United States Patent 1191
`Cox et al.
`
`Patent Number:
`[11]
`[45] Date of Patent:
`
`5,043,582
`Aug. 27, 1991
`
`[54] X-RAY IMAGING SYSTEM AND SOLID
`STATE DETECTOR THEREFOR
`
`4,672,454 6/l987 Cannella et a1‘ ......... .. 250/3701] X
`4.675,?30 6/1987 Catchpole et a1. .... ..
`358/21311
`
`[75] Inventors: John D. Cox; Alan M. Jacobs;
`Stephen A. Scott; Yi-Shung Juang, all
`
`4,807,000 ,2/l989 Gurnee el al. . 1 . . . . .
`
`4,810,881 5/1989 Berger et al.
`4,905,265 2/1990 Cox et al. . . . . .
`
`. . . . . . . . .. 357/30
`
`250/370.01
`. . . . . . . . .. 378/99
`
`of Gainesville, Fla_
`
`_
`[73] Asslgnoor General Imagining Corporation,
`Gainesville, Fla.
`
`[21] APPL N°~= 462,042
`[22] Filed:
`Jan. 8, 1990
`
`[63i
`
`Related U'S' Apphcanon Data
`Continuation-impartM4561‘:NO- 1513235"Fe§- 1, 1983'
`Pat‘ No' 4905364‘ Wh‘ch ‘5 a commuanon'm'pan of
`Ser. No. 807,650, Dec. 11, 1985, abandoned.
`
`[51] Int. Cl.5 ......................... .. G01T U29; 6011' 1/16
`[52] US. Cl. ........................ .. 250/370.09; 250/37011;
`250/336.1; 250/366
`[58] Field of Search ........... .. 250/3361, 336.2, 370.09,
`250/370“, 366; 357/32
`References Cited
`
`[56]
`
`U.S. PATENT DOCUMENTS
`
`. . . .. 357/32 X
`4,927,771 5/1990 Ferreh . . . . . . . . . .
`250/37009
`4,940,901 7/1990 Henry et a1. .
`4,948,972 8/1990 Guyot ........................... .. 250/37011
`_
`Primary Examiner-Constantine Hannaher
`Assistant Examiner-Edward J. Glick
`Attorney. Agent. or Firm-Foley & Lardner
`[57]
`ABSTRACT
`The x-ray detector comprises a plurality of closely adja
`cent radiation sensing arrays facing a plurality of pre
`processing circuits. Burnp bonds are connected directly
`f
`.
`.
`.
`.
`.
`.
`rom said sensing arrays to said preprocessmg c1rcu1ts.
`Added absorber material may be included next to the
`sensing arrays for converting impinging x-rays to free
`electrons for affecting a charge on the sensing arrays.
`Alternately, a scintillator may be positioned between
`the sensing arrays and the preprocessing circuits and
`sensing cells in the sensing arrays may be exposed to the
`scintillator.
`
`4,660,066 5/1987 Reid .................................... .. 357/30
`
`20 Claims, 11 Drawing Sheets
`
`X- RAY RADIATION
`
`Raytheon2062-0001
`
`Sony Corp. v. Raytheon Co.
`IPR2015-01201
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 1 of 11
`
`5,043,582
`
`25::
`$28: 8a 5%
`
`(a
`
`a “2%.. @2182
`
`65E 5:; E N
`
`22:5 E855 _
`
`a gas:
`
`
`av . age: as: $1‘
`
`
`
`
`
`13:82 225.831 .652
`
`a 2:2 5;: )0.
`
`Q. m, s “23%
`
`~ , _\ H c
`
`E.
`
`222255
`\. “55%.?
`
`2
`
`Raytheon2062-0002
`
`

`
`U.S. Patent
`
`Aug. 27, 1991
`
`Sheet 2 of 11
`
`5,043,582
`
`’"_“--7
`
`3
`
`'
`
`I
`
`I
`
`L___._..._____r~\
`:
`2’:
`0
`-J
`"""___'I’
`
`H.:<%_.__=8
`
`
`
`S.(cm
`
`I I
`
`____HiJ gl
`
`2.III.<'\:$2:
`
`2.2,
`
`IL
`
`.5.:=._
`
`:8
`
`fir
`
`88>8
`
`1344:;
`
`NTNS.LI
`.Ea_<S:5LI\
`
`ca2.2,
`:32:/3
`
`Raythe0ni062-0003
`
`#H_..s_H_.w:__....____3“&joofij_.__II.__II.ME:
` a_......J.flII_
`
`
`=25IE 8s
`
`Raytheon2062-0003
`
`

`
`U.S. Patent
`
`Aug. 27, 1991
`
`Sheet 3 of 11
`
`5,043,582
`
`!IDDDDDUm
`
`\\
`
`I
`"'L_J
`- ~..{_}
`100
`gmbmmmamm
`moo
`Dabamammm
`
`It
`-~‘l|
`
`—
`
`Fms
`
`,
`
`S14
`
`EI4
`
`EM
`
`Raythe0n2062-0004
`
`Raytheon2062-0004
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 4 of 11
`
`5,043,582
`
`F|G.7
`
`H0
`
`Hi4
`
`l4’\.
`
`0
`
`.
`I? 22 v
`T__'___
`
`Raytheon2062-0005
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 5 of 11
`
`5,043,582
`
`START
`
`‘20
`
`INITIALIZE NR2
`SET m=|-,n'l
`
`READ \‘(m,n) “'24
`
`LOOK UP f(m,n)
`
`DETERMINE
`{(m,n) GRAY SCALE
`INTERVAL 1
`
`~ I28
`
`E
`LOOK UP 1* (j) N150
`
`E
`
`f (m,n) "B2
`
`DETERMINE GRAY
`SCALE INTERVAL
`FOR +* (m,n)
`
`ASSlGN GRAY
`SCALE T0 (m,n)
`
`END
`
`INCREMENT
`m,n
`
`Raytheon2062-0006
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 6 of 11
`
`5,043,582
`
`:dE
`
`O- .91
`
`22% 3 03m
`
`39.23%
`
`_ _ . _ _ _ _
`
`PIXEL INTEGRATION TIMEImilliseconds)
`
`00.0w
`
`Oom
`
`_
`
`O
`
`PIXEL LOGIC HOLDTIME (seconds)
`AT ROOM TEMPERATURE
`
`Raytheon2062-0007
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 7 of 11
`
`5,043,582
`
`Raytheon2062-0008
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 8 of 11
`
`5,043,582
`
`206/\_
`
`2l2/\.
`
`204%
`
`X-ray radia’rion
`
`d-Si Substrate
`
`x-ray absorber lay
`
`2'5”
`
`Pryp->-*-pi’raxial layer
`
`2MP»
`
`2167
`Si02’\' l
`
`25m.
`
`2l8
`
`\ ‘
`
`220
`
`N+ ) \‘ N+I
`
`I
`
`/77
`SENSING
`AREQ
`
`2l9'-\-
`6
`Y MOS
`Switch
`
`22l»-\
`o
`XMOS .
`Switch
`
`\224
`d SIGNAL READ
`
`Raytheon2062-0009
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 9 of 11
`
`5,043,582
`
`8
`
`3528
`
`2:38
`
`ME:
`
`Raytheon2062-0010
`
`

`
`US. Patent
`
`Aug. 27, 1991
`
`Sheet 10 of 11
`
`5,043,582
`
`
`
`
`
`852:...5389:.13:28
`
`m_.o_n_
`
`£_Em.__.m5Lo
`
`5222:>21.%
`
`
`
`.33__._s:._i.te:._
`
`aw
`
`
`
`.S.:=o:85”
`
`.s_:>_§._
`
`§(
`
`New
`
`
`
`:52.§$§.E._
`
`Raythe0n2062-0011
`
`Raytheon2062-0011
`
`
`

`
`U.S. Patent
`
`Aug. 27, 1991
`
`Sheet 11 of 11
`
`5,043,582
`
`
`
`
`
`3...:t2_%w_8_=2$:
`
`25:.
`
`Raythe0n2062-0012
`
`
`
`P6:tsasm_8_._2.$z
`
`
`
`£.5+$__ow_.£_:__s3
`
`9___.._§2:$5
`
`m_.o_.._
`
`as.IL
`
`
` .£2“.m2__oE3lav:
`
`.2._$__.=:B25
`
`....._=_
`
`.5;_._om
`
`om fl
`
`95.3:
`
`_8_.=.$:
`
`:33
`
`33388:.
`
`$.54
`
`Raytheon2062-0012
`
`
`

`
`_
`
`"
`
`1
`
`5,043,582
`
`2
`Yet another object of the present invention is provide
`an x-ray detector which is relatively inexpensive to
`fabricate so as to enable its use in ?xed locations for ease
`of periodic x-ray analysis of mechanical structures and
`the like.
`In accordance with the above and other objects, the
`present invention is an x-ray imaging system comprising
`an x-ray source for producing an x-ray ?eld, and an
`x-ray detector. The x-ray detector comprises a solid
`state integrated circuit having a plurality of charge
`storage devices and a circuit for placing a charge on the
`charge storage devices. The charge storage devices are
`disposed in an x-ray permeable material and the detec
`tor is positioned in the x-ray ?eld such that the charge
`is dissipated by secondary radiation produced by inter
`action of the x-ray ?eld in the silicon substrate and
`metallization layer of the solid state integrated circuit.
`The charge storage devices may be divided into
`groups to form pixels. Each pixel comprises one or a
`plurality of charge storage devices and the exposure
`times for discharging the charge storage devices in a
`single pixel can be different from one another to provide
`a gray scale.
`In accordance with other aspects of the invention, the
`integrated circuit may be a dynamic random access
`memory.
`'
`Each charge storage device comprises a single cell of
`the integrated circuit. The cells are spaced from each
`other such that dead space exists therebetween. Also,
`the cells are produced in banks of 32,000 with about 5
`mm dead space between banks. A plurality of detectors
`may be stacked with the cells of the detectors staggered
`such that each cell of one detector is positioned behind
`the gap between cells of another detector so as to elimi
`nate all dead space.
`The imaging system also includes processing cir
`cuitry for accessing the cells of a detector. The process
`ing circuitry may include a system for normalizing the
`outputs of all of the cells to compensate for various
`inherent differences in radiation sensitivities of the vari
`ous cells.
`One of the most important aspects of the digital radi
`ography technique employed in the present invention
`compared to conventional systems using silver halide
`?lm is the ability to perform quantitative radiography.
`This is achieved practically through image digitization
`and makes subtraction of radiographic images an ex
`tremely useful enhancement technique.
`The x-ray image detection system according to the
`present invention is based on direct acquisition of digital
`information, utilizing solid-state silicon and hybrid de
`tectors. An x-ray image of an object is projected di~
`rectly onto the sensor without any intermediate x~ray
`to-light conversion and signal magni?cation. Secondary
`electrons produced by x-ray interactions with the sili
`con substrate and metallization layer are collected and
`digitized using techniques similar to those employed for
`visible light detection.
`One of the major concerns of direct x-ray sensing is
`designing a solid-state sensor that can withstand the
`radiation dose accumulation suf?ciently_to justify the
`cost. of replacing the degraded detectors. The sensor
`must have good x-ray sensitivity compared to other
`systems with typical x-ray spectra (30 kVp to 200 kVp),
`and should have a capability of sensing a continuous
`large format image. To solve this problem, the sensor
`used in the present invention is a conventional DRAM
`device. The cost of producing such a device is orders of
`
`X-RAY IMAGING SYSTEM AND SOLID STATE
`DETECTOR THEREFOR
`‘
`
`CROSS REFERENCE TO RELATED
`APPLICATIONS
`This is a continuation-in-part of US. application Ser.
`No. 151,235, ?led Feb. 1, 1988, now US. Pat. No.
`4,905,265, which is a continuation-in-part of US. appli
`cation Ser. No. 807,650, ?led Dec. 11,1985, now aban
`doned.
`
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`This invention relates to x-ray imaging systems and
`particularly to an x-ray imaging system which utilizes a
`solid state x-ray detector.
`2. Discussion of Related Art
`Presently, x-ray imaging systems are utilized in a
`variety of applications, both as medical diagnostic tools
`and for industrial quality control. The most common
`form of x-ray detection resides in the use of silver halide
`?lm. However, the use of such ?lm requires the perfor
`mance of several wet, control requiring chemical devel
`oping steps. In addition, this ?lm is expensive, thus
`increasing the cost of x-ray images produced in this
`manner.
`It would be highly desirable, therefore, to produce an
`x-ray imaging system which does not require the use of
`silver halide ?lm. Several detectors have been proposed
`for this purpose.
`For example, US. Pat. No. 4,471,378 to Ng discloses
`a light and particle image intensi?er which includes a
`scintillator and photocathode unit for converting inci
`dent image conveying light or charged particles to
`photoelectrons and a charge coupled device for detect
`ing the photoelectrons and transmitting to data process
`ing and video equipment information relating to the
`quantity or energy level as well as the location of the
`electrons impinging on the sensing areas of the charge
`couple device.
`US. Pat. No. 4,413,280 to Adlerstein et al discloses an
`x-ray imaging apparatus which includes a transducer
`for converting incident x-radiation to a corresponding
`pattern of electrical charges. The charges generated by
`the transducer are accelerated onto an array of charge
`detecting or charge storing devices which store the
`charges in the form of an electrical signal corresponding
`to the charge pattern.
`SUMMARY OF THE INVENTION
`One object of the present invention is to provide a
`solid state imaging system and detector which are
`highly sensitive to x-radiation and can produce highly
`accurate x-ray images.
`Another object of the present invention is to provide
`an x-ray imaging system and detector which can be
`produced by conventional solid state fabrication tech
`nology.
`A further object of the present invention is to provide
`a solid state imaging detector which can be produced in
`such small sizes as to enable its use in very con?ned
`
`15
`
`20
`
`45
`
`50
`
`55
`
`Another object of the present invention is to provide
`an x-ray imaging detector which can be substituted
`directly for x-ray ?lm used in conventional x-ray imag
`ing systems.
`
`65
`
`Raytheon2062-0013
`
`

`
`5,043,582
`3
`magnitude less than producing other types of sensors,
`such as CCD and CID arrays.
`‘
`
`4
`Detector 14 may be a dynamic random access mem
`ory such as the I832 OpticRAM sold by Micron Tech
`nology, Inc. of Boise, Id. This device is an integrated
`circuit DRAM having 65,536 elements and is used as a
`solid state light sensitive detector. The Micron DRAM
`is speci?cally adapted to sense light inasmuch as there is
`no opaque surface covering the integrated circuit.
`However, any type of dynamic random access memory
`may be used for detector 14 as long as the covering is
`transparent to x-radiation. In fact, as will become appar
`ent, any type of dynamic memory element may be used
`as detector 14. The memory element does not have to
`be a random access memory, although the use of a ran
`dom access memory facilitates preprocessing and image
`processing routines.
`Moreover, random access circuitry decouples one
`pixel from another. Unlike charge coupled architecture
`employed in CCD arrays where one dead pixel affects
`others in the same row, pixels that are randomly accessi
`ble do not affect others should they become defective.
`This phenomenon plays an important role in the longev
`ity and cost of large-scale imagers where tens of mi]
`lions of pixels are employed. From the standpoint of
`longevity, random bit (pixel) failure is a common side
`effect of radiation damage, as well as manufacturing
`processes. A ZO-million pixel random access imager
`with 14" X 17” dimensions could have as many as ten or
`twenty thousand dead pixels so long as they were ran
`domly distributed. If one failed pixel could affect oth
`ers, random bit (pixel) failure would propagate. causing
`entire rows of pixels to fail, greatly reducing the ima
`ger’s lifetime and manufacturing yield. Both of these
`parameters have a direct impact in the cost to produce
`the imager and the cost per image, respectively.
`The I832 OpticRAM image sensor is a solid-state
`device capable of sensing an image and translating it to
`digital computer-compatible signals. The chip contains
`two arrays each'of which contains 32,768 sensors ar
`ranged as l28 rows by 256 columns of sensors (4,420
`microns><876.8 microns). Each pixel, 6.4 microns on a
`side, consists of two elements, a MOS capacitor and a
`MOS switch. The fill factor is 50 percent. The sensor is
`a random access device and thus, pixels may be individ
`ually accessed.
`The detector 14 operates by the projection of radia
`tion penetrating the object onto the 65,536 radiation
`sensitive elements of each array-pair. Radiation striking
`a particular element will cause the capacitor, which is
`initially charged to ?ve volts, to discharge toward zero
`volts. The capacitor will discharge at a rate propor
`tional to the intensity of the radiation ?eld to which it is
`exposed.
`~
`To determine whether a particular element is black or
`white, one can read the‘ appropriate row and column
`address associated with the physical location of the
`element. The sensor reads the voltage value of the ca
`pacitor and performs a digital comparison between the
`voltage of the capacitor and a ?xed externally applied
`threshold voltage bias. A white pixel indicates the ca
`pacitor is exposed to a radiation field suf?cient to dis
`charge the MOS capacitor below the threshold point,
`whereas a black pixel has not received suf?cient expo
`sure.
`The output of detector 14 is passed to preprocessor 16
`which serves the function of normalizing the outputs of
`all of the cells of detector 114. That is, the sensitivity of
`the cells of detector 14 will inherently vary. A normal
`ization value can be stored in preprocessor l6 so as to
`
`15
`
`25
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`The above and other objects of the present invention
`will become more readily apparent as the invention is
`more clearly understood from the detailed description
`to follow, reference being had to the accompanying
`drawings in which like reference numerals represent
`like parts throughout, and in which:
`FIG. 1 is a block diagram of the x-ray imaging system
`of the present invention;
`FIG. 2 is a circuit diagram of an integrated circuit
`detector used in the imaging system of FIG. 1;
`FIG. 3 is an enlarged schematic showing one charge
`storage capacitor of the circuit diagram of FIG. 2;
`FIG. 4 is a cross section of a chip showing the struc
`ture depicted schematically in FIG. 3;
`FIG. 5 is a view of a portion of the detector of the
`present invention stacked over additional detectors to
`?ll up the dead space between cells;
`FIG. 6 is an end elevational view of the stacked de
`tectors of FIG. 5;
`FIG. 7 is a diagrammatic representation showing the
`system of the present invention used in place of x-ray
`film;
`FIG. 8 is a flow diagram depicting a method of nor
`malizing the cells of the present invention;
`FIG. 9 shows the orientation of the detector and the
`detector leads for backscattering images;
`FIG. 10 is a graph showing pixel logic hold time as a
`function of accumulated radiation exposure based on
`data taken at 120 kVp ?ltered through 0.25 mm A1 with
`50% of total detecting pixels discharged beyond the
`threshold point;
`FIG. 11 is a graph showing pixel integration time as
`a function of accumulated exposure based on data taken
`at 120 kVp ?ltered through 0.25 mm Al;
`FIG. 12 shows a thin ?lm detector embodiment of the
`present invention;
`FIG. 13 is a cross section of one detector of the em
`bodiment of FIG. 12;
`FIG. 14 is a circuit diagram of the detector of FIG.
`12;
`FIG. 15 is a cross section showing the interconnec
`tion between the sensing layer and the preprocessor
`portion of the detector of FIG. 12; and
`FIG. 16 is a perspective cross sectional view of a
`further embodiment of the present invention in which a
`scintillator is used to convert x-rays to visible light.
`
`35
`
`45
`
`DESCRIPTION OF THE PREFERRED
`EMBODIMENT
`FIG. 1 shows the x~ray system 10 to comprise a high
`energy x-ray source 12 and a detector 14 positioned to
`receive the radiation from source 12. Source 12 can be
`any standard high energy x-radiation source having an
`output in the range of 8 Kev or higher. Sources such as
`this are well known and manufactured by, for example,
`GB. or Siemens. Alternatively, source 12 be an ultra
`small focal spot source such as manufactured by Ridge
`or Magnaflux, also having an output in the range of 8
`Kev or higher. Any size focal spot source can be used.
`Currently, the smallest focal spot available is one mi
`cron. Also, the source 12 and detector can be placed as
`close to the object 0 to be x-rayed as desired due to the
`con?guration of the detector M, as will become readily
`apparent.
`
`55
`
`65
`
`Raytheon2062-0014
`
`

`
`20
`
`5
`normalize the output of each of the cells to ensure a
`?eld describing reading.
`The normalized output of preprocessor 16 is passed
`to image processor 18 which manipulates the data using
`conventional image processing programs as well as new
`image processing programs which will be made possible
`by the present invention, such as “zoom” programs
`which are not currently in existence. This image can be
`displayed on a high resolution monitor 20, can be stored
`on a laser disc recorder 22, can be printed using a dry
`silver printer 24, or can be sent via satellite to remote
`image processors (not shown). A menu driven program
`is displayed on a computer monitor 26 prompting ap
`propriate instructions and data which can be entered
`into the image processor 18 using a keyboard 28.
`FIG. 2 shows a schematic diagram of a portion of a
`typical DRAM used in detector 14. The circuit 30 com
`prises a plurality of cells 32, each of which contains a
`memory capacitor 34 and an access transistor 36. The
`individual cells are accessed through left and right digit
`lines 38 and 40, respectively, as well as word lines 42
`and 44. A sense ampli?er 46 is provided in the form of
`a cross coupled MOSFET detector circuit. The sense
`ampli?er 46 has nodes A and B which are coupled to
`the left digit line 38 and to the right digit line 40, respec
`tively. The cells 32 are divided into a left array 50 and
`a right array 52. The left array 50 is accessed by the left
`digit line 38 and the right array 52 is accessed by the
`right digit line 40. The word lines 42 access the individ
`ual cells of array 50 and the word lines 44 access cells of
`the array 52.
`A pair of equilibrate transistors 56 and 58 couple the
`digit lines together to allow equalization of the digit
`lines at the end of a refresh cycle and during the re
`charge state of the next cycle.
`The common drains of the cross coupled sense ampli
`?er transistors at node C are connected through an
`isolation transistor 60 to a pad 62 on the periphery of the
`integrated circuit chip. The pad 62 is bonded to one of
`the leads of the circuit chip package.
`A pair of pull up circuits 66, 68 are coupled, respec
`tively, to the nodes A and B. The pull up circuits 66, 68
`are voltage divider circuits operable to control the volt
`age level of the digit lines 38 and 40.
`FIG. .3 shows one cell of the circuit 30. For conve
`nience, the cell is shown to be one of the array 52 but it
`could be any of the cells. As shown, the capacitor 34 has
`two plates 70 and 72 between which a charge is stored.
`Initially, the capacitor is charged by applying a high
`potential on word line 44 and a high potential on right
`digit line 40. This corresponds to a .“1” state of the cell.
`In the presence of incident x-radiation, the charge on
`capacitor 34 is dissipated as will be discussed below.
`With reference to FIG. 4, the portion of the inte
`grated circuit containing the cell shown .in FIG. 3 is set
`forth in cross section. The circuit comprises a p-type
`silicon substrate 80 onto which an n+region 82 has
`been added. A silicon dioxide layer 84 is deposited over
`the substrate and n+region 82 to form an insulating
`layer. The lead 40 is connected to the n+region to form
`the drain of transistor 34. A metal plate 86 is formed on
`the oxide layer 84 to form an insulated gate of transistor
`36. The capacitor 34 is formed by a metal plate 72 and
`_the interface 70 between p-type substrate 80 and oxide
`layer 84, which forms the other capacitor plate.
`When the cell of FIGS. 3 and 4 is set to the “1” state,
`charge is built up on the interface 70 to charge the
`capacitor 34. The gate voltage is then lowered so as to
`
`5,043,582
`6
`discontinue communication between one drain voltage
`at line 40 and the capacitor 34. This charge is dissipated
`due to the absorption of x-ray photons in the substrate
`80. In FIG. 4, the direction of incident x-radiation is
`shown by the arrow 88.
`The x-radiation can produce free electrons in the
`substrate 80 either by photoelectric effect, Compton
`scattering, or pair production. However, because of the
`high energy of the source used in the present system,
`the number of electrons produced through photoelec
`tric effect is negligible. The x-ray energy is in the range
`where compton scattering and pair production have the
`highest probabilities of producing free electrons.
`It is noted that interconnections between components
`of a cell and between cells are provided on the oxide
`layer of the semi-conductor. This is indicated in FIG. 4
`by showing the leads 40, 44 and 45 extending out of the
`oxide layer. Such leads represent the interconnections
`produced by the metalization layer of an integrated
`circuit.
`Irradiation, of the cell 30 from either side results in
`virtually all of the x-radiation being received in the
`substrate 80 so that the total electron production for any
`given energy level of radiation is achieved. The free
`electrons produced by interaction between the substrate
`and the x-radiation decrease the charge at junction 70
`and thus decrease the charge on capacitor 34.
`In conventional x-ray systems, compton scattering
`and the photoelectric effect are the relevant interactions
`causing the appearance of free electrons. Rayleigh scat
`tering, a type of coherent scattering may also be respon
`sible for the production of some free electrons. The
`relative occurrences of the different reactions depends
`on the energy of the x-ray. As discussed above, the
`present invention uses a high energy source. Rayleigh
`scattering and photoelectric effect are low energy inter
`actions so that the number of free electrons produced by
`the these effects in the present invention is negligible.
`There is little direction sensitivity in any of the interac
`tions relating to the production of free electrons except
`in the case of Rayleigh scattering, which is predomi
`nantly forward but also yields the least free electrons.
`Referring again .to FIG. 2, it will be understood that
`the circuit 30 is of the type which employs a dynamic
`/active/restore sense ampli?er of the type described in
`US. Pat. No. 4,397,002, issued Aug. 2, 1983 to Wilson et
`al, and US. Pat. No. 4,291,392, issued Sept. 22, 1981 to
`Proebsting.
`In operation of circuit 30, during one cycle, a given
`word line 42, 44 in FIG. 2 is brought to a logic one level
`to enable the addressed access transistor 36. The respec
`tive cell capacitor 34 is discharged into the appropriate
`digit line (e.g. digit line 40 for a capacitor of array 32)
`changing its value above the equalized value. Then, a
`latch signal from the pad 62 becomes a logic low state to
`enable operation of the cross coupled transistors and the
`sense ampli?er 46 during absence of the equilibrate
`signal. The sense ampli?er 46 responds to the latch
`signal by reducing the opposite digit line (in this case
`digit line 38)_ to a ground potential. The digit lines are
`connected by input/output circuitry (not shown),
`which provides a digital signal representing the content
`of the selected memory capacitor 36. The pull up cir
`cuits cause the right digit line to be pulled up to the
`level of the supply voltage. At approximately this time,
`the storage capacitor 36 which has been connected to
`the bit line has been restored to its original logic one
`state. The word line is then returned to ground to iso
`
`35
`
`45
`
`65
`
`Raytheon2062-0015
`
`

`
`5
`
`15
`
`25
`
`35
`
`45
`
`5,043,582
`7
`late the charge on the respective memory cell. The digit
`lines are then permitted to go low and the equilibration
`signal becomes a logic high to render the equilibrate
`transistors 56, 58 conductive to allow the digit lines to
`be connected for equalization. This permits the charge
`on the digit lines 38, 40 to be shared such that the digit
`lines equilibrate to a voltage approximately half way
`between the supply voltage and ground. A new cycle is
`thereupon ready to commence.
`A factor affecting the performance of the detector is
`the length of the time which the MOS capacitors are
`exposed to the radiation ?eld. This period of time is
`measured from the initial exposure of an element until
`the time the particular element is read or refreshed.
`Accessing any pixel in a row causes the entire row to be
`refreshed. This sets all the row cells that have not
`leaked below threshold to ?ve volts and sets all cells in
`the row that have leaked below threshold to zero volts.
`Optimal imaging conditions exist when the absorbed
`dose rate is much greater than the discharge rate cre
`ated by dark current in the capacitor circuit.
`There are two factors which have profound effect on
`the dark current: operating temperature and exposure
`history. Generally, the lower the operating tempera
`ture, the lower the dark current; and the longer the
`exposure history, the higher the dark current. The dark
`current of the device directly affects the logic holdtime,
`or the ability of the sensor to integrate images over time.
`At room temperature, the logic holdtime or integrating
`time is approximately 20 seconds. At 40 degrees Fahr
`enheit the logic holdtime increases to 200 seconds. This
`has the impact of allowing the sensor to take images in
`radiation ?elds 20 times weaker. The sensor, however,
`is 20% less sensitive at this lower temperature, which is
`minor compared to the factor of 20 increase in integra
`tion time. The sensor also has a much greater resistance
`to radiation damage at lower temperatures. At room
`temperature, the chip fails at an absorbed dose of 10
`KRads. The chip is considered failed when it can no
`longer integrate an image ‘more than two seconds. AT
`40 degrees Fahrenheit, it takes more than 150 KRads, a
`factor of 15. This is an unexpected result and no expla
`nation for this phenomenon is available.
`The best approach to cooling the imager during oper
`ation is to place the imager array in direct contact with
`the cold side of a single or dual stage thermoelectric
`cooler (such as a thermoelectric cooler manufactured
`by Melcor, Materials Electronis Products Corporation
`of 990 Spruce Street, Trenton, NJ. 08648). To maintain
`a ?xed (lowered) operating temperature of the array, a
`heat sink must be employed on the hot side of the ther
`moelectric cooler. The heat sink could be in the form of
`an array of metal ?ns or a liquid coolant.
`The latch signal is placed on pad 62 sucli that during
`the equilibrate signal, a voltage potential may be applied
`to the digit lines when they are connected to effect
`equilibrium. As will become apparent, the voltage
`which is applied to the pad 62 allows adjustment to the
`sensitivity of the image sensor. In particular, the digit
`line potential acts as a threshold to determine if a partic
`ular memory cell 36 is at a high or low voltage level. By
`raising the potential, cells may leak less before they are
`considered to be decayed from a logic one to a logic
`zero value. Accordingly, the sensitivity of the cells can
`be adjusted by adjusting this potential.
`Another way of adjusting the sensitivity of the cells is
`by etching in different thicknesses of oxide material in
`the capacitor 34- shown in FIG. 4. In other words, the
`
`8
`thickness of the oxide between interface 70 and metal
`plate 72 determines the sensitivity of the capacitor to
`x-radiation. The thicker this layer, the more sensitive
`the capacitor is to radiation discharge. Accordingly, to
`make a more sensitive detector, the layer should be
`made thicker and to make a less sensitive detector, the
`layer should be made thinner. There are, however,
`limitations to this technique because the impedance of
`the capacitor must lie in a certain range in order for the
`circuit to function properly (varying the thickness
`changes the impedance).
`A further technique for adjusting sensitivity is to
`adjust the exposure time of the detector cells. Using
`conventional x-ray source controlling equipment, the
`response of the cells can be adjusted either by using
`higher energy x-rays or by increasing the intensity of
`the x-rays. Clearly, in either case, the cells will react
`more quickly.
`The OpticRAM has a broad spectral sensitivity
`range. In the UV-visible-IR portion of the spectrum
`(300—1,2OO nanometers) the sensitivity is fairly uniform
`requiring a fluence of about 2 microjoules per square
`centimeter to discharge the sensor to threshold (typi
`cally 2.5 volts). In the x-ray region of the spectrum the
`fluence required to reach threshold is about 0.2 mi
`crojoules per square centimeter (at 120 kVp). The de
`tector responds almost linearly with x-ray energy spec
`tra from 20 kVp to 120 kVp. The response of the sensor
`below 15 kVp is low totally due to the strong absorp~
`tion of the ?lter employed. In fact, most solid~state
`planar devices have optimum sensitivity in the photon
`energy range from 1 keV to 30 keV.
`The inherent x-ray sensitivity of a silicon sensor is a
`function of the device structure, process parameters and
`cell con?guration. The usual IS32 OpticRAM is not in
`the optimum condition in any of those three aspects.
`The dielectric layer of the device was modi?ed to ex
`amine its effect on sensitivity. Test results show that
`increasing the gate oxide layer thickness from 300 Ang
`stroms to 600 Angstroms increases the detector sensitiv
`ity by an order of magnitude. This is interesting since,
`by doubling gate oxide thickness, one not only doubles
`the active volume ‘for x-ray interaction in the oxide
`region (assuming depletion region remained unaffected)
`but also reduces the total number of events of x-ray
`interaction necessary for the MOS capacitor to dis
`charge below the threshold point. The expected im
`provement is about four fold, rather than the 10 fold
`_ observed in the measurement. Experiments on the opti
`cal sensitivity show only a three fold improvement with
`50
`the same change. It is not clear what constitutes the
`dramatic x-ray sensitivity improvement (two-and-one
`half times greater than expected). There is an indication
`that this method may have a limit. In an experiment
`conducted on an IS64lO OpticRAM, and increase in
`oxide layer from 600 Angstroms to 800 Angstroms
`improved detector sensitivity by only 25 percent.
`One of the device structural parameters which has
`been changed is removal of a silicon nitride layer nor
`mally deposited on the gate oxide layer of the capacitor.
`The nitride layer is common in DRAMs to reduce bit
`soft-errors, but also appears to reduce x-ray sensitivity
`by about two orders of magnitude. A possible explana
`tion for this observation is tha