Method for programming a reference cell

Information

  • Patent Grant
  • 6584017
  • Patent Number
    6,584,017
  • Date Filed
    Thursday, April 5, 2001
    23 years ago
  • Date Issued
    Tuesday, June 24, 2003
    21 years ago
Abstract
A method for programming one or more reference cells is described. The reference cell is programmed a predetermined amount, its program state is sensed relative to a prescribed cell on the same die (e.g., a memory cell or a golden bit cell), and the programming process continues until the reference cell fails a preselected read operation. In one preferred embodiment, the memory cell used during the reference cell programming process is the cell in the memory array having the highest native threshold value. In another preferred embodiment, the memory cell used during the reference cell programming process is a native cell that is onboard the die containing the memory array, but not a cell within the memory array.
Description




FIELD OF THE INVENTION




The present invention relates to sensing schemes for read operations on semiconductor devices, and, more particularly, to a method for programming a reference cell for use in a read operation.




BACKGROUND OF THE INVENTION




Memory devices, such as random access memory (RAM), read-only memory (ROM), non-volatile memory (NVM) and like, are known in the art. These devices provide an indication of the data which is stored therein by providing an output electrical signal. A device called a sense amplifier is used for detecting the signal and determining the logical content thereof. U.S. Pat. No. 4,916,671 to Ichiguchi describes one such sense amplifier.




In general, prior art sense amplifiers determine the logical value stored in a cell by comparing the output of the cell with a fixed reference voltage level. The aforementioned U.S. Pat. Nos. 6,134,156 and 6,128,226 describe an alternative circuit architecture in which a reference cell is used to generate a reference signal in lieu of a fixed reference voltage value.




When a memory cell is programmed or erased, the signal it generates differs from the reference signal by some margin. Since reading the cell's state should always result the same (i.e., either programmed or erased depending on the cell's state), introducing such margin is necessary to overcome imperfections in the reading process and to compensate for drifts in the cell's threshold voltage (e.g., caused by retention loss or disturbs). A reduction in the original margin due to imperfections in the reading process (e.g., due to operation at different operational conditions) is referred to as “margin loss.”




It is well understood that the placement of a reference signal to which an array cell signal can be compared during sensing can be achieved in a number of ways. When close to ground signals are sensed as in the aforementioned U.S. Pat. Nos. 6,134,156 and 6,128,226 patents, the reference cell signal develops at an intermediate rate between that of a programmed cell and an erased cell. When set this way, the array cells' signals on one side of the reference signal are determined to be programmed cells, while signals on the other side of the reference signal are determined to be erased cells. For example, array cells generating signals smaller than the reference signal are considered to be programmed and array cells generating signals larger than the reference signal are considered to be erased. Conventionally, such placement is achieved by using a reference cell whose current is between the erased and programmed cells' current levels. The reference cell's current level can be controlled by the reference cell's size, its programming level, or its gate voltage level. Furthermore, if voltage signals are used to detect the cells' contents, then the reference signal placement can be controlled by providing a different load capacitance on the reference cell compared to that of the array cells. However, if the array and the reference cells differ in their sizes, in their operating gate voltages, or in their loads then some margin loss will be introduced to the sensing scheme. On the other hand, placing the reference cells' signals by properly programming the reference cells (while operating the array and reference cells at identical conditions) minimizes the sensing scheme sensitivity to operating conditions, process parameters and environmental conditions variations, thereby minimizing the margin loss, if any, that is introduced to the sensing system.




When reference cell placement is by programming, it must be programmed a precise amount in order to achieve its intended purpose. There are difficulties attendant with reliable programming of a reference cell so as to minimize operating margin loss, as well as accurate placement of a programmed reference cell relative to the memory array cells. The present invention provides a method for programming reference cells to minimize margin loss and maximize cycling performance.




SUMMARY OF THE PRESENT INVENTION




The present invention provides a method for programming one or more reference cells, with the programming being performed relative to a prescribed cell on the same die as the reference cell (e.g., a memory cell or a golden bit cell).




According to one aspect of the invention, a method for programming a reference cell for use in an integrated circuit memory having an array of memory cells each exhibiting a native threshold voltage value is described. That method comprising the steps of first locating an address for the memory cell in the array that has the highest native threshold voltage value (VTNH). A reference cell is programmed a predetermined amount and its program state is sensed relative to the VTNH memory cell. The programming and sensing steps are repeated until the sensing step indicates that the reference cell has been programmed an amount sufficient to fail a first preselected read operation.




In a more particular methodology in accordance with this aspect of the invention, the locating step can include the steps of iteratively increasing a gate voltage applied to the memory cells and performing the first preselected read operation at each such applied gate voltage until a final gate voltage is identified at which all the memory cells in the array pass the first preselected read operation. Further, the first preselected read operation can exclude memory cells that have already passed the first preselected read operation at a previously applied gate voltage.




According to another aspect of the invention, a method for programming a reference cell for use in an integrated circuit memory having a plurality of memory cells each exhibiting a native threshold voltage value is described. That method locates an address for the VTNH cell by applying a first gate voltage value at which at least one memory cell fails a first preselected read operation and increasing the applied gate voltage until a final gate voltage value is reached at which each of the memory cells can just pass the first preselected read operation. The reference cell is programmed a predetermined amount and the program state of the reference cell relative to the VTNH memory cell is sensed by performing a second preselected read operation on the reference cell. The programming and sensing steps are repeated until the sensing step indicates that the reference cell has been programmed an amount sufficient to fail the second preselected read operation.




According to still another aspect of the invention, a method for programming a set of reference cells for use in performing respective read operations on an integrated circuit memory having a plurality of memory cells is described. That method locates the VTNH cell, determines a placement for a reference voltage read signal relative to the VTNH cell, places a reference voltage erase verify signal relative to the reference voltage read signal, and places a reference voltage program verify signal relative to the reference voltage read signal.




The foregoing methods can have their sensing steps performed relative to the VTNH memory cell and also relative to a native cell (a golden bit cell) on-board the same die.




The inventive method can be utilized to program a reference cell used with a memory array, a sliced array having one or more columns of memory cells, and redundant or auxiliary arrays.




These and other more specific aspects and features of the present invention can be appreciated from the acompanying Drawing Figures and Detailed Description of Certain Preferred Embodiments.











BRIEF DESCRIPTION OF THE DRAWINGS




The present invention will be understood and appreciated more fully from the following detailed description taken in conjunction with the drawings in which:





FIG. 1

illustrates a spread of native threshold voltage values for a number of memory cells on a die.





FIG. 2

illustrates target threshold voltage values for programming reference cells in order to achieve exemplary and desired operating margins.





FIG. 2A

illustrates a practical example of operating margins and a preferred sequence of steps to establish the operating margins relative to a particular cell within the memory array.





FIG. 3A

plots signal development through an array cell and a reference cell when both are driven by the same VCCR voltage supply.





FIG. 3B

is the plot of

FIG. 3A

, showing the array cell being driven by a trimmed external supply voltage;





FIG. 3C

is the plot of

FIG. 3B

, showing the reference cell at various program states and the results of the read operations.





FIG. 4

illustrates an overall process for programming a selected reference cell according to a first preferred method of the present invention.





FIG. 4A

illustrates a process for locating the memory cell within a memory array having the highest native threshold voltage value (denoted “VTNH”) among the memory cells in the array.





FIG. 5

illustrates a process for programming a selected reference cell to a target value in accordance with the first preferred method.





FIG. 6

illustrates a process for more accurately determining the threshold voltage of a reference cell placed in accordance with the embodiments of the present invention.





FIG. 7

illustrates a process for programming a reference cell relative to another reference cell.





FIG. 8A

plots signal development for a “golden bit” cell when driven by a trimmable external supply voltage, together with the VTNH cell and an exemplary reference cell.





FIG. 8B

is the plot of

FIG. 8A

, showing the reference cell at various program states relative to a golden bit's signal development that has been placed using the trimmable external supply voltage and the results of various read operations.











DETAILED DESCRIPTION OF CERTAIN PREFERRED EMBODIMENTS




By way of overview and introduction, the present invention is described in connection with a methodology for programming a reference cell to enable sensing of the contents of a memory cell from close to ground level. Such a memo array is described in the aforementioned U.S. Pat. No. 6,134,156 and U.S. application Ser. No. 09/827,756, filed on even date herewith and entitled “Architecture And Scheme For A Non-Strobed Read Sequence.” By using a reference cell instead of a fixed threshold for comparison, a low voltage signal can be reliably processed irrespective of any changes in temperature or power supply level. The present invention bas applicability to other sensing schemes, including A.C. and D.C. sensing techniques, and with read operations from either the source or drain side of a transistor, as can be appreciated by those of skill in the art.




Reference is made to

FIG. 1

which illustrates the memory array cells native threshold voltage distribution. The memory cells on the die include both array cells that are addressed in a conventional manner, and additional cells such as auxiliary and reference cells that are used for a variety of purposes including quality control during manufacturing and read operations when the array cells are put into service. The native reference cells can lie anywhere along the threshold voltage (VT) axis. By way of example, the native threshold values for two reference cells are indicated as lying within the distribution curve, though that is not required.




The methods of the present invention enable precision programming of reference cells relative to a memory cell on the die, for example, relative to an array cell, another reference cell, or to a golden bit. Each of these techniques defines a related but different method described herein with reference to a flow diagram of a preferred embodiment for each such technique. These techniques are performed after manufacturing the devices, prior to placing the memory array into service. Generally, there are a number of target threshold voltage values that are set in a corresponding number of reference cells as a result of the reference cell programming process of the invention. That is to say, different types of read operations can be performed on the memory cell such as program verify, erase verify and operations in which temporary states of a cell in the progress of an erase or a program operation should be detected. The basic difference presented by these in-progress read cycles is the placement of the reference cell signal. Since the reference cell's state is not changed according to the type of the read operation, different reference cells should be used for each of these types of operations.




As illustrated in

FIG. 2

, target threshold voltage values are spaced relative to one another by predetermined margins M


1


, M


2


, M


3


and M


4


which are established to ensure reliable reads, read verifies, erase verifies, program verifies, or other read operations on the array memory cells. To minimize margin loss and maximize product endurance and reliability, it is critical that the margins M


1


. . . M


4


provide a reliable buffer and a tight operating window. As well, it is essential that the placement of the programmed reference cells is defined (at least for one of the reference cells) relative to the array threshold voltages (VT) distribution.





FIG. 2

illustrates an example of the margins M


1


. . . M


4


established and positioned relative to the memory cell in the array having the highest native threshold value (“VTNH”). As will be appreciated from the discussion below, the margins can be established and positioned relative to any cell, such as the VTNH cell, a golden bit cell, or another reference cell.





FIG. 2A

illustrates a practical example of the margins M


1


. . . M


4


that can be well defined using the techniques of the present invention. Four margins are illustrated: cycling margin (“CM”), erase margin (“EM”), retention margin (“XM”) and refresh margin (“FM”). These margins are maintained between the reference cells, once their threshold voltage value have been programmed. In other words, the margins are controlled based on the response of each reference cell when driven by a standard voltage source (“VCCR”). Since the reference programming procedure results in reference threshold voltages that may not be exactly as the original target levels, it is necessary that the actual reference cells threshold voltages maintain the following margins:








VT









RD









REF


_actual−


VTNH>=CM+EM












VT









RD









REF


_actual−


VT









EV









REF


_actual>=


EM












VT









RV









REF


_actual−


VT









RD









REF


_actual>=


XM












VT









PV









REF


_actual−


VT









RV









REF


_actual>=


FM








where VT_RD_REF_actual is the threshold voltage value of a reference cell programmed to implement a cell-contents read operation, VT_EV_REF_actual is the threshold voltage value of a reference cell programmed to implement an erase verify operation concerning the contents of a cell, VT_RV_REF_actual is the threshold voltage value of a reference cell programmed to implement a refresh verify operation concerning the contents of a cell, and VT_PV_REF_actual is the threshold voltage value of a reference cell programmed to implement a program verify operation concerning the contents of a cell.





FIG. 2A

also illustrates a preferred sequence of steps to establish the operating margins relative to a particular cell within the memory array. The steps in

FIG. 2A

are shown relative to the VTNH cell, though other cells can be used such as the golden bit cell, as described below. Use of the VTNH cell facilitates discussion because it clearly illustrates the relationship between the required threshold voltages to be programmed into the reference cells and the required operating margins. Thus, with reference to the VTNH cell, a reference cell is programmed to be the RD_REF cell by programming it to have a margin M of CM+EM above VTNH. If a golden bit cell were used as the base line threshold voltage, then the margin M would most likely be greater by the difference between the threshold voltage of the golden bit cell (V_OTPG) and the VTNH cell.




Next, another reference cell is programmed to be the erase verify cell by programming that cell to have a margin EM below the actual threshold voltage of the RD_REF cell, as can be appreciated from FIG.


2


A. Likewise, the RV_REF cell can be programmed to be XM above the actual threshold voltage of the RD_REF cell, and another cell is programmed to be the PV_REF cell by programming it to have a margin FM above the actual threshold voltage of the RV_REF. It will be appreciated that other arrangements of the reference cells' threshold voltages and of the sequence of programming the reference cells are possible and are part of the present invention. Furthermore, the placement of each reference cell can be made relative to an array cell, a “golden” cell, or another reference cell.




When put into service, it is desirable to drive both the array cells' and the reference cells' gates with the same voltage supply VCCR during all standard read operations. However, during reference programming, a trimmable external voltage supply (EXT_VCCR) driving the array cells' gates can be used to facilitate the reference cell programming procedure. The effect of trimming the external supply is illustrated in

FIGS. 3A

,


3


B and


3


C. In

FIG. 3A

, the development over time of voltage signals from a given, native reference cell is plotted together with the VTNH cell when both are driven by VCCR. We can define a “pass read ‘1’” operation to be when the array cell has a greater signal than the reference cell and a “fail read ‘1’” operation to be when the array cell signal is lower than the reference signal. For example, the result of a read operation between the native VTNH cell and the native reference cell shown in

FIG. 3A

will be “fail read ‘1’”.




In

FIG. 3B

, the gate voltage applied to the VTNH cell has been reduced by a margin M, and the lower gate voltage results in slower signal development in that cell. Consequently, when compared to VTNH driven at the trimmed supply voltage level, the reference cell REF


1


(driven by the standard gate voltage VCCR) must be programmed further in order to fail the read ‘1’ operation.

FIG. 3C

illustrates the reference cell REF


1


at various program states along the process of being programmed including its native state in which it fails the read ‘1’ operation, an interim state in which the cell REF


1


has been partially programmed, yet still fails the read ‘1’ operation, and a state in which the further programmed reference cell signal is smaller than the array cell's signal and so it passes the read ‘1’ test. At this last state, it is ensured that the threshold voltage of the reference cell has been raised by at least the M margin.




A read operation or “sensing” of a cell can be perform as described in the aforementioned, co-pending U.S. patent application Ser. No. 09/827,756, filed on even date herewith and entitled “Architecture And Scheme For A Non-Strobed Read Sequence,” which describes the steps taken to sense a close to ground signal sensing methods (e.g., DC current sensing).





FIG. 4

illustrates the overall process for placing a selected reference cell with regard to a first preferred method of the present invention. At step


410


, the VTNH memory cell within the memory cell array is located. That cell is then used as a reference for verifying the programming of the selected reference cell, as indicated at step


420


. All margins are then established with respect to the threshold voltage for that cell, which is the highest threshold voltage of any cell in the memory array. This method places the VT for the selected reference cell by programming the cell in accordance with a prescribed criterion at step


430


. Generally, the criterion is that the VT for the reference cell is VTNH+M, where M is a prescribed margin.




As a specific example, if the selected reference cell is the RD_REF cell, then M is CM+EM and target reference cell VT is set to be greater than or equal to VTNH+CM+EM. The reference cell is preferably programmed a predetermined amount until placed, as described in detail with regard to the flow diagrams of

FIGS. 5-8

.




With reference now to

FIG. 4A

, step


410


of

FIG. 4

for locating the VTNH cell proceeds by first setting an initial, low voltage level (EXT VCCR) that is externally applied to the gates of respective memory cells, in succession, as the array is parsed, as shown at step


450


. Meanwhile, a predetermined gate voltage level such as the standard gate voltage level at read (“read VCCR”), is applied to the gate of the reference cell. It will be recalled that the reference cell has a similar structure (e.g., size, access path, environment, etc.) to the array cells and has a matched load; thus, as was illustrated in

FIG. 3B

, the comparatively low voltage applied to the gate of the still native array cells by EXT_VCCR at step


452


causes the array cell to develop a smaller signal than the reference cell. The memory array is then parsed starting with a first array cell address, as shown at step


454


.




At a prescribed moment after applying these gate voltages, for example, as described in the aforementioned patent application entitled “Architecture And Scheme For A Non-Strobed Read Sequence,” a read operation is performed on consecutive memory cells in the array against the reference cell (see step


456


) until a fail-to-read-1 condition is detected. Assuming that EXT_VCCR has been initially set to a low enough value (e.g., 2 Volts), then the first read operation is expected to fail a read ‘1’ operation, as tested at step


458


, since both cells are native and are expected to have similar characteristics. If the read does not fail “read ‘1’” (i.e., it passes), then if the entire array has not already been parsed, as tested at step


460


, then the next array cell is selected at step


462


and the process is repeated, starting again at step


456


, until a fail read ‘1’ is detected. The address of the array cell that failed the test is stored as well as the EXT_VCCR level, at step


464


. The voltage EXT_VCCR is increased at step


466


, preferably by a fixed small increment J (typically smaller than 100 mV), and, if the array has not been completely parsed, as tested at step


468


, this read operation is repeated starting from the next array cell address at step


470


until the fail read ‘1’ is detected (at step


458


). On the other hand, if the array has been completely parsed yet the present address has failed to read ‘1,’ then the die is considered as a bad die, and the operator is advised of this at step


469


. A bad die can also be detected if the level of EXT_VCCR has been incremented beyond a prescribed maximum, as shown at step


471


. The maximum level for EXT_VCCR can be established based on empirical data for processing similar devices.




After an array cell address is stored at step


464


and a valid next address is obtained at step


470


, further addresses are examined to determine whether EXT_VCCR must be incremented and to store a new array cell address. As additional array cells, if any, fail the read ‘1’ test at step


458


, the already stored address and EXT_VCCR level can be replaced by the present address and EXT_VCCR level for the presently read array cell, and the EXT_VCCR level can be increased again, with the process continuing from the next array cell address, assuming there are more addresses to parse (as tested at step


468


).




As described above, if the read operation repeats starting at the last stored cell address and continues to pass the read ‘1’ operation at step


458


then the last address of the array will eventually be reached, as tested at step


460


. Once the entire array has already parsed, the flow proceeds to step


480


at which point a test is made to ensure that the die is not defective by examining the current setting for EXT_VCCR. If EXT_VCCR is above an established minimum setting, then the last stored address is the VTNH memory cell address because it was the last cell that failed to be read as a ‘1’. Also, the last stored level for EXT_VCCR is the V_VTNH level. These values are preserved for use in reference cell programming, as indicated at step


482


. On the other hand, if EXT_VCCR is not above an established minimum level, the operator is advised that the die is defective, as indicated at step


484


.




This process enables each cell to be read only once, in other words, a single parse of the array locates the VTNH memory cell address, as indicated at step


462


. This is a preferred method although other flows for parsing the array can be used although they will probably result in a longer process due to the larger number of read operations (e.g., for each EXT_VCCR level start the read operations at a specific first address and parse the array till a fail read ‘1’ is detected).




The increment step of the EXT_VCCR supply in this process determines the accuracy of the VTNH cell identification. Accuracy of identification means that there is no cell in the array whose threshold voltage is larger than VTNH plus the EXT_VCCR increment step level, where VTNH is the threshold voltage of the cell identified as the VTNH cell. Assume for example that a 100 mV increment step is used. At the end of the process, the stored address indicates the cell that failed read ‘1’ at the stored EXT_VCCR level. The remaining cells that were read in the next loop of read cycles, applying to their gates the stored EXT_VCCR level plus the 100 mV increment step, pass the read ‘1’ test (otherwise the address of the first failing one would have been stored). However, if a smaller increment step was used, for example 50 mV, then one of those remaining cells could have failed the read ‘1’ test. Thus, if a more accurate identification of the VTNH cell is required, then either one of the following two options can be adopted:




1. Use a smaller EXT_VCCR increment step for the whole process.




2. Use a coarse EXT_VCCR increment step for a first pass of the whole array and a finer EXT_VCCR increment step for a second pass of the cells starting at the stored VTNH cell address.




A separate aspect of this procedure is that a final voltage value for EXT_VCCR which is arrived at once the array has been parsed provides an indication as to whether the threshold voltage of the reference cell or an array cell is outside of a standard distribution of values. This value is therefore useful in determining the quality of the memory array. If at the end of the process the final EXT_VCCR level is the starting level (EXT_VCCR_min) then this indicates that the reference cell threshold voltage is significantly larger than the array cells' threshold voltages. As well, by setting a maximum level for the EXT_VCCR supply (EXT_VCCR_max), if this level is reached then it indicates that the reference cell threshold voltage is significantly lower than at least one of the array cells' threshold voltages. Any indication of such non-standard distribution of threshold voltages can be used as a quality check and as a basis for a decision such as to reject the part (steps


469


and


484


in FIG.


4


A).




Moreover, while the foregoing steps are operative to locate the VTNH memory cell, they also can serve as a blank test for the memory array, in which all cells are read once, and so the locating process does not increase the sort time during manufacturing.




Optionally, the final value of EXT_VCCR is stored, and, if desired, the value of EXT_VCCR at each cell at which a read ‘1’ operation fails can be stored.




With reference now to

FIG. 5

, a process for programming a selected reference cell REF


1


to a target value in accordance with the first preferred method is described. At step


510


, the VTNH cell, once located, is driven by a trimmable voltage source EXT_VCCR at a voltage value of








EXT









VCCR=VCCR−M








where M is a prescribed margin suitable for the selected reference cell. For purposes of illustration, assume that the reference cell REF


1


is the read reference cell RD_REF shown in

FIG. 2A

for which








M=CM+EM.








Meanwhile, as indicated at step


520


, the reference cell REF


1


is driven with a standard supply voltage having a supply voltage value of VCCR. As should be appreciated, the trimmable voltage source EXT_VCCR permits the VTNH array cell signal to be temporarily placed at the target location of the reference cell REF


1


signal, by applying a reduced gate voltage during this reference cell programming phase test as compared to the VCCR voltage ordinarily applied to the VTNH (and other array cells) when the memory cell is put into service.




At step


530


, a program pulse is applied to the reference cell REF


1


. A test is then made at step


540


to determine whether the result of a read operation of the VTNH cell (driven by the trimmed gate voltage EXT_VCCR) against the reference cell REF


1


(driven by the standard gate voltage at read, VCCR) is PASS read ‘1’ or FAIL. This test is performed by sensing the program state of the reference cell relative to the VTNH cell (driven by the trimmed gate voltage EXT_VCCR) preferably using the standard sensing scheme that is used when the device is put into service. If it does not pass “read ‘


1


’” then the reference cell has not been programmed to the location at which the VTNH cell signal has been temporarily placed (by the trimmed gate voltage EXT_VCCR), and farther programming pulses are required. The process loops back to step


530


so that an additional program pulse can be applied, and then a read test as at step


540


is again performed. The program pulses can be applied in a fixed increment (i.e., a predetermined amount), or an algorithm can be employed to enhance the process. For one suitable algorithm, see U.S. application Ser. No. 09/827,596, filed on even date herewith and entitled “Programming and Erasing Methods for A Reference Cell of An NROM Array,” which application is a continuation-in-part application of U.S. application Ser. No. 09/730,586, filed Dec. 7, 2000, which is a continuation-in-part application of U.S. Ser. No. 09/563,923, filed May 4, 2000. The process flow ends when the test at step


540


indicates that the reference cell REF


1


has been placed at the target location, that is, when it has been programmed an amount sufficient to pass the read ‘1’ test.




The incremental programming of the reference cell REF


1


is illustrated in

FIG. 3C

, in which the VTNH cell is driven by the trimmed gate voltage EXT_VCCR (step


510


) while an initially native reference cell REF


1


is driven by the standard gate voltage at read (step


520


). The initially native reference cell REF


1


exhibits signal development that is much more rapid than that of the VTNH cell, as shown by the curve “VREF


1


(native)”, when the cells are driven with such applied gate voltages. That curve fails to read ‘1,’ and so the process of

FIG. 5

continues by applying one or more program pulses to the reference cell REF


1


at step


530


. These program pulses have the effect of partially programming the reference cell and reducing its rate of signal development. That is reflected in the curve “VREF


1


(partial PGM'ing)”, which also fails to read ‘1.’ Consequently, additional program pulses are applied at step


530


and the test at step


540


is repeated, and this loop continues until a PASS read ‘1’ is detected, namely, when the signal development at the reference cell REF


1


is akin to the curve “VREF


1


(done PGM'ing)”. Of course, different tests can be arranged to provide a boundary condition to gauge when to stop programming the reference cell (e.g., external read of the cells currents, or other tests).




The process flow of

FIG. 5

permits the reference cell REF


1


to be positioned in the vicinity of the target; however, since each program pulse increments the VT of the reference cell by a predetermined amount, the actual VT of the reference cell after being programmed by the process flow of

FIG. 5

is typically above the target level. Consequently, if further reference cells REF


2


. . . REFx are to be programmed relative to the VT of that reference cell REF


1


, e.g., relative to the RD_REF reference cell, then knowledge of the actual level of the REF


1


threshold voltage, e.g., VT_RD_REF_actual, maybe desired. The VT level of the reference cell REF


1


relative to the VTNH cell can be refined by iteratively decreasing EXT_VCCR starting from EXT_VCCR=VCCR−M by a controlled amount which is a smaller value than the VT increment as a result of a programming pulse.





FIG. 6

illustrates a process flow for refining the position of reference cell REF


1


. It should be understood that the just-programmed reference cell itself is not being partially erased, but rather the trimmable voltage source EXT VCCR is adjusted so that the actual VT of the reference cell REF


1


can be more accurately determined.




At step


610


, the VTNH cell is driven with EXT_VCCR, that is with the voltage value that was applied at step


510


described above. Namely,








EXT









VCCR=VCCR−M








Meanwhile, at step


620


, the reference cell REF


1


is driven with a standard gate voltage VCCR. At step


630


, a counter N is set to zero for use in flow of

FIG. 6

to monitor the number of times that the trimmable voltage source has been trimmed a prescribed amount. Other techniques can be used as will be appreciated for tracking the number of adjustments or the overall adjustment to the voltage source EXT_VCCR.




At step


640


, the voltage value of EXT_VCCR is decreased by a prescribed amount K, which amount is preferably 50 mV or less, and at step


650


the counter N is incremented. At step


660


the VTNH array cell is read against the reference cell REF


1


and a test is made to determine whether the result is PASS read ‘1’. It is expected that in the first loop that the VTNH cell will pass this test, with the process flow looping to step


640


to again decrease the voltage value of EXT_VCCR and then increase the counter N (at step


650


). Ultimately, however, EXT_VCCR will be reduced such that the VTNH cell no longer passes the read ‘1’ test, and, at that point, the actual VT of the reference cell REF


1


is determined at step


670


to be:








VT


_REF


1


_actual=


VTNH+VCCR−EXT









VCCR


  (1)






Expressed another way, the actual VT of the reference cell REF


1


is:








VT


_REF


1


_actual=


VTNH+M+N*K,








where N is the number of loops in which the voltage value emanating from the EXT_VCCR source was decreased by the constant amount K.




A further process flow can be used in like manner to increment EXT_VCCR from its new-found value to a higher value by selecting a new voltage interval E, where E<K (e.g., E is 25 mV or less). Yet a further process flow can again decrement EXT_VCCR, and so on, to more accurately place the threshold voltage of the reference cell REF


1


.




With reference now to the flow diagram of

FIG. 7

, the steps taken for programming further reference cells REF


2


. . . REFx relative to the reference cell REF


1


is described. Preferably, the actual threshold voltage of the reference cell REF


1


has been identified through the process flow of FIG.


6


. In this first preferred embodiment, the reference cells are located relative to the VTNH cell, which was located in connection with the process flow of

FIG. 4

, described above.




The threshold voltage for the reference cell REF


2


is to be placed at:








VT


_REF


2


=


VT


_REF


1


_actual+M


2


,  (2)






where M


2


is the margin between REF


1


and REF


2


. M


2


can be positive or negative. If the threshold voltage of the reference cell REF


2


is lower than that of the reference cell REF


1


, then M


2


is a positive voltage value; otherwise, M


2


is a negative voltage value. For example, if REF


1


is the RD_REF cell and REF


2


is the RV REF cell, then the margin M


2


is XM, as shown in FIG.


2


A. For the EV_REF cell in

FIG. 2A

a negative M


2


margin (EM) should be used. Using equation (1) to solve equation (2) for EXT_VCCR, we see that








EXT









VCCR=VCCR−VT


_REF


1


_actual+


VTNH−


M


2




EXT









VCCR=VCCR−


M


1





N*K−


M


2








Where, M


1


is the original target margin of the REF


1


reference cell, N*K is the amount of over programming introduced to the REF


1


cell past the original target, and M


2


is a positive or negative target margin between the REF


1


and REF


2


threshold voltages.




For example, if REF


2


to be programmed relative to REF


1


is the EV_REF cell shown in

FIG. 2A

then M


2


=−EM, i.e., a negative margin. Similarly, if REF


2


is the RV_REF cell shown in

FIG. 2A

then M


2


=XM, i.e., a positive margin.




Thus, to place REF


2


relative to REF


1


, the gate of the VTNH cell is driven by EXT_VCCR while the gate of REF


2


is driven by the standard voltage VCCR, as indicated at steps


710


and


720


, respectively. At step


730


, a program pulse of a predetermined amount is applied to the reference cell REF


2


to move it from its native state and place it at its target programmed state relative to REF


1


. The cell REF


2


is being placed relative to REF


1


by sensing its value relative to the VTNH cell, while the VTNH cell is driven by a gate voltage which places the VTNH signal taking into account the actual placement of REF


1


and the required margin between the respective VTs of REF


1


and REF


2


. At step


740


, the partially programmed reference cell REF


2


is tested to see if it still passes a read ‘1’ test, in which case it has not been sufficiently programmed to place it M


2


from the reference cell REF


1


. If the cell fails the read ‘1’ test, as is initially expected since both cells are essentially native, then a further program pulse is applied at step


730


and the test at step


740


is repeated until the reference cell REF


2


fails the read ‘1’ test. At that point, the target placement for REF


2


has been achieved.




As described above in connection with the target programming of the reference cell REF


1


, a more accurate location of the actual placement of the reference cell REF


2


can be achieved using a process flow as in

FIG. 6

, this time applied to the reference cell REF


2


and this time setting the actual VT at step


650


to be:








VT


_REF


2


=


VTNH


+M


1


+N


1


*


K


+M


2


+N


2


*


K.








This refinement may be necessary if the reference cell REF


2


is to be used for placing yet a further reference cell, or for other purposes, such as accurate monitoring of the reference cell retention loss after bake.




Further programming of reference cells will follow one of the flows described above, depending if they are programmed relative to the VTNH cell or relative to another, already programmed, reference cell.




It is know in the art that, after manufacturing, memory devices are tested to detect malfunctioning or marginal devices. As part of these tests the array cells may be programmed and then the device is introduced to a relatively high temperature cycle. After this high temperature cycle, the threshold voltage of the array cells may drop by some amount. This threshold voltage drop is known as the “retention loss.” Since the reference cells are also programmed, their threshold voltage may also drop by some amount (which may be different than that of the array cells due to differences in the programming levels of the array cells and the reference cells). Since the threshold voltages of all the programmed memory cells are affected, whether a reference cell or an array cell (including the VTNH cell), there is no relative way to determine the cell's state. A non-relative way to determine the cells threshold voltage state, such as by an external measurement of the cells currents, is very expensive in terms of test time. Thus, an internal relative measurement of the array cells and the reference cells threshold voltages is desired.




A native cell which has never been programmed and which is on-board the same die is utilized to provide an internal relative measurement, in accordance with another aspect of the present invention. This native cell, referred to herein as a “golden cell” or “golden bit,” permits the internal read mechanism to be used not only for the reference cells' programming flow as described above in connection with

FIGS. 4-7

, but also for detection of array cell and reference cell threshold voltage changes after a high temperature cycle. The process flows described above can each be modified such that the reference cells' signal placements are performed relative to this golden bit in conjunction with the native cell having the maximum threshold voltage (that is, the VTNH cell). Now, even after a high temperature cycle, the reference cells' threshold voltages can be re-estimated against the golden bit because the threshold voltage of this unprogrammed (native) cell does not change as a result of any high temperature cycle(s).




The golden bit is a memory cell having the same size, the same environment, similar loads and matched or similar access paths as an array memory cell. However, the golden bit is usually not among the memory cells in the memory array, but rather is typically included in an auxiliary array. Referring back to

FIG. 4A

, the process flow for locating the VTNH cell, the V_VTNH voltage level found at step


462


is the EXT_VCCR level at which the VTNH cell signal just passed a read ‘1’ (i.e., became higher than) the reference cell signal.

FIG. 8A

plots signal development for the golden bit cell when driven by the trimmable external supply EXT_VCCR, together with the VTNH cell and an exemplary reference cell. In particular, the native golden bit signal can be placed by trimming the voltage applied to its gate using a method similar to that of

FIG. 4A

for finding V_VTNH. The voltage value of a trimmable gate voltage source EXT_VCCR is adjusted until the golden bit just passes a read ‘1’ test, and that defines the value of V_GB. The difference D between the gate voltages applied to the VTNH cell and the golden bit cell,








D




VTNH-GB




=V









VTNH−V









GB








can be used to program the reference cells and to now determine the actual threshold voltages of the programmed reference cells relative to the golden bit instead of the VTNH cell. As shown in

FIG. 8B

, reference cells can be programmed relative to the golden bit using a target threshold voltage level of:








VT


_REF


1


=


VTNH


+M


1


.






The method consists of placing the golden bit signal at the target place of the programmed reference cell, applying a programming pulse to the reference cell, and reading the contents of the cell to determine its contents. These steps are repeated until a PASS read ‘1’ is detected, as described above in connection with

FIGS. 4-7

. Thus, the flow is as described above for programming the reference cell using the VTNH cell, except that in the previous discussion the VTNH cell signal was placed at the target place of the programmed reference cell whereas now the golden bit cell's signal is placed at the target place of the programmed reference signal. The golden bit cell's signal is placed by driving its gate with








EXT









VCCR=VCCR−


M


1





D




VTNH-GB


.






Preferably, the reference cells are contained in a reference unit which includes a selector that can select which of several reference cells to use for a given operation, wherein each reference cell is programmed as described above to a prescribed level. Further details concerning the use of a selector and multiple reference cells can be found in the aforementioned U.S. Patent Application Serial No. to be assigned, attorney docket No. 2671/01148, filed on even date herewith, entitled “Method For Selecting a Reference Cell.”




If the reference cells are NROM cells, they can be used as either single or two bit cells. When used as single bit cells one of the two bits in the cell is programmed as explained above while the other bit can be maintained native, programmed or erased. On the other hand, if the two bits of the NROM reference cell are used as references, then each of them is programmed as explained above.




Persons skilled in the art will appreciate that the present invention is not limited to what has been particularly shown and described hereinabove. Rather the scope of the present invention is defined only by the claims which follow.



Claims
  • 1. In an integrated circuit memory having an array of memory cells each exhibiting a native threshold voltage value, a method for programming a reference cell comprising the steps of:a) locating an address for the memory cell in the array which has the highest native threshold voltage value (VTNH); b) programming the reference cell a predetermined amount; c) sensing the program state of the reference cell relative to the VTNH memory cell; and d) repeating steps b) and c) until the sensing step indicates that the reference cell has been programmed an amount sufficient to pass a first preselected read operation.
  • 2. The method as in claim 1, wherein the locating step includes the steps of iteratively increasing a gate voltage applied to the memory cells and performing the first preselected read operation at each such applied gate voltage until a final gate voltage is identified at which all the memory cells in the array pass the first preselected read operation.
  • 3. The method as in claim 2, wherein the step of performing the first preselected read operation excludes memory cells that have already passed the first preselected read operation at a previously applied gate voltage.
  • 4. The method as in claim 2, wherein the locating step is accomplished in a single pass through the array.
  • 5. The method as in claim 4, including the additional step of performing a second pass through a portion of the array while applying to the memory cells in that portion a gate voltage which is less than the final gate voltage.
  • 6. The method as in claim 5, wherein the portion is delimited at one end by the address of the VTNH memory cell.
  • 7. The method as in claim 1, wherein the sensing step includes comparing signals respectively generated by the reference cell and the VTNH memory cell, the gate of the reference cell being driven with a standard gate voltage level at read (“read VCCR”) for the first preselected read operation and the VTNH memory cell being driven with a gate voltage level equal to VCCR less a predetermined margin, the compared signals indicative of whether the reference cell signal is higher or lower than the VTNH memory cell signal.
  • 8. The method as in claim 7, wherein the signal output by the differential comparator indicates that the first preselected read operation has failed if the reference signal is higher than the VTNH memory cell signal and wherein the signal output by the differential comparator indicates that the first preselected read operation has passed if the reference signal is lower than the VTNH memory cell signal.
  • 9. The method as in claim 1, wherein the sensing step is performed relative to the VTNH memo cell and to a native cell within the same array.
  • 10. The method as in claim 1, wherein the locating step includes the steps of:a. setting an initial, low voltage level to be applied to the gate of the memory cells; b. applying a predetermined gate voltage level to the gate of the reference cell; c. performing a read operation on consecutive memory cells against the reference cell until a fail-to-read-1 condition is detected; d. storing the address of the memory cell at which the read operation of step (c) stopped; e. increasing the voltage level applied to the gates of the memory cells; and f. repeating steps c, d and e until the entire array has been read, wherein the VTNH memory cell is the memory cell in the array last stored at step (d).
  • 11. The method as in claim 10, wherein the repeating step commences with the memory cell address stored at step (d).
  • 12. The method as in claim 10, wherein the step of performing a read operation makes one complete pass through the array.
  • 13. The method as in claim 10, wherein the step of increasing the voltage level applied to the gates of the memory cells results in a final value of the voltage level, the final value providing an indication as to whether the threshold voltage of the reference cell is outside of a standard distribution of values.
  • 14. The method as in claim 10, wherein the step of increasing the voltage level applied to the gates of the memory cells results in a final value of the voltage level, the final value providing an indication as to whether the threshold voltage of one or more of the array cells is outside of a standard distribution of values.
  • 15. The method as in claim 10, wherein step of locating the VTNH memory cell provides a blank test for the memory array.
  • 16. The method as in claim 10, including the additional step, for at least the VTNH memory cell, of storing the voltage level applied to a low voltage level (“EXT VCCR”) of the gate of the memory cell.
  • 17. In an integrated circuit memory having a plurality of memory cells each exhibiting a native threshold voltage value, a method for programming a reference cell comprising the steps of:a) locating an address for the memory cell having the highest native threshold voltage value (VTNH) by applying a first gate voltage value at which at least one memory cell fails a first preselected read operation and increasing the applied gate voltage until a final gate voltage value is reached at which each of the memory cells can just pass the first preselected read operation; b) programming the reference cell a predetermined amount; c) sensing the program state of the reference cell relative to the VTNH memory cell by performing a second preselected read operation on the reference cell; and d) repeating steps b) and c) until the sensing step indicates that the reference cell has been programmed an amount sufficient to fail the second preselected read operation.
  • 18. The method as in claim 17, wherein the sensing step comprises comparing signals of the reference cell and the VTNH memory cell, the reference cell being driven with a standard gate voltage value at read VCCR, the VTNH memory cell being driven with the final gate value plus a predetermined margin, the compared signals indicative of whether the reference cell has been programmed in an amount sufficient to fail the first preselected read operation.
  • 19. The method as in claim 17, wherein the locating step is part of a blank test to ensure that each of the plurality of memory cells can be programmed.
  • 20. The method as in claim 17, wherein the gate voltage applied during the locating step is increased at one of a fixed increment and a non-fixed increment.
  • 21. The method as in claim 20, wherein the increment is a fixed increment on the order of 0.05 volts.
  • 22. The method as in claim 17, wherein the sensing step is performed relative to the VTNH memory cell and to a native cell within the same array.
  • 23. A method for programming a set of reference cells for use in performing respective read operations on an integrated circuit memory having a plurality of memory cells, comprising the steps of:a) locating the cell in the memory array having the highest native threshold voltage value (VTNH); b) determining a voltage level for a reference voltage read signal relative to the VTNH cell; c) adjusting a reference voltage erase verify signal relative to the reference voltage read signal; and d) adjusting a reference voltage program verify signal relative to the reference voltage read signal.
  • 24. The method as in claim 23, wherein the determining step is performed relative to the VTNH memory cell and to a native cell within the same array.
CROSS-REFERENCE TO RELATED APPLICATIONS

This application is related to U.S. Pat. No. 6,128,226, issued Oct. 3, 2000 and entitled “Method and Apparatus for Operating with a Close to Ground signal,” to U.S. Pat. No. 6,134,156, issued Oct. 17, 2000 and entitled “Method for Initiating a Retrieval Procedure in Virtual Ground Arrays,” to U.S. application Ser. No. 09/827,756, filed on even date herewith and entitled “Architecture And Scheme For A Non-Strobed Read Sequence,” and to U.S. application Ser. No, 09/827,596, filed on even date herewith and entitled “Programming and Erasing Methods For A Reference cell of An NROM Array,” the foregoing patents and patent applications being incorporated by reference in their entireties as if set forth herein.

US Referenced Citations (132)
Number Name Date Kind
4016588 Ohya et al. Apr 1977 A
4017888 Christie et al. Apr 1977 A
4151021 McElroy Apr 1979 A
4173766 Hayes Nov 1979 A
4173791 Bell Nov 1979 A
4257832 Schwabe et al. Mar 1981 A
4281397 Neal et al. Jul 1981 A
4306353 Jacobs et al. Dec 1981 A
4342149 Jacobs et al. Aug 1982 A
4360900 Bate Nov 1982 A
4380057 Kotecha et al. Apr 1983 A
4388705 Sheppard Jun 1983 A
4448400 Harari May 1984 A
4471373 Shimizu et al. Sep 1984 A
4521796 Rajkanan et al. Jun 1985 A
4527257 Cricchi Jul 1985 A
4586163 Koike Apr 1986 A
4630085 Koyama Dec 1986 A
4667217 Janning May 1987 A
4742491 Liang et al. May 1988 A
4769340 Chang et al. Sep 1988 A
4780424 Holler et al. Oct 1988 A
4847808 Kobatake Jul 1989 A
4870470 Bass, Jr. et al. Sep 1989 A
4916671 Ichiguchi Apr 1990 A
4941028 Chen et al. Jul 1990 A
5021999 Kohda et al. Jun 1991 A
5042009 Kazerounian et al. Aug 1991 A
5075245 Woo et al. Dec 1991 A
5104819 Freiberger et al. Apr 1992 A
5142496 Van Buskirk Aug 1992 A
5159570 Mitchell et al. Oct 1992 A
5168334 Mitchell et al. Dec 1992 A
5172338 Mehrotra et al. Dec 1992 A
5175120 Lee Dec 1992 A
5214303 Aoki May 1993 A
5241497 Komarek Aug 1993 A
5289412 Frary et al. Feb 1994 A
5295108 Higa Mar 1994 A
5305262 Yoneda Apr 1994 A
5311049 Tsuruta May 1994 A
5315541 Harari et al. May 1994 A
5324675 Hayabuchi Jun 1994 A
5345425 Shikatani Sep 1994 A
5349221 Shimoji Sep 1994 A
5350710 Hong et al. Sep 1994 A
5359554 Odake et al. Oct 1994 A
5393701 Ko et al. Feb 1995 A
5394355 Uramoto et al. Feb 1995 A
5399891 Yiu et al. Mar 1995 A
5412601 Sawada et al. May 1995 A
5414693 Ma et al. May 1995 A
5418176 Yang et al. May 1995 A
5418743 Tomioka et al. May 1995 A
5422844 Wolstenholme et al. Jun 1995 A
5424567 Chen Jun 1995 A
5424978 Wada et al. Jun 1995 A
5426605 Van Berkel et al. Jun 1995 A
5434825 Harari Jul 1995 A
5436481 Egawa et al. Jul 1995 A
5440505 Fazio et al. Aug 1995 A
5450341 Sawada et al. Sep 1995 A
5450354 Sawada et al. Sep 1995 A
5455793 Amin et al. Oct 1995 A
5467308 Chang et al. Nov 1995 A
5477499 Van Buskirk et al. Dec 1995 A
5496753 Sakurai et al. Mar 1996 A
5518942 Shrivastava May 1996 A
5521870 Ishikawa May 1996 A
5523251 Hong Jun 1996 A
5553018 Wang et al. Sep 1996 A
5563823 Yiu et al. Oct 1996 A
5581252 Thomas Dec 1996 A
5583808 Brahmbhatt Dec 1996 A
5592417 Mirabel Jan 1997 A
5599727 Hakozaki et al. Feb 1997 A
5606523 Mirabel Feb 1997 A
5623438 Guritz et al. Apr 1997 A
5654568 Nakao Aug 1997 A
5656513 Wang et al. Aug 1997 A
5712814 Fratin et al. Jan 1998 A
5748534 Dunlap et al. May 1998 A
5754475 Bill et al. May 1998 A
5760445 Diaz Jun 1998 A
5768192 Eitan Jun 1998 A
5777919 Chi-Yung et al. Jul 1998 A
5784314 Sall et al. Jul 1998 A
5787036 Okazawa Jul 1998 A
5793079 Georgescu et al. Aug 1998 A
5801076 Ghneim et al. Sep 1998 A
5812449 Song Sep 1998 A
5812456 Hull et al. Sep 1998 A
5825686 Schmitt-Landsiedel et al. Oct 1998 A
5828601 Hollmer et al. Oct 1998 A
5834851 Ikeda et al. Nov 1998 A
5836772 Chang et al. Nov 1998 A
5841700 Chang Nov 1998 A
5847441 Cutter et al. Dec 1998 A
5864164 Wen Jan 1999 A
5870335 Khan et al. Feb 1999 A
5886927 Takeuchi Mar 1999 A
5946558 Hsu Aug 1999 A
5949728 Liu et al. Sep 1999 A
5963412 En Oct 1999 A
5963465 Eitan Oct 1999 A
5969989 Iwahashi Oct 1999 A
5973373 Krautschneider et al. Oct 1999 A
5991202 Derhacobian et al. Nov 1999 A
6011725 Eitan Jan 2000 A
6018186 Hsu Jan 2000 A
6020241 You et al. Feb 2000 A
6028324 Su et al. Feb 2000 A
6030871 Eitan Feb 2000 A
6034403 Wu Mar 2000 A
6034896 Ranaweera et al. Mar 2000 A
6063666 Chang et al. May 2000 A
6097639 Choi et al. Aug 2000 A
6108241 Chevallier Aug 2000 A
6128226 Eitan et al. Oct 2000 A
6134156 Eitan Oct 2000 A
6163048 Hirose et al. Dec 2000 A
6169691 Pasotti et al. Jan 2001 B1
6185143 Perner et al. Feb 2001 B1
6201282 Eitan Mar 2001 B1
6205056 Pan et al. Mar 2001 B1
6219277 Devin et al. Apr 2001 B1
6222768 Hollmer et al. Apr 2001 B1
6282145 Tran et al. Aug 2001 B1
6304485 Harari et al. Oct 2001 B1
6353554 Banks Mar 2002 B1
20020034097 Banks Mar 2002 A1
20020071313 Takano et al. Jun 2002 A1
Foreign Referenced Citations (11)
Number Date Country
0693781 Jan 1996 EP
0751560 Jan 1997 EP
1073120 Jan 2001 EP
1297899 Nov 1972 GB
2157489 Oct 1985 GB
04291962 Oct 1992 JP
05021758 Jan 1993 JP
07193151 Jul 1995 JP
09162314 Jun 1997 JP
WO 8100790 Mar 1981 WO
WO 9625741 Aug 1996 WO
Non-Patent Literature Citations (24)
Entry
U.S. patent application Ser. No. 08/902,890, Eitan, filed May 4, 2000.
U.S. patent application Ser. No. 08/905,286, Eitan, filed Jul. 30, 1997.
U.S. patent application Ser. No. 09/082,280, Eitan, filed May 20, 1998.
U.S. patent application Ser. No. 09/348,720, Eitan, filed Jul. 6, 1999.
U.S. patent application Ser. No. 09/413,408, Eitan, filed Oct. 6, 1999.
U.S. patent application Ser. No. 09/536,125, Eitan et al., filed Mar. 28, 2000.
Chan et al., “A True Single-Transistor Oxide-Nitride-Oxide EEPROM Device,” IEEE Electron Device Letters, vol. EDL-8, No. 3, Mar. 1987.
Chang, J., “Nonvolatile Semiconductor Memory Devices,” Proceedings of the IEEE, vol. 64 No. 7, pp. 1039-1059, Jul., 1976.
Eitan et al., “Hot-Electron Injection into the Oxide in n-Channel MOS Devices,” IEEE Transactions on Electron Devices, vol. ED-28, No. 3, pp. 328-340, Mar. 1981.
Glasser et al., “The Design and Analysis of VLSI Circuits,” Addison Wesley Publishing Co., Chapter 2, 1988.
Lee, H., “A New Approach For the Floating-Gate MOS Nonvolatile Memory”, Applied Physics Letters, vol. 31, No. 7, pp. 475-476, Oct. 1977.
Ma et al., “A Dual-bit Split-Gate EEPROM (DSG) Cell in Contactless Array for Single-Vcc High Density Flash Memories,” IEEE, pp. 3.5.1-3.5.4, 1994.
Oshima et al., “Process and Device Technologies for 16Mbit Eproms with Large-Tilt-Angle Implanted P-Pocket Cell,” IEEE, CH2865-4/90/0000-0095, pp. 5.2.1-5.2.4, 1990.
Ricco, Bruno et al., “Nonvolatile Multilevel Memories for Digital Application,” IEEE, vol. 86, No. 12, pp. 2399-2421, issued 1998.
Roy, Anirban, “Characterization and Modeling of Charge Trapping and Retention in Novel Multi-Dielectric Nonvolatile Semiconductor Memory Devices,” Doctoral Dissertation, Sherman Fairchild Center, Department of Computer Science and Electrical Engineering, pp. 1-35, 1989.
“2 Bit/Cell EEPROM Cell Using Band-To-Band Tunneling For Data Read-Out,” IBM Technical Disclosure Bulletin, U.S. IBM Corp. NY vol. 35, No. 4B, ISSN:0018-8689, pp. 136-140, Sep., 1992.
Tseng, Hsing-Huang et al., “Thin CVD Stacked Gate Dielectric for ULSI Technology”, IEEE, 0-7803-1450-6, pp. 321-324, 1993.
Pickar, K.A., “Ion Implantation in Silicon,” Applied Solid State Science, vol. 5, R. Wolfe Edition, Academic Press, New York, pp. 151-249, 1975.
Bhattacharyya et al., “FET Gate Structure for Nonvolatile N-Channel Read-Mostly Memory Device,” IBM Technical Disclosure Bulletin, U.S. IBM Corp. vol. 18, No. 6, p. 1768, Nov., 1975.
Martin et al. “Analog Integrated Circuit Design”, John Wiley and Sons Inc., Chapter 10, Jun. 1997.
Martin, K. “Improved Circuits for Realization of Switched-capacitor Filters,” John Wiley and Sons Inc., Chapter 10, Jun. 1997.
Bude et al., “EEPROM/Flash Sub 3.0 V Drain-Source Bias Hot carrier Writing”, IEDM 95, pp. 989-992.
Bude et al., “Secondary Electron Flash—a High Performance, Low Power Flash Technology for 0.35 μm and Below”, IEDM 97, pp. 279-282.
Bude et al., “Modeling Nonequilibrium Hot Carrier Device Effects”, Conference of Insulator Specialists of Europe, Sweden, Jun. 1997.