Active matrix liquid crystal display

[0002] 1. Field of the Invention

[0003] The present invention relates to an active matrix liquid crystal display and, more particularly, to an active matrix liquid crystal display consuming less electric power than a conventional display.

[0004] 2. Description of the Related Art

[0005] In an active matrix liquid crystal display, a pixel is disposed at each intersection in a matrix construction. Every pixel is equipped with a switching device. Information about the pixels is represented by turning on and off the switching devices. A liquid crystal is used as a display medium in such a display device. In the present invention, thin-film transistors (TFTs) each having three terminals, i.e., gate, source, and drain, are used as switching devices.

[0006] In the present specification, a row of the matrix construction means scanning lines (gate lines) which extend parallel to the row and are connected with the gate electrodes of the TFTs in the row. A column of the matrix construction means signal lines (source lines) which run parallel to the column and are connected with the source (or drain) electrodes of the TFTs in the column. A circuit for driving the scanning lines is referred to as a scanning line driver circuit. A circuit for driving the signal lines is referred to as a signal line driver circuit.

[0007] FIG. 2 shows one conventional active matrix liquid crystal display. A signal line driver circuit 21 is mounted at the top and a scanning line driver circuit 22 is mounted on the left side to drive signal lines 23 and scanning lines 24, respectively. The scanning line driver circuit 22 and the signal line driver circuit 21 receive signals such as clock pulses from a signal-generating circuit such as a clock generator.

[0008] A scanning line driver circuit 22 using shift registers 35 as shown in FIG. 3(a) is normally used. Whenever a clock pulse (CL1,CL2) is entered, the output pulse is shifted by one position. The output pulse is fed to one scanning line 32 via a NAND gate 33 and a buffer circuit 34. In this way, the scanning lines 32 are successively driven. FIG. 3(b) shows timing charts of the scanning line driver circuit 22.

[0009] In the case of a video graphics array (VGA), each scanning line is scanned in about 31 μs.

[0010] One example of the signal line driver circuit 21 is shown in FIG. 4. Signal line driver circuits 21 normally use shift registers 41 in the same way as scanning line driver circuits 22. However, the signal line driver circuit 21 does not directly drive signal lines 44, unlike a scanning line driver circuit 22. The output signal from a shift register 41 drives a sampling analog switch 43 via a buffer circuit 42. The analog video signal 45 is sampled and fed to the signal lines 44.

[0011] In the case of VGA, the ideal sampling time is about 40 nsec. Where the signal line driver circuit 21 is composed of TFTs, the sampling time is set to 320 nsec or 640 nsec, taking account of the performance of the TFTs. In this case, 4-phase or 8-phase clock pulses which are shifted from each other in phase by 40 nsec are used.

[0012] Another conventional active matrix liquid crystal display is described in Japanese Patent Laid-Open No. 186281/1992 and shown in FIG. 5. In this construction, signal lines 50 are divided into plural groups. The signal lines 50 are driven from both ends of the display device. Since the load capacitance and load resistance for signals are halved, it is easy to drive the signal lines 50.

[0013] Examples of commercial products using active matrix liquid crystal displays include notebook computers and portable intelligent terminals. These commercial products are required to be driven by batteries. However, the service time of the existing active matrix liquid crystal display is limited by the amount of electric power consumed by the display. Accordingly, it is important to reduce the electric power consumed by the active matrix liquid crystal display in obtaining a longer service time.

[0014] The current worldwide trend is toward saving of resources. Active matrix liquid crystal displays which are considered to be promising next-generation display devices must accomplish lower power consumption.

[0015] One conceivable method of reducing the electric power consumption is to reduce the applied voltage or the operating frequency. However, this method deteriorates the performance. Therefore, a method of reducing the electric power consumption while maintaining the performance has been sought for.

[0016] In the conventional method described already in connection with FIG. 5, the signal line capacitance is halved. Let P2 be the electric power consumed when signal lines are driven. Let P1 be the electric power consumed when signal lines are driven by the construction described previously in connection with FIG. 2. The following relations hold:

[0017] P2=C1/2×V2×f

[0018] P1/2

[0019] where C1 is the capacitance of the signal lines, V is the amplitude of the signal, and f is the operating frequency.

[0020] In this way, the electric power consumed by the configuration shown in FIG. 5 can be halved compared with the electric power consumed by the configuration shown in FIG. 2. However, two driver circuits are required to be disposed at opposite ends, respectively, of the display device. Therefore, the total electric power consumed by the driver circuits is doubled compared with the electric power consumed by the driver circuit shown in FIG. 2. Hence, the electric power consumed is increased accordingly. With respect to the driver circuits, the load is halved. However, each driver circuit must have the same number of stages of shift registers as the stages of shift registers of the driver circuit in the configuration shown in FIG. 2. Therefore, the electric power for driving the shift registers is doubled. Furthermore, the electric power needed to drive the common clock terminal for applying clock pulses to the shift registers is doubled. In addition, the electric power required to drive the video signal input terminal is doubled. These electric powers are comparable to or greater than the electric power for driving the signal lines.

[0021] The method described in the above-cited Japanese Patent Laid-Open No. 186281/1992 was originally developed to drive a large area display. Therefore, this method is a disadvantage in reducing electric power consumption.

[0022] It is an object of the present invention to provide an active matrix liquid crystal display having signal lines which create a frame of image on a viewing screen and are divided into upper and lower groups to reduce the electric power consumed by the load capacitance of the signal lines, in the same way as in the conventional configuration shown in FIG. 5.

[0023] We have noticed that when the viewing screen is being scanned successively from top to below, e.g., when the upper half of the screen is being scanned, the driver circuit for the lower half is not required to be driven. Accordingly, the invention provides a means for halting the driver circuit for the lower signal lines or putting this driver circuit on standby. Obviously, when the lower half of the display screen is being scanned, the driver circuit for the upper signal lines is halted or put on standby.

[0024] Other objects and features of the invention will appear in the course of the description thereof, which follows.

[0025] FIG. 1 is a block diagram of an active matrix liquid crystal display according to the invention;

[0026] FIG. 2 is a block diagram of a known active matrix liquid crystal display;

[0027] FIG. 3(a) is a diagram of a general scanning line driver circuit;

[0028] FIG. 3(b) is a time chart explaining the operation of the circuit shown in FIG. 3(a);

[0029] FIG. 4 is a diagram of a general signal line driver circuit;

[0030] FIG. 5 is a block diagram of another known active matrix liquid crystal display;

[0031] FIG. 6(a) is a circuit diagram of a state-switching circuit for alternately halting two signal line driver circuits according to the invention;

[0032] FIG. 6(b) is a time chart for explaining the operation of the circuit shown in FIG. 6(a);

[0033] FIG. 7 is a circuit diagram of another state-switching circuit for alternately deactivating two signal line driver circuits according to the invention;

[0034] FIG. 8 is a circuit diagram of signal line driver circuits using decoder circuits according to the invention;

[0035] FIG. 9 is a circuit diagram of a further state-switching circuit for stopping supply of an address signal to two signal line driver circuits alternately according to the invention;

[0036] FIGS. 10(A)-10(D) are cross-sectional views of an active matrix liquid crystal display, illustrating some low-tempera ture polysilicon process steps for fabricating the display;

[0037] FIGS. 11(A) and 11(B) are cross-sectional views, illustrating following low-temperature polysilicon process steps;

[0038] FIGS. 12(A)-12(D) are cross-sectional views, illustrating high-temperature polysilicon process steps carried out after the steps illustrated in FIGS. 11(A) and 11(B); and

[0039] FIGS. 13(A) and 13(B) are cross-sectional views, illustrating following high-temperature polysilicon process steps.

[0040] Referring to FIG. 1, there is shown an active matrix liquid crystal display embodying the concept of the present invention. This liquid crystal display has signal lines 10 which drive an upper pixel matrix construction 11aand a lower pixel matrix construction 11bon opposite sides of the center line of the frame of image displayed on the viewing screen of the liquid crystal display. The signal lines 10afor the upper pixel matrix 11aare driven by an upper signal line driver circuit 12a. The signal lines 10bfor the lower pixel matrix 11bare driven by a lower signal line driver circuit 12b. The liquid crystal display further includes a scanning line driver circuit 13 and a state-switching circuit 14 for halting the two signal line driver circuits 12a,12balternately.

[0041] FIG. 6 shows an example of the state-switching circuit according to the invention. In this example, clock pulses 60 to be applied to the scanning line driver circuits are first supplied to the state-switching circuit. This state-switching circuit comprises a frequency division circuit 61, an upper AND gate 62aconnected to the output of the frequency division circuit 61, an inverter 63 connected to the output of the frequency division circuit 61, and a lower AND gate 62bconnected to the inverter 63. In the case of a VGA (video graphics array), the frequency division circuit 61 is a divided-by-240 frequency division circuit. The output of the upper AND gate 62ais connected with the clock input terminal 64aof the upper signal line driver circuit 12a. Similarly, the output of the lower AND gate 62bis connected with the clock input terminal 64bof the lower signal line driver circuit 12b. This switching circuit 14 controls the clock pulses 65 supplied to the signal line driver circuits 12.

[0042] As shown in timing charts of in FIG. 6(b), when the upper half of the image is being scanned, the clock pulses 65 are prevented from entering the lower signal line driver circuit 12b. When the lower half of the frame of image is being scanned, the clock pulses 65 are prevented from entering the upper signal line driver circuit 12a. By the addition of this switching circuit 14, unwanted consumption of electric power by one signal line driver circuit 12 can be eliminated.

[0043] FIG. 7 shows another example of the state-switching circuit 14 according to the invention. Electric power to be supplied to two signal line driver circuits 70a, 70bis switched between scanning of the upper half of the frame of image on the viewing screen and the scanning of the lower half. In the same way as the method described already in connection with FIG. 6, a signal for switching between the upper half of the frame of image and the lower half is used to deactivate shift registers alternately.

[0044] FIG. 8 shows an example in which decoder circuits are used in signal line driver circuits.

[0045] FIG. 9 shows a further example of the state-switching circuit 14 according to the invention. The operation of the signal line driver circuit which is not presently used can be stopped by cutting off supply of an address signal to the decoder circuit.

[0046] A method of fabricating TFT substrates of a liquid crystal display, using an active matrix circuit according to the invention, is described below.

[0047] FIGS. 10(A)-10(D) and 11(A)-11(B) illustrate low-temperature polysilicon process steps for fabricating a monolithic active matrix circuit of the present example. The process sequence for fabricating TFTs forming a peripheral logic circuit is shown on the left sides of FIGS. 10(A)-10(D) The process sequence for fabricating the active matrix circuit is shown on the right sides. First, a silicon oxide film 1002 is formed as a buffer oxide film 1002 on a glass substrate 1001 to a thickness of 1000 to 3000 Å. This silicon oxide film may be formed in an oxygen ambient by sputtering or plasma CVD.

[0048] Then, an amorphous silicon film is formed to a thickness of 300 to 1500 Å, preferably 500 to 1000 Å, by plasma CVD or LPCVD. The amorphous film is thermally annealed at a temperature higher than 500° C., preferably 500-600° C., to crystallize the amorphous silicon film or to enhance the crystallinity. After the crystallization, the crystallinity may be further enhanced by carrying out photo-annealing making use of laser light. Furthermore, during the crystallization making use of thermal annealing, an element (or, a catalytic element) such as nickel for promoting crystallization of silicon may be added, as described in Japanese Patent Laid-Open Nos. 244103/1994 and 244104/1994.

[0049] Then, the silicon film is etched to form islands of an active layer 1003 for P-channel TFTs forming a driver circuit, islands of an active layer 1004 for N-channel TFTs, and islands of an active layer 1005 for pixel TFTs forming a matrix circuit. Furthermore, a gate-insulating film 1006 of silicon oxide is formed to a thickness of 500 to 2000 Å by sputtering in an oxygen ambient. The gate-insulating film may be formed by plasma CVD. Where the silicon oxide film is formed by plasma CVD, it is desired to use nitrogen monoxide (N₂O) as a gaseous raw material. Alternatively, oxygen (O₂) and monosilane (SiH₄) may be employed.

[0050] Subsequently, an aluminum layer having a thickness of 2000 to 6000 Å is formed by sputtering over the whole surface of the laminate. The aluminum may contain silicon, scandium, palladium, or other material to prevent generation of hillocks in later thermal processing steps. The gate-insulating film 1006 is etched to form gate electrodes 1007, 1008, and 1009 (FIG. 10(A)).

[0051] Thereafter, the aluminum layer is anodized to form aluminum oxide, 1010, 1011, and 1012, on the surface of the aluminum layer. These aluminum regions act as insulator (FIG. 10(B)).

[0052] Then, a photoresist mask 1013 which covers the active layer of the P-channel TFTs is formed. Phosphorus ions are introduced by ion doping while using phosphine as a dopant gas. The dose is 1×10¹²to 5×10¹³atoms/cm². As a result, heavily doped N-type regions, or source 1014 and drain 1015, are formed (FIG. 10(C)).

[0053] Thereafter, a photoresist mask 1016 for covering both active layer for the N-channel TFTs and active layer for the pixel TFTs is formed. Boron ions are introduced again by ion doping, using diborane (B₂H₆) as a dopant gas. The dose is 5×10¹⁴to 8×10₁₅atoms/cm². As a result, P-type regions 1017 are formed. Because of the doping steps described thus far, heavily doped N-type regions (source and drain 1014 and 1015) and heavily doped P-type regions (source and drain 1017) are formed (FIG. 10(D)).

[0054] Then, the laminate is thermally annealed at 450-850° C. for 0.5 to 3 hours to repair the damage created by the doping. In this way, the dopants are activated. At the same time, the crystallinity of the silicon is recovered. Thereafter, as shown in FIG. 11(A), a silicon oxide film having a thickness of 3000 to 6000 Å is formed as an interlayer dielectric 1018 over the whole surface by plasma CVD. This may be a silicon nitride film or a multilayer film of silicon oxide layers and silicon nitride layers. The interlayer dielectric 1018 is etched by a wet etching process or a dry etching process to form contact holes in the source/drain regions.

[0055] Then, an aluminum film or a multilayer film of titanium and aluminum is formed to a thickness of 2000 to 6000 Å by sputtering techniques. This film is etched so as to create electrodes/interconnects, 1019, 1020, and 1021, for a peripheral circuit and pixels/interconnects, 1022 and 1023, for pixel TFTs (FIG. 11(A)).

[0056] Subsequently, a silicon nitride film 1024 is formed as a passivation film having a thickness of 1000 to 3000 Å by plasma CVD. This silicon nitride film is etched to create contact holes extending to the electrodes 1023 of the pixel TFTs. An ITO (indium-tin oxide) film having a thickness of 500 to 1500 Å is formed by sputtering. Finally, the ITO film is etched to form pixel electrodes 1025. In this manner, the peripheral driver circuit and active matrix circuit are formed integrally (FIG. 11(B)).

[0057] The process sequence of the present example is described by referring to FIGS. 12(A)-12(D), taking as an example a high-temperature process for fabricating silicon gate polysilicon TFTs.

[0058] The process sequence for fabricating the TFTs forming the peripheral logic circuit is shown on the left sides of FIGS. 12(A)-12(D). The process sequence for fabricating the active matrix circuit is shown on the right sides. First, a silicon oxide film is formed as a buffer oxide film 1102 on a quartz substrate 1101 to a thickness of 1000 to 3000 Å. This silicon oxide film may be formed in an oxygen ambient by sputtering or plasma CVD.

[0059] Then, an amorphous or polycrystalline silicon film is formed to a thickness of 300 to 1500 Å, preferably 500 to 1000 Å, by plasma CVD or LPCVD. The silicon film is thermally annealed at a temperature higher than 500° C., preferably 800-950° C., to crystallize the silicon film or to enhance the crystallinity. After the crystallization, the crystallinity may be further enhanced by carrying out photoannealing. Furthermore, during the crystallization making use of thermal annealing, an element (or, a catalytic element) such as nickel for promoting crystallization of silicon may be added, as described in Japanese Patent Laid-Open Nos. 244103/1994 and 244104/1994.

[0060] Then, the silicon film is etched to form islands of an active layer 1103 for P-channel TFTs forming a driver circuit, islands of an active layer 1104 for N-channel TFTs, and islands of an active layer 1105 for pixel TFTs forming a matrix circuit. Furthermore, a gate-insulating film 1106 of silicon oxide is formed to a thickness of 500 to 2000 Å by sputtering in an oxygen ambient. The gate-insulating film may be formed by plasma CVD. Where the silicon oxide film is formed by plasma CVD, it is desired to use nitrogen monoxide (N₂O) as a gaseous raw material. Alternatively, oxygen (O₂) and monosilane (SiH₄) may be employed.

[0061] Subsequently, a polycrystalline silicon film having a thickness of 2000 Å to 5 μm, preferably 2000 to 6000 Å, is formed by LPCVD over the whole surface of the laminate. To enhance the electrical conductivity, a trace amount of phosphorus is added to the polycrystalline silicon film. This polysilicon film is etched to form gate electrodes 1107, 1108, and 1109 (FIG. 12(A)).

[0062] Then, phosphorus ions are introduced into all the islands of the active layers by self-aligned ion implantation techniques, using phosphine (PH₃) as a dopant gas. At this time, the gate electrodes are used as a mask. The dose is 1×10¹²to 5×10₁₃atoms/cm². As a result, lightly doped N-type regions 1110, 1111, and 1112 are formed (FIG. 12(B)).

[0063] Thereafter, a photoresist mask 1113 for covering the active layer for the P-channel TFTs is formed. Another photoresist mask 1114 for covering the active layer for the pixel TFTs up to the portions which are spaced 3 μm from the ends of the gate electrodes is formed. Phosphorus ions are introduced again by ion doping, using phosphine (PH₃) as a dopant gas. The dose is 1×10¹²to 5×10₁₃atoms/cm². As a result, heavily doped N-type regions, or source and drain, 1115 and 1116, are formed. Those regions of the lightly doped N-type regions of the active layer for the pixel TFTs which are capped with the mask are not implanted with phosphorus ions at this time and so these regions remain lightly doped N-type (FIG. 12(C)).

[0064] Then, a photoresist mask 1117 for covering the active layer for the N-channel TFTs is formed. Boron ions are introduced again by ion doping, using diborane (B₂H₆) as a dopant gas. The dose is 5×10¹⁴to 8×10¹⁵atoms/cm². Consequently, the dose of boron is in excess of the dose of phosphorus. The previously formed, lightly doped N-type regions turn into heavily doped P-type regions 1118. As a result of these doping steps, heavily doped regions (source and drain, 1115 and 1116), heavily doped P-type regions (source and drain, 1118), and the lightly doped N-type region 1112 are formed (FIG. 12(D)).

[0065] Then, the laminate is thermally annealed at 450-850° C. for 0.5 to 3 hours to repair the damage created by the doping. In this way, the dopants are activated. At the same time, the crystallinity of the silicon is recovered. Thereafter, as shown in FIG. 13(A), a silicon oxide film having a thickness of 3000 to 6000 Å is formed as an interlayer dielectric 1119 over the whole surface by plasma CVD. This may be a silicon nitride film or a multilayer film of silicon oxide layers and silicon nitride layers. The interlayer dielectric 1119 is etched by a wet etching process or a dry etching process to form contact holes in the source/drain regions. Then, an aluminum film or a multilayer film of titanium and aluminum is formed to a thickness of 2000 to 6000 Å by sputtering. This film is etched so as to create electrodes/interconnects, 1120, 1121, and 1122, for a peripheral circuit and electrodes/interconnects, 1123 and 1124, for pixel TFTs (FIG. 13(A)).

[0066] Subsequently, a silicon nitride film 1125 is formed as a passivation film having a thickness of 1000 to 3000 Å by plasma CVD. This silicon nitride film is etched to create contact holes extending to the electrodes 1124 of the pixel TFTs. An ITO (indium-tin oxide) film having a thickness of 500 to 1500 Å is formed by sputtering. Finally, the ITO film is etched to form pixel electrodes 1126. In this manner, the peripheral driver circuit and active matrix circuit are formed integrally (FIG. 13(B)).

[0067] In the examples described thus far, the driver circuit and the pixel matrix circuit can be formed integrally. Therefore, if two separate signal line driver circuits for activating two sets of signal lines, respectively, which are assigned to upper and lower halves, respectively, of the frame of image displayed on the viewing screen are provided, a large area is not needed. Hence, a liquid crystal display of reduced size can be accomplished. Furthermore, since the signal lines are vertically divided into two groups, the load capacitance and load resistance of the signal lines are halved. As a consequence, the display device can be driven with small driving capability and in a short time. The driver circuits can be built with a point-at-a-time scanning system. This dispenses with analog buffer and large sample-and-hold capacitors which would be normally necessitated where the line-at-a-time scanning drive method is adopted. Consequently, the area occupied by the driver circuits themselves can be reduced. This is a further advantage to miniaturization.

[0068] In the above example, the driver circuits are of the monolithic construction. The invention can also be applied to a display device comprising an active matrix circuit consisting of amorphous TFTs, together with outside driver circuits attached to the outside of glass substrates.

[0069] As described thus far, in the present invention, signal lines are divided into two sets corresponding to the upper and lower halves of the frame of image displayed. The two sets of signal lines are driven by two signal line driver circuits, respectively. When one of the two driver circuits is operating, operation of the other is stopped. As a consequence, a great reduction in the electric power consumed can be accomplished.