Method of reducing wordline recovery time

This application is related to co-pending patent application titled, “SYSTEM FOR REDUCING WORDLINE RECOVERY TIME” by Pan filed Dec. 22, 2006 having application Ser. No. 11/615,873 that is incorporated by reference in its entirety for all purposes.

The present invention relates generally to memory devices. More particularly, the present invention relates to using charge pumps operating in a highly efficient mode to reduce wordline recovery time.

As microprocessor systems increase in size and speed, there is need for larger and faster memory arrays. These high-speed memory arrays may contain a large number of memory cells. However, as the number of memory elements increases, the time needed to read and/or write to the individual memory cells may also increase. This may be due to the fact that as the number of memory elements increases in an individual array, the length of the wordline between the supply voltage and the individual memory cells may also increase. The increased length of the wordline may directly relate to an increase in the resistance of the wordline. Therefore, as the size of the memory array increases so may the required voltage needed to read the individual memory cells. Additionally, in large memory arrays, the capacitance of the wordline may require an increase in the time required charge to the desired level in order to access a given memory cell. The increase in the time required to charge the wordline may lead to a large cycle time, which may directly limit the access time for the memory. Additionally, the increase in the time required to charge the capacitance may also limit the length of the wordline. As the length of the wordline increases so does the resistance value of the wordline wire. Therefore, the time required to charge the capacitance may be limited by the large RC value associated with a longer wire.

For example, FIG. 1A shows a representative prior art arrangement whereby wordline 100 having wordline capacitance C wl is coupled to high voltage charge pump 102 by way of switch circuit 104 . As shown in FIG. 1B , at time t 1 , charge pump 102 goes into regulation providing pump current I pump at charge pump output node 106 . In order for wordline voltage V wl to recover to V reg , switch circuit 104 connects wordline 100 to charge pump output node 106 at time t 2 (causing the momentary decrease in V reg due to charge sharing). At time t 3 , wordline voltage V wl reaches target wordline voltage V reg . In this arrangement, the wordline recovery time (Δt=t 3 −t 2 ) is determined as Eq(1):

Δt =( C wl ×V reg )/ I pump   Eq(1)

There have been many attempts at improving the wordline recovery time Δt. One such attempt relies upon using what is referred to as stacked memory cell in which a large single array is broken into a number of smaller arrays thereby reducing the resistive/capacitive loading (i.e., RC time constant) of the array as a whole. However, this approach requires a substantial increase in the amount of silicon used since the amount of silicon required increases proportionally with the number of memory arrays included in the stacked architecture as does manufacturing costs. Other attempts to reduce wordline time recovery rely upon simply increasing charge pump current I pump . Unfortunately, in order to increase the capacity of high voltage charge pump 102 , not only does the size of high voltage charge pump 102 increase but charge pump 102 operates at a higher output voltage V ref and is therefore operating less efficiently. Each of these factors contribute in their own way to an overall increase in power consumption of the memory device as well as the amount of silicon required for fabrication of the larger capacity charge pumps.

Therefore, what is required is a method of reducing wordline recovery time.

The invention can be implemented in numerous ways. Several embodiments of the invention are discussed below. As a method of increasing a wordline voltage level V WL of a wordline having a wordline capacitance V WL from an initial wordline voltage level V WLI to a pre-determined target voltage level V WLT is described. One embodiment of the invention includes at least the acts of: configuring the wordline capacitance C WL and a load capacitance C L in a series configuration with respect to a first node; supplying a charging current at the first node; and charging up the wordline capacitance C WL by way of the charging current thereby increasing the wordline voltage level.

In a memory device having a number of wordlines, a method of substantially reducing an amount of time required for a wordline voltage level (V WL ) to reach a pre-determined operating voltage level, one embodiment of the invention includes at least the acts of: selecting one of the wordlines having a wordline capacitance C WL ; signaling a first switch to connect the selected wordline to a wordline voltage supply through a load capacitance C L at a first node; signaling a second switch to disconnect the selected wordline from the wordline voltage supply at a second node in conjunction with the first switch resulting in load capacitance C L being in series with wordline capacitance C WL with respect to the wordline supply voltage; applying a wordline charging voltage by the wordline voltage supply at the first node; and increasing the wordline voltage V WL by charging up a wordline capacitance (C WL ) in series with a load capacitance (C L ) by way of the applied wordline charging voltage.

In the described embodiments the wordline is connected to non-volatile memory cells arranged to form a non-volatile memory array suitable for storing data. The non-volatile memory array is arranged in a NAND-type memory array architecture having a number of wordlines and bitlines. Additionally, the described method is contemplated for use on a multilevel type memory array that when programmed stores data in the form of at least one lower page and at least one associated upper page.

The invention, together with further objects and advantages thereof, may best be understood by reference to the following description taken in conjunction with the accompanying drawings.

FIG. 1 shows a representative prior art arrangement whereby a wordline having wordline capacitance C wl is coupled to charge pump by way of switch circuit.

FIG. 2 illustrates an exemplary MOSFET having a floating gate.

FIG. 3 illustrates an exemplary non-volatile memory string.

FIGS. 4A-4C illustrates a representative wordline charging circuit in accordance with an embodiment of the invention.

FIG. 5 shows operating regimes of output current capabilities for a high output voltage and low output voltage charge pumps in accordance with an embodiment of the invention.

FIG. 6 shows a graph describing a relationship between amount of charge ΔQ at a particular load L provided by the high voltage and low voltage charge pumps of FIG. 5 as a function of charge pump output voltage V and time t.

FIG. 7 shows a flowchart detailing a process of substantially reducing an amount of time required for a wordline voltage level to reach a pre-determined operating voltage level.

In the drawings, like reference numerals designate like structural elements. Also, it should be understood that the depictions in the figures are not to scale.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without some or all of these specific details. In the description below, a memory cell is programmed using a next neighbor wordline as an injector for providing hot carriers. The objective is to use low voltage biasing to prevent and therefore substantially eliminate programming problems related to high voltage, small geometries, and resultant high electric fields.

With reference to FIGS. 2 and 3 , a representative non-volatile memory storage array will be described. In the described embodiment, the non-volatile memory storage system is a flash memory chip having a NAND architecture, although the present invention is applicable to other forms of non-volatile memory including EPROM and EEPROM, and additionally to NOR architectures as well. In the described embodiment, the array consists of non-volatile memory storage elements that are p-type substrate floating gate MOSFETs. FIG. 2 illustrates a representative floating gate MOSFET, such as that used in this architecture, having a substrate 202 , a source 204 , a drain 206 , a control gate 208 , a floating gate 210 , and a dielectric 212 surrounding the floating gate.

In NAND flash memory, a number of such transistors, also known hereafter as cells, memory cells or memory storage elements, are arranged in series. On either side of the series of memory cells is an additional transistor known as a select transistor, or gate. Collectively, the memory cells and two select gates are referred to as a NAND string. By way of illustration, the equivalent circuit of a 4-transistor NAND string is illustrated in FIG. 3 . The four memory cells are labeled 306 , 308 , 310 and 312 . The first select gate 304 , referred to as the select gate drain (SGD), connects the NAND string to bitline 302 . The second select gate 314 , referred to as the select gate source (SGS), connects the NAND string to source line 316 . Select gates 304 and 314 are controlled by control gates 318 and 336 , respectively. An SGD line controls the control gate 318 for the select gate 304 , while an SGS line controls the control gate 336 for the select gate 314 . The four series transistors 306 - 312 each have both a control gate and a floating gate. For example, transistor 306 has a control gate 320 and a floating gate 322 ; transistor 308 has a control gate 324 and a floating gate 326 ; transistor 310 has a control gate 328 and a floating gate 330 ; and transistor 312 has a control gate 332 and a floating gate 334 . Control gates 320 , 324 , 328 and 332 are connected to wordlines WL( 3 ), WL( 2 ), WL( 1 ) and WL( 0 ), respectively. It should be noted that although the example provided describes a NAND string that includes four memory cells, other numbers of cells may be arranged in series and are permissible as well.

FIG. 4A is a block diagram of a circuit 400 for charging a high-speed wordline in accordance with some embodiments of the present invention. The circuit 400 includes a wordline voltage supply 402 for providing a wordline supply voltage V supply . (It should be noted that wordline supply voltage 402 can be implemented as a charge pump or as a voltage supply such as V cc ). In the described embodiment, switch 404 responds to control signal S by connecting wordline supply voltage 402 to wordline 406 at node 1 and switch 414 responds to control signal S by disconnecting wordline supply voltage 402 from node 2 of wordline 406 . In this way, load capacitance C L and wordline capacitance C WL are in series configuration with respect to wordline supply voltage 402 (see FIG. 4B ). It should be noted that generally switch 104 responds to signals provided by a wordline select circuit (such as a wordline decoder) to select wordline 406 to node 106 . In some embodiments of the present invention, switch 404 is a complimentary metal oxide semiconductor (CMOS) inverter that includes a p-type MOS (PMOS) transistor 408 and an n-type MOS (NMOS) transistor 410 . The source of the PMOS transistor 408 is connected to wordline voltage supply 402 , the drain of the PMOS transistor 408 is connected to the drain of the NMOS transistor 410 while the source of the NMOS transistor 410 is connected to the a reference voltage, such as V ss . A control signal line 412 for carrying control signal S is connected to the gate of both the PMOS transistor 408 and the NMOS transistor 410 (it should be noted that the control signal S can be level shifted up). In operation, when the control signal S is a logic “low”, the NMOS transistor 410 is “OFF” and the PMOS transistor 408 may be “ON” and switch 414 is open so that wordline supply voltage 402 is disconnected from wordline 406 at node 2 and connected to the bottom plate of load capacitance C L through switch 404 .

It should be noted that implementing switch 404 as a CMOS inverter provides several advantages. For example, the “HIGH” and “LOW” output levels of control signal S are such that high noise margins can be realized and the logic levels of the switch 404 is not dependent upon the relative device sizes. Furthermore, the CMOS inverter has a low output impedance, which may make it less sensitive to noise and other disturbances.

As shown in FIG. 4B , when switch 404 connects wordline voltage supply 402 to wordline 406 at node 1 and switch 414 disconnects wordline voltage supply 402 from wordline 406 at node 2 , charging current I charge charges up both load capacitance C L and wordline capacitance C wl . In this way, an amount of time required to restore wordline voltage V wl to the target wordline voltage V reg is substantially reduced than that with conventional arrangements (as shown in FIG. 4C ).

It should be noted that in those cases where wordline voltage supply 402 is a charge pump, disconnecting charge pump 402 from wordline 406 at node 2 and re-connecting charge pump 402 to wordline 406 at node 1 , allows charge pump 402 to start charge pump output from approximately 0V which is much more efficient than operating with charge pump output at very high potential. As shown in FIG. 5 , by operating the high efficiency regime (represented as OP 2 ), a charge pump is able to provide a substantially larger amount of charge ΔQ at a given V out for a given load than that provided by a charge pump operating with charge pump output at very high potential (20V, for example) under the same conditions (represented as OP 1 ) as shown in FIG. 6 . Therefore, even operating at a lower V out , the low voltage, high efficiency charge pump provides for higher ΔQ/Δt for a given load thereby substantially reducing wordline recovery time Δt wl over that provided by a high voltage, lower efficiency charge pump, such as charge pump 102 .

In this way, circuit 400 utilizes a low voltage, high efficiency charge pump (or other appropriate voltage supply) to provide the requisite charge ΔQ (i.e, (C wl ×V supply )) required to charge wordline capacitor C w l consistent with wordline voltage V wl being target wordline voltage V reg . The relationship between the various parameters can be described as shown in equations 2-4:

C load =C L ×C wl /( C L +C wl )  Eq(2)

ΔV wl =( C L /(C L +C wl )  Eq(3)

Q total =V supply ×C wl ;  Eq(4)

if C L =C wl , then C load =0.5 C wl and ΔV wl =0.5 V supply .

FIG. 7 shows a flowchart detailing a process 700 of substantially reducing an amount of time required for a wordline voltage level to reach a pre-determined operating voltage level. Accordingly, process 700 includes the following operations: at 702 , providing a wordline charging voltage at a first node and at 704 charging up a wordline capacitance in series with a load capacitance connected in series with the first node. In the described embodiment, the wordline charging voltage is provided by a low voltage, high efficiency charge pump or equivalent and any high voltage, low efficiency charge pumps are disconnected. At 706 , wordline voltage level is compared to a pre-determined wordline voltage level. At 708 , if the wordline voltage level is approximately equal to the pre-determined wordline voltage level, then at 710 , the wordline charging voltage is no longer applied to the first node. In the described embodiment, the high voltage charge pump is then re-connected.

Although only a few embodiments of the invention have been described in detail, it should be appreciated that the invention may be implemented in many other forms without departing from the spirit or scope of the invention. Although specific features and conditions have been described, it should be appreciated that a wide variety of implementations, such as bias conditions and method combinations, may be modified and employed as well. In addition, although in physical construction the present invention is applied to a conventional non-volatile memory system without modification, it should be appreciated by one skilled in the art that the structure could be modified to enhance the effects of the present invention. Accordingly, the present embodiments are to be considered as illustrative and not restrictive and the invention is not to be limited to the details given herein, but may be modified within the scope of the appended claims.