Multiplexing pixel circuits

Information

  • Patent Grant
  • 6414665
  • Patent Number
    6,414,665
  • Date Filed
    Wednesday, November 4, 1998
    25 years ago
  • Date Issued
    Tuesday, July 2, 2002
    22 years ago
Abstract
An active matrix display in accordance with the present invention includes a plurality of pixels arranged in an array. At least two transistors associated with each pixel are included. The transistors are serially connected to each other and disposed within the array for switching the pixels on and off according to data and gate signals. A data line is coupled to a first end of the serially connected transistors for each pixel. A second end of the serially connected transistors is coupled to a storage device. The serially connected transistors provide multiplexing capability for at least one of data signal multiplexing and gate signal multiplexing.
Description




BACKGROUND OF THE INVENTION




1. Field of the Invention




The present invention relates to pixel display circuits and, more particularly, to a circuit for providing integrated data and gate multiplexing without impact to acceptable standards for displays.




2. Description of the Related Art




Due to poor charging ability in amorphous silicon thin film transistors (a-Si TFTs) resulting from inherently low TFT transconductance, all commercially available a-Si TFT liquid crystal displays (LCD) include an array of pixel elements connected with row and column metal lines. The row and column drivers require higher transconductance devices. The row and column drivers typically include crystalline silicon technology and are separately fabricated and attached to the a-Si TFT LCDs. Over the years, there have been attempts at integrating some level of multiplexing between the attached crystalline silicon drivers and the pixel array. See for example, U.S. Pat. No. 5,175,446 to R. Stewart. In this way, the number of crystalline drivers needed could be reduced. These prior art designs follow a circuit approach that is commonly used in crystalline silicon circuit designs. Even simple 2:1 level multiplexing schemes at the edge of a pixel array have not been implemented for a-Si TFT LCD circuits. Although not realized for direct view a-Si TFT LCDs, multiplexer circuits have been implemented with some success in smaller displays for example, in light valves and in poly-silicon technology. Poly-silicon TFTs make it possible to realize a higher transconductance TFT. However, implementing poly-silicon technology on larger and/or high resolution TFT LCDs leads to an unacceptably higher RC load and/or higher bandwidth rates of the rows and columns.




Therefore, a need exists for a circuit for providing integrated data and gate multiplexing for active matrix LCDs without impacting acceptable display limits. A further need exists for a reduction in data drivers and gate drivers to reduce costs of these displays.




SUMMARY OF THE INVENTION




An active matrix display in accordance with the present invention includes a plurality of pixels arranged in an array. At least two transistors associated with each pixel are included. The transistors are serially connected to each other and disposed within the array for switching the pixels on and off according to data and gate signals. A data line is coupled to a first end of the serially connected transistors for each pixel. A second end of the serially connected transistors is coupled to a storage device. The serially connected transistors provide multiplexing capability for at least one of data signal multiplexing and gate signal multiplexing.




Another active matrix display in accordance with the present invention includes a plurality of pixels arranged in an array including rows and columns. At least two transistors associated with each pixel, the transistors being serially connected and positioned within the array for switching the pixels on and off. A plurality of data lines run substantially parallel to the columns. A plurality of scan lines run substantially parallel to the rows. The data lines and scan lines being coupled to the transistors of the pixels such that the data lines provide data multiplexing for each pixel and the scan lines provide gate multiplexing for each pixel.




In alternate embodiments of the displays in accordance with the invention, one of the at least two transistors may be shared between adjacent pixels to further reduce gate or data drivers. The pixels may modulate light in a transmissive mode and/or a reflective mode. The array preferably includes rows and columns and the control lines may select the pixels in different rows simultaneously. The simultaneously selected pixels may share a data line. Control lines may include data lines and the simultaneously selected pixels may each use a different data line. The control lines may includes scan lines and/or capacitance storage lines. The control lines may be coupled to the transistors by a low impedance contact path which may include a metal, polycrystalline silicon, a capacitor or a combination thereof. The display may include a liquid crystal display, and the transistors preferably include thin film transistors.




The data multiplexing may include L:1 multiplexing where L is an integer greater than one. The gate multiplexing may include m:1 multiplexing where m is an integer greater than one. The data lines may select pixels in different rows simultaneously, or the pixels may share a data line. The display may further include logic circuitry for controlling the multiplexing in accordance with control signals. The transistors are preferably disposed on a substrate and the pixels are formed over the transistors




These and other objects, features and advantages of the present invention will become apparent from the following detailed description of illustrative embodiments thereof, which is to be read in connection with the accompanying drawings.











BRIEF DESCRIPTION OF DRAWINGS




The invention will be described in detail in the following description of preferred embodiments with reference to the following figures wherein:





FIG. 1

is a schematic diagram of a pixel circuit showing two TFTs in series per pixel in accordance with the present invention;





FIG. 2

is a timing diagram for the circuit of

FIG. 1

in accordance with the present invention;





FIG. 3

is a schematic diagram of a pixel circuit showing three TFTs in series per pixel and additional scan lines in accordance with the present invention;





FIG. 4

is a timing diagram for the circuit of

FIG. 3

in accordance with the present invention;





FIG. 5

is a schematic diagram of another pixel circuit showing two TFTs in series per pixel in accordance with the present invention;





FIG. 6

is a timing diagram for the circuit of

FIG. 5

in accordance with the present invention;





FIG. 7

is a schematic diagram of a pixel circuit effectively having one and one half TFTs per pixel due to a shared TFT between adjacent pixels in accordance with the present invention;





FIG. 8

is a timing diagram for the circuit of

FIG. 7

in accordance with the present invention;





FIG. 9

is a schematic diagram showing a fanout wiring for a demultiplexer in an active matrix array in accordance with the present invention;





FIG. 10

is a schematic diagram of another embodiment of the pixel circuit of

FIG. 5

showing an additional scan line to achieve 4:1 multiplexing in accordance with the present invention;





FIG. 11

is a timing diagram for the circuit of

FIG. 10

in accordance with the present invention;





FIG. 12

is a schematic diagram of a pixel circuit having a four pixel layout with a common data line and three TFTs in series per pixel in accordance with the present invention;





FIG. 13

is a timing diagram for the circuit of

FIG. 12

in accordance with the present invention;





FIG. 14

is a schematic diagram of a pixel circuit effectively having one and one half TFTs per pixel due to a shared TFT between adjacent pixels for 4:1 multiplexing in accordance with the present invention;





FIG. 15

is a timing diagram for the circuit of

FIG. 14

in accordance with the present invention; and





FIG. 16

is a top plan view of a semiconductor device layout implementing the present invention and showing logic circuitry.











DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS




The present invention relates to liquid crystal display circuits and, more particularly, to a circuit for providing integrated data and gate multiplexing without impact to acceptable standards of the displays. The present invention provides a novel design for introducing gate and data line multiplexing functions from circuitry implemented within the pixel rather than from circuitry at the data or gate line ends. The present invention does not follow the prior art schemes of implementing crystalline silicon multiplexer designs with amorphous silicon thin film transistors (TFTs), nor does the present invention follow the prior art scheme of placing the multiplexers between the pixel array and the externally attached crystalline silicon drivers.




The present invention provides an RC load that is reduced to about a load within a pixel. This may be implemented by employing a TFT gate of minimum length and width as described hereinbelow. Whereas the prior art needed to drive a full load of the gate or data line to accomplish gate or data multiplexing.




The present invention employs more than one TFT in a pixel. All TFTs within the pixel are connected in series. This connection includes a first source connected to a data line and a last drain in the series of TFTs connected to a storage capacitor and/or an LC capacitor. Multiplexing as provided by the present invention places the multiplexer circuitry within each pixel and employs several serially connected TFTs to provide the multiplexing function. A data line is connected to the source of one TFT, and the drain of this TFT is connected to the source of another TFT and so on until the last serially connected TFT has its drain connected to one electrode of a storage capacitor and a liquid crystal flag. The liquid crystal flag (not shown) is one of the electrodes forming a liquid crystal capacitor. The liquid crystal flag is preferably formed from indium-tin oxide (ITO).




The number of serially connected TFTs may be two or more. The more TFTs per pixel, the higher the level of multiplexing that may by implemented. However, the more TFTs per pixel the lower the aperture ratio for that pixel since the amount of available chip area is reduced by the presence of the TFTs and enable lines connected thereto. There is no minimum aperture ratio for reflective displays since the circuitry may be placed below reflector pixels. Referring now in detail to the figures in which like numerals represent the same or similar elements and initially to

FIG. 1

, a schematic diagram of a two pixel circuit


10


in accordance with the present invention is shown. Pixel circuit


10


includes two pixels


12


with a common data line (DATA LINE) and two thin film transistors (TFTs)per pixel


12


, M


2


and M


3


and M


5


and M


6


, respectively. Enabling TFTs M


2


and M


5


are activated by scan line B and scan line E, respectively. Data line signal transfer TFTs M


3


and M


6


provide access from DATA LINE to pixel capacitors CLC


1


and CLC


2


. M


3


and M


6


are enabled by scan line A and scan line F, respectively. Advantageously, all TFTs shown in

FIG. 1

are located at pixels


12


. An illustrative timing diagram is shown in

FIG. 2

to show the states of pixel circuit


10


of FIG.


1


.




Referring to

FIG. 3

, a schematic diagram shows an alternate embodiment of a pixel circuit shown as pixel circuit


20


. Circuit


20


includes additional enabling TFTs M


1


and M


4


in series with the other TFTs. TFTs M


1


and M


4


are enabled by signal line C and signal line D, respectively. In a similar way, additional enabling TFTs may be added in series with the TFTS to provide a similar function. The additional TFTs in series provide, among other things, a higher level of gate line multiplexing, more scan lines through each pixel


22


and an integer number of threshold drops from scan line C (or D) voltage to a gate of TFT M


3


(or M


6


). It is to be understood that if a data voltage at the source of a first TFT (M


1


or M


4


), connected to DATA LINE, is comparable to an ENABLE voltage, a cumulative threshold voltage drop in the data voltage after each TFT will result. This TFT threshold voltage alteration on the data voltage may be rendered nonexistent if the pixel TFTs are overdriven such that the ENABLE voltage is at least one threshold voltage larger than the maximum data line voltage. In one example for amorphous silicon TFT LCDs, this condition is easily satisfied since the data voltage is typically in the range of 10 V or less (i.e., ±5 V centered around a top plate common voltage of about 7 V), and the “on” gate voltage (that would be comparable to the ENABLE voltage) is approximately 13 V or greater than the maximum data voltage, or approximately 25 V.

FIG. 4

illustratively presents a timing diagram for circuit


20


of FIG.


3


.




Referring to

FIG. 5

, another embodiment of the present invention is shown as pixel circuit


30


. The number of scan lines are reduced to effectively two per pixel. Scan lines between pixels are shared to reduce their number. As shown in circuit


30


, scan line D is shared with a previous row of pixels where scan line C is shared with a next row of pixels. Circuit


30


provides a more efficient layout of pixels and thus maximizes aperture area.




An illustrative timing diagram is shown for circuit


30


in FIG.


6


. Referring now to

FIGS. 5 and 6

, a mode of operation is provided where scan line A pulse high width overlaps scan line C and scan line D pulses. The significance of this is that the falling edge of scan line A pulses are completed after that of scan line C of scan line D falling edges so as to ensure that a gate node of M


3


or M


6


is discharged and will not act as a charge storage node which would prevent M


3


and/or M


6


from turning off. Further, only two scan lines are needed to turn a pixel on, for example, scan line A and D or scan line A and C. The third scan line provides a multiplexing capability in accordance with the present invention. In this example, a shared common DATA LINE between pixels


32


provides a 2:1 data multiplexing function, and the three scan lines A, C and D provide an m:1 gate demultiplexing function, where m is an integer greater than 1.




The pulse width on scan line A is larger than that on scan line C or D. This implementation provides scan line pulse flexibility and in general, scan line pulse widths and relative positions may be different between scan line A and scan line D and scan line C. For example, when scan line A is high, enable TFTs M


2


and MS are turned on. Scan line C and D voltages are placed on gates of TFTs M


3


and M


6


, respectively. If scan line C (or D) is high, TFT M


3


(or M


6


) conducts and transfers a data voltage to storage capacitors CS


1


(or CS


2


). If scan line C (or D) is low, TFT M


3


(or M


6


) does not conduct, and the data line voltage is not transferred to CS


1


(or CS


2


). The timing diagram of

FIG. 6

is for normal scan line multiplexing. For simplicity, a liquid crystal voltage is not shown, however the liquid crystal voltage is connected across a source of TFT M


3


(M


6


) where CS


1


(CS


2


) is connected and the other node is connected to a common plate voltage. The common plate extends over a surface of the device and is preferably formed from ITO.




Referring to

FIG. 7

, a circuit


40


is shown which is another embodiment of circuit


30


of FIG.


5


. In circuit


40


, the serially accessed and redundant enable TFTs (M


5


and M


2


) of circuit


30


are replaced with a single TFT M


2


′ which is shared between two pixels.

FIG. 8

illustratively presents a timing diagram for circuit


40


of FIG.


7


.




Referring to

FIG. 9

, an illustrative example of row fan-out wiring for gate demultiplexing is shown. An m:1 gate multiplexing function is provided. Signal line labeled “EN


1


o/e” through EN m*n o/e” represent pulses to the gates of the enable TFTs M


2


and M


5


. The odd (M


2


) and even (M


5


) pixel access TFTs are designated by the nomenclature “o” and “e”. Also, “row #”1 through m*n represent scan lines to which the storage capacitors (i.e., CS


1


and CS


2


) overlap onto. In one example, if m=n, and XGA and SXGA color displays are used, gate driver outputs may be multiplexed to approximately 28:1 and 32:1, respectively.




The above figures have shown gate driver outputs being multiplexed 2:1, however, higher data line multiplexing is possible through the introduction of other scan lines.




Referring to

FIG. 10

, circuit


30


of

FIG. 5

is shown as circuit


50


having an additional scan line G to provide increased data multiplexing from 2:1 to 4:1. Circuit


50


includes four pixels


52


. The TFTs M


8


, M


9


, M


11


and M


12


function similarly to TFTs M


2


, M


3


, M


5


and M


6


. CS


3


and CS


4


function similarly to CS


1


and CS


2


. By repeating the circuit pattern of

FIG. 10

, a circuit is provided in accordance with the present invention which incorporates m:1 gate demultiplexing and L:1 data demultiplexing, where m and L are integers greater than 1.

FIG. 11

illustratively shows a timing diagram for circuit


50


.




Referring to

FIG. 12

, a schematic diagram of a circuit


60


is shown. Circuit


60


includes a common DATA LINE and three TFTs per pixel


62


. Circuit


60


is a 4:1 data demultiplexing representation of circuit


20


of FIG.


3


. Scan lines G, H, I and J are added to accomplish this. The TFTs are serially connected in groups of three (i.e., M


1


, M


2


, M


3


; M


4


, M


5


, M


6


; M


7


, M


8


, M


9


; and M


10


, M


11


, M


12


), each group functions similarly. CS


1


, CS


2


, CS


3


and CS


4


are storage capacitor nodes.

FIG. 13

illustrates an example of a timing diagram for circuit


60


.




Referring to

FIG. 14

, a schematic diagram of a circuit


70


is shown. Circuit


70


includes a common DATA LINE and two TFTs per pixel


72


in which one TFT (M


2


, M


8


) is shared between two pixels (an efficient one and one half (1½)TFT per pixel layout). Circuit


70


is a 4:1 data demultiplexing representation of circuit


40


of FIG.


7


. Scan line G is added to accomplish this. A common DATA LINE is used for four pixels


72


.

FIG. 15

illustrates an example of a timing diagram for circuit


70


.




It is understood that the circuits described above may be implemented on a semiconductor device. Advantageously, circuit


30


yields a high aspect ratio and provides a reduction in number of both data and gate drivers when compared to the prior art X-Y addressed displays. Table 1 compares the number of lines needed for a display.




Table 1.




X=number of rows




Y=number of columns (i.e., RGB columns)




R


xy


=Y/X information content aspect ratio of a panel (typically 4:3).

















TABLE 1













total #




total #










of array




of array










connec-




connec-







# of




# of




total #




tions for




tions for






Display




horizontal




vertical




of array




aspect




aspect






Address




lines in




lines in




connec-




ration




ration






Type




array




array




tions




4:3




16:9











Prior art




  X




3Y or




(1 + 3R


xy


)*X




  5X




˜6.333X








3R


xy


*X






Prior




1.5X




2Y or




(1.5 + 2R


xy


)*X




˜4.167X




˜5.056X






Art with





2R


xy


*X






delta-






RGB






pixel






layout






Circuit




  2X + 1




1.5Y or




(2 + 1.5R


xy


)*X




  4X + 1




˜4.667 +






40





1.5R


xy


*X






1











X = number of rows










Y = number of columns (i.e., RGB columns)










R


xy


= Y/X information content aspect ratio of a panel (typically 4:3).













Table 1 illustratively demonstrates a reduction the present invention yields for vertical lines. The result is a 2:1 data demultiplexing effect. The total number of horizontal lines in the display may be increased, however, if this is the case the total number of connections (lines for tabbing external drivers) for the 4:3 or the 16:9 aspect ratio displays are minimized in accordance with the present invention. Further improvements in demultiplexing may be realized if the fanout structure as shown in

FIG. 9

is employed. Gate drivers are further reduced in numbers and the data demultiplexing becomes m:1 where m is an integer greater than one and equal to the number of stages implemented as described in FIG.


9


.




Referring to

FIG. 16

, an illustrative layout is shown for an implementation of the present invention. Pixels


110


are shown. Between rows of pixels a circuit region


112


is provided in accordance with the present invention. Circuit region


112


may include TFTs, scan lines, data lines, storage node lines, connections, etc. as described. Alternately, circuit region


112


may include logic circuitry including AND gates or other logic gates such as OR, NOR, NAND and/or XOR gates. Thus, the selection of control lines such as scan lines will be performed using logic gates having control or enable signals which are logically combined to transmit an appropriate control signal to TFTs. Logic gates may also be included externally to the pixel areas.




Pixels


110


may include a transmissive mode and a reflective mode. Transmissive mode includes modulating light from a surface of pixel


110


by modulating a capacitive voltage to the pixel to transmit light directly therefrom. Reflective mode includes preparing the pixel to modulate light therefrom by reflecting light incident on its surface.




It is to be understood that the present invention may be implemented with various semiconductor technologies, for example crystalline silicon, amorphous silicon, polysilicon, organic materials, Si—Ge and/or CdS. The embodiments of the present invention may be implemented on any active matrix display without impacting conventional fabrication processes. In preferred embodiments, the displays are used in lap top computers or other electronic devices having LCDs. Further, the present invention implements multiplexing/demultiplexing capability while reducing components and cost. Layouts may have reduced area where the transistors are disposed on a substrate and the pixels are formed over the transistors in accordance with the invention.




Having described preferred embodiments of multiplexing pixel circuits (which are intended to be illustrative and not limiting), it is noted that modifications and variations can be made by persons skilled in the art in light of the above teachings. It is therefore to be understood that changes may be made in the particular embodiments of the invention disclosed which are within the scope and spirit of the invention as outlined by the appended claims. Having thus described the invention with the details and particularity required by the patent laws, what is claimed and desired protected by Letters Patent is set forth in the appended claims.



Claims
  • 1. An active matrix display comprising:a plurality of pixels arranged in an array, wherein the array comprises rows and columns and control lines select the pixels in different rows simultaneously; at least two transistors associated with each pixel, the transistors serially connected to each other and disposed within the array for switching the pixels on and off according to data signals and gate signals; and a data line coupled to a first end of the serially connected transistors and shared between the transistors of different pixels; a second end of the serially connected transistors coupling to a storage device and the pixel; the serially connected transistors providing multiplexing for at least one data signal multiplexing and gate signal multiplexing of each pixel by selectively enabling the at least two transistors for each pixel in accordance with the gate signals.
  • 2. The display as recited in claim 1, wherein one of the at least two transistors is shared between adjacent pixels.
  • 3. The display as recited in claim 1, wherein the pixels modulate light in a transmissive mode.
  • 4. The display as recited in claim 1, wherein the pixels modulate light in a reflective mode.
  • 5. The display as recited in claim 1, wherein the control lines include data lines and the simultaneously selected pixels share a data line.
  • 6. The display as recited in claim 1, wherein the control lines include data lines and the simultaneously selected pixels each use a different data line.
  • 7. The display as recited in claim 1, further comprising scan lines for connecting to gates of the transistors for activating the transistors.
  • 8. The display as recited in claim 7, wherein the scan lines include capacitance storage lines.
  • 9. The display as recited in claim 1, wherein control lines are coupled to the transistors by a low impedance path.
  • 10. The display as recited in claim 9, wherein the low impedance path includes one of a metal, a doped amorphous silicon and a polycrystalline silicon.
  • 11. The display as recited in claim 9, wherein the low impedance path includes a capacitor.
  • 12. The display as recited in claim 1, wherein the display includes a liquid crystal display.
  • 13. The display as recited in claim 1, wherein the transistors include thin film transistors.
  • 14. The display as recited in claim 1, further comprising logic circuitry for controlling the multiplexing in accordance with control signals.
  • 15. The display as recited in claim 1, wherein the transistors are disposed on a substrate and the pixels are formed over the transistors.
  • 16. The display as recited in claim 15, wherein the transistors are disposed on a substrate and the pixels are formed over the transistors.
  • 17. An active matrix display comprising:a plurality of pixels arranged in an array including rows and columns; at least two trnasistors associated with each pixel, the transistors being serially connected and positioned within the array for switching the pixels on and off; a plurality of data lines running substantially parallel to the columns; a plurality of scan lines running substantially parallel to the rows; and the data lines and scan lines being coupled to the transistors of the pixels and shared between the transistors of different pixels such that the data lines provide data multiplexing for each pixel and the scan lines provide gate multiplexing for each pixel by selectively enabling the at least two transistors for each pixel in accordance with data signals on the data lines and gate signals on the gate lines, wherein the data signal multiplexing comprises L:1 multiplexing, and the gate signal multiplexing comprises M:1 multiplexing where L and M each comprise an integer greater than one, respectively.
  • 18. The display as recited in claim 17, wherein one of the at least two transistors is shared between adjacent pixels.
  • 19. The display as recited in claim 17, wherein the pixels modulate light in a transmissive mode.
  • 20. The display as recited in claim 17, wherein the pixels modulate light in a reflective mode.
  • 21. The display as recited in claim 17, wherein control lines select pixels in different rows simultaneously.
  • 22. The display as recited in claim 17, wherein the pixels share a data line.
  • 23. The display as recited in claim 17, wherein the pixels each have a different data line.
  • 24. The display as recited in claim 17, wherein the scan lines include capacitance storage lines.
  • 25. The display as recited in claim 17, wherein the control lines are coupled to the transistors by a low impedance path.
  • 26. The display as recited in claim 25, wherein the low impedance path includes a capacitor.
  • 27. The display as recited in claim 17, wherein the display includes a liquid crystal display.
  • 28. The display as recited in claim 17, wherein the transistors include thin film transistors.
  • 29. The display as recited in claim 17, further comprising logic circuitry for controlling the multiplexing in accordance with control signals.
  • 30. The display as recited in claim 17, wherein the array comprises rows and columns and control lines select the pixels in different rows simultaneously.
  • 31. The display as recited in claim 30, wherein the control lines include data lines and the simultaneously selected pixels share a data line.
  • 32. The display as recited in claim 30, wherein the control lines include data lines and the simultaneously selected pixels each use a different data line.
  • 33. An active matrix display comprising:a plurality of pixels arranged in an array; at least two transistors associated with each pixel, the transistors serially connected to each other and disposed within the array for switching the pixels on and off according to data signals and gate signals; and a data line coupled to a first end of the serially connected transistors and shared between the transistors of different pixels; a second end of the serially connected transistors coupling to a storage device and the pixel; the serially connected transistors providing multiplexing for at least one of data signal multiplexing and gate signal multiplexing of each pixel by selectively enabling the at least two transistors for each pixel in accordance with the gate signals, wherein the data signal multiplexing comprises L:1 multiplexing, and the gate signal multiplexing comprises M:1 multiplexing where L and M each comprise an integer greater than one, respectively.
  • 34. The display as recited in claim 33, wherein the array comprises rows and columns and control lines select the pixels in different rows simultaneously.
US Referenced Citations (11)
Number Name Date Kind
5120964 Wieczorek Jun 1992 A
5122792 Stewart Jun 1992 A
5175446 Stewart Dec 1992 A
5581273 Yoneda et al. Dec 1996 A
5641974 den Boer et al. Jun 1997 A
5691783 Numao et al. Nov 1997 A
5844538 Shiraki et al. Dec 1998 A
5903249 Koyama et al. May 1999 A
5926158 Yoneda et al. Jul 1999 A
6023308 Takemura Feb 2000 A
6104067 Zhang et al. Aug 2000 A