish to acknowledge the contribution of many colleagues at R.S.R.E.
`(Malvern) and Mullard Limited, particularly 1. Blenkinsop of R.S.R.E.
`for the imaging,
`M. D. Jenner and his team for the hybrid fabrication, G. Crimes for the diodes results,
`R. A. Lockett for the RALSA work and J. Parsons for the CCD work, all of Mullard.
`References
`
`‘
`
`Solid State Electronics
`'
`for the 8-14 m band“
`
`Iwasa, S.
`Direct cou lin of 5 micron H CdTe hotovoltaic detector and CCD
`l.
`lexer" Opt.Eng. IE, 533,
`multi
`I§77
`Longo, S. T. et al_ “Infrared focal plane in intrinsic semiconductors“
`2.
`I.R.E.E. In. of Solid State C rcu ts
`- -
`,
`,
`,
`3.
`Baker, I. M.
`"Recombination in CdHgTe Ehotodetectors"
`Vol. 21 pp 1475-1480
`'
`"Electronicall
`4.
`Ballingall, R. A.
`scanned C.M.T. detector arra
`Blec; Letters, Vol. 18, 7, 555-257
`5.
`van Santen
`"Solid state ima e sensors usin
`In. App. Phys. Vol.
`I3
`I977 p333-37$
`
`the char e transfer
`
`rinci
`
`le" JPA,
`
`. SPIE Vol. 5 70 Infrared Technology X (I 984) / I2.9
`
`1026-01 1
`
`1026-011