This application claims priority to German Application Nos. 10 2004 052 586.2 filed Oct. 29, 2004 and 10 2005 045 312.0 filed Sep. 22, 2005 which are incorporated herein, in their entirety, by reference.
The invention relates to a semiconductor memory with a combination of volatile and non-volatile memory cells. The invention further relates to the operation, the design, and to differing layout concepts for a semiconductor memory with a combination of volatile memory cells and non-volatile polymer memory cells.
A semiconductor memory device usually comprises a cell field consisting of a plurality of memory cells, and a matrix of column and line supply lines or word and bit lines, respectively. The memory cells are each positioned at the crosspoints of the electroconductive supply lines that are each electrically connected with the memory cell via an upper electrode or top electrode and a lower electrode or bottom electrode. To perform a change of the information content in a particular memory cell at the addressed crosspoint, or to recall the content of the memory cell, the corresponding word and bit lines are selected and impacted either with a write current or with a read current.
Different kinds of semiconductor memories are known, e.g. a RAM (Random Access Memory). A RAM memory device is a memory with optional access, i.e. data can be stored under a particular address and can be read out again under this address later. By selectively applying a voltage at a corresponding selection transistor via the column and row supply lines, it is possible to store an information unit (bit) in a capacitor during a write process and to recall it again during a read process via the selection transistor.
A particular kind of semiconductor memories are DRAMs (Dynamic Random Access Memory) which comprise in general only one single, correspondingly controlled capacitive element, e.g. a trench capacitor, with the capacitance of which one bit each can be stored as charge. DRAM memory cells stand out by especially short access times. The charge or the information stored, however, remains for a relatively short time only in a DRAM memory cell, so that a “refresh” must be performed regularly, wherein the corresponding information content is written in the memory cell again or is refreshed, respectively. The problem of the leaking currents in the memory capacitor existing with the DRAM memory concept, which may result in a loss of charge or a loss of information, respectively, has so far been solved insufficiently only by the permanent refreshing of the stored charge.
In contrast to DRAMs, no “refresh” has to be performed in the case of SRAMs (Static Random Access Memories) since the data stored in a SRAM memory cell remain stored as long as an appropriate supply voltage is fed to the SRAM. To this end, every memory cell of the SRAMs comprises as a rule a larger number of, for instance, 6 transistors, which entails the need for a larger space on a silicon substrate. It is, however, generally intended to accommodate as many memory cells as possible in a memory device, so that they have to be realized, or to be scaled, respectively, as simple as possible and on the smallest possible space.
Various non-volatile memory devices are known which are based on different physical principles. It is only with non-volatile memory devices (NVMs) such as EPROMs, EEPROMs, and flash memories, that the stored data remain stored even if the supply voltage is switched off. The flash memory concept, however, has the problem of restricted write and read cycles. Moreover, relatively high voltages are needed with flash devices since the charges have to overcome a barrier layer.
In addition to the above-described memory devices, memory types on the basis of polymers or specific molecules have also been suggested. The concept of the polymer memory cells deals with complex molecules that are adapted to assume two different states that are connected by an intramolecular charge flow. Such polymer memory cells may be addressed, written and read out electrically. In the case of modern polymer memory cells, an electrochemically active material of at least two different molecule or polymer layers that are each adapted to be reversibly transferred from an oxidized form to a reduced form and thus form an electrochemical Red/Ox pair is positioned in a volume between an upper electrode or top electrode and a lower electrode or bottom electrode. These molecule or polymer layers are in electric connection both with one another and with the respectively adjacent electrode layers of the top or bottom electrode, respectively.
As has been explained above, the DRAM semiconductor memory devices have the advantage of short write and read times, but the disadvantage of a volatile data content, which necessitates a permanent refreshing of the information stored. Contrary to this, the non-volatile polymer memory cells have the advantage that the information stored therein remains stored for a relatively long time even without any voltage supply. In prior art, e.g. in US 2004/0016947 A1, combinations of DRAM semiconductor memories and FLASH memory devices have already been suggested, which, however, have the above-mentioned disadvantages of the FLASH memory devices.
In the following, the basic functioning of a polymer memory cell is described. A typical structure of a polymer memory cell comprises, for instance, a first layer of an electroconductive material, a second layer being arranged on the first layer and being in electric connection therewith, the second layer comprising a first chemical compound that is adapted to be reversibly transferred from an oxidized form to a reduced form, a third layer arranged on the second layer, the third layer comprising a second chemical compound that is adapted to be reversibly transferred from a reduced form to an oxidized form, and a fourth layer of an electroconductive material being arranged on the third layer and being in electric connection therewith.
The memory cell thus comprises, with the second and third layers, at least two different molecule or polymer layers forming an electrochemical Red/Ox pair. If an appropriate voltage is applied to the electroconductive first layer and the electroconductive fourth layer, the first chemical compound comprised in the second layer yields electrons to the electroconductive first layer, thus oxidizing the first chemical compound. Simultaneously, electrons flow from the electroconductive fourth layer into the third layer, so that the second chemical compound comprised therein is transferred to the reduced form by the acceptance of electrons. If the polarity of the voltage is reversed, the memory cell may be written back to the original state. For equalizing the charges generated by the oxidation or reduction, respectively, of the first and second chemical compounds, protons flow from the second layer into the third layer, so that the memory cell altogether remains electrically neutral.
The information content of the polymer memory cell is determined by the state of oxidation of the first and second chemical compounds that are comprised in the second or third layer, respectively, of the memory cell. In the first state, the first chemical compound is in its reduced form and the second chemical compound is in its oxidized form. By applying a voltage, an oxidation or reduction, respectively, of the chemical compounds is effected. In the second state, the first chemical compound is then in its oxidized form whereas the second chemical compound is in its reduced form. By reversing the polarity of the voltage, swapping between the two states may be effected.
If the first chemical compound in its oxidized form and the second chemical compound in its reduced form are electrically neutral, the two logic states in the memory cell may be characterized in that in the first state the molecules are available in their neutral form and in the second state the molecules are available in ionic form. Thus, a purely electric reading out of the cell state is possible.
Another type of polymer memory devices is described in IEDM, 2003, Paper #10.2, “Organic Materials for High-Density Non-Volatile Memory Applications.”, by R. Sezi et al, Infineon Technologies. Such polymer memory cells are good to integrate in electronic circuits in that they are arranged on a substrate in which integrated circuits are structured. To this end, the polymer memory cells may be arranged in a memory field or array, respectively, with the word and bit lines being arranged at right angles to each other, so that they form crosspoints at each of which a polymer memory cell is formed.
The present invention combines the advantageous characteristics of volatile memory devices, on the one hand, and of non-volatile memory devices, on the other hand, without the disadvantages of restricted write and read cycles and of high voltages for the write and read processes.
According to one embodiment of the invention, there is a semiconductor memory with a volatile memory device, in particular a DRAM memory device, and with a non-volatile memory device, wherein the volatile memory device is electrically coupled with the non-volatile memory device, and wherein the non-volatile memory device comprises a polymer memory device adapted to be switched between two states of information.
According to another embodiment of the present invention, volatile DRAM memory devices are used in combination with resistively switching memory devices in the form of non-volatile polymer memory devices instead of FLASH memory cells. Thus, the present invention provides the combination between volatile memory devices with short access times in the form of DRAM memory devices and non-volatile memory devices in the form of resistively switching polymer memory cells that are adapted to immediately reload the information stored in the DRAM at the time it was switched off.
This way, on the one hand, the cell size of the memory is reduced since, depending on the embodiment, a reduced number of control gate lines or no control gate line at all is needed anymore. Due to the reduced number of control gate lines, a simpler operation is also achieved than in prior art. Furthermore, distinctly lower voltage values than with conventional semiconductor memories that are operated in combination with flash memory cells are required in the case of the inventive semiconductor memory. Another advantage of the inventive semiconductor memory consists in that the writing rate of a polymer memory is higher than that of a flash memory cell.
The present invention is primarily based on the combination of a volatile DRAM semiconductor memory with non-volatile polymer memory cells which enable to load information stored in the polymer memory cells into the DRAM semiconductor memory. Likewise, the information from the volatile DRAM semiconductor memory be transferred to the non-volatile polymer memory. By this, it may be achieved that the information stored in the DRAM semiconductor memory, or the state of the DRAM semiconductor memory available, respectively, prior to the switching off is stored in the non-volatile polymer memory. On switching on of the DRAM semiconductor memory, the information stored in the non-volatile polymer memory or the state available, respectively, prior to the switching off of the DRAM semiconductor memory may be again taken over in the DRAM semiconductor memory directly after the switching on. Thus, an instantaneous switching on of a system may be achieved without substantial time delay.
In the following, the invention will be explained in more detail with respect to several preferred embodiments and the enclosed drawings. In the drawings:
FIGS. 6 to 9 show the layout for semiconductor memories according to preferred embodiments of the present invention.
The inventive semiconductor memory comprises a matrix consisting of word lines and bit lines that are preferably in orthogonal orientation to each other. According to a first preferred embodiment of a semiconductor memory with a combination of volatile and non-volatile memory devices according to the present invention, the resistively switching polymer memory device is coupled with an additional transistor each.
For better understanding of the circuits illustrated in
In the circuit illustrated in
The embodiments of electric circuits illustrated in
word line WL
passing word line PWL
control gate line CG
plate line Plate
passing plate line PPlate
passing control gate line PCG
These electric lines are crossed orthogonally by bit lines BLj and BLj+1. The volatile DRAM memory device comprises a word line transistor or selection transistor 3, respectively, and a capacitor 9. As to the bit line plane BLj, the selection transistor 3 and the capacitor 9 of the DRAM memory device are each controlled via the word lines WLi or WLi+1, respectively, and the bit line BLj. To this end, the word lines WLi or WLi+1, respectively, are each connected with the gates of the selection transistors 3 while the bit line BLj is connected with the source/drain path of the selection transistors 3.
The non-volatile memory device comprises a polymer memory device P and an additional transistor 6. On the one side, the polymer memory device P is connected with the additional transistor 6, and on the other side with the plate line Plate. The control gate of the additional transistor 6 is contacted and controlled via the control gate line CG. Via a node, the additional transistor 6 is connected with the capacitor 9 of the DRAM memory device.
In the bit line plane BLj+1, the selection transistor 3 and the capacitor 9 of the DRAM memory device are controlled via the passing word lines PWLi or PWLi+1, respectively, and the bit line BLj+1. To this end, the passing word lines PWLi or PWLi+1, respectively, are each connected with the gates of the selection transistors 3 while the bit line BLj+1 is connected with the source/drain path of the selection transistors 3. The polymer memory device P is, on the one side, again connected with the additional transistor 6, but on the other side with the passing plate line PPlate. Accordingly, the control gate of the additional transistor 6 is contacted and controlled via the passing control gate line PCG. In a further bit line plane BLj+2 (not illustrated), the above-described structure of the bit line plane BLj would recur. This structure recurs as many times as desired, which is indicated by the orders i−1, i and i+1 or j and j+1, respectively.
In the following, the operating mode of the circuits illustrated in
Subsequently, a sense amplifier (not illustrated) amplifies this voltage value and draws the corresponding bit line 2 and a reference bit line to the respective logic level. To this end, the selected bit line 2 is, if a voltage of 0.9V is measured there, taken to a low level, and the pertinent reference bit line to a high level. If a voltage of 1.1V is measured at the selected bit line, the sense amplifier takes it to a high level and the pertinent reference bit line to a low level. This voltage value of the bit line that has been amplified by the sense amplifier effects that the voltage at the capacitor again assumes the value of prior to the reading process, which corresponds to a refreshing of the information stored in the corresponding DRAM memory cells 3, 9. Subsequently, the word line 1 may again be closed or deactivated by eliminating the selection, this causing the bit line 2 to be disconnected from the capacitor.
Writing of the information contained in the DRAM memory device 3, 9 into the non-volatile polymer memory device 5 may be effected as follows (with the word line 1 being in the low state and the control gate 7 of the additional transistor 6 being in the high state): By applying a negative voltage at the plate connection 4 of the polymer memory device 5 of, for instance, Vplate=−1.7V, it is effected that a voltage of Vc−Vplate=3 is generated between the capacitor voltage Vc at the node 8 to the plate connection 4. The capacitor voltage Vc of the DRAM memory device 3 is approx. 1.3V to 1.8V (high state of the DRAM memory device), so that the voltage lies above the writing voltage of the polymer memory device 5. By the charging of the capacitor 9, the polymer memory resistance is written, i.e. it becomes lowly resistive.
If there is no charge on the capacitor 9 of the DRAM memory device (low state of the DRAM memory device), the voltage Vc−Vplate=1.7V is not sufficient to write the polymer memory device 5, and the electric resistance of the polymer memory device 5 remains highly resistive and is therefore not written. This transfer of the information contained in the DRAM memory device 3, 9 to the polymer memory cells 5 may be effected in parallel on a chip altogether on which there is a number of circuits illustrated in
In the following, the operating mode of the circuits illustrated in
By applying a positive voltage that is lower than the deleting voltage VDel of the polymer memory device 5 at the plate connection 4, the corresponding value is transferred from the polymer memory device 5 to the DRAM memory device 3, 9. The polymer memory cell 5 is lowly resistive with an electric resistance of approx. 105 Ω. The capacitor 9 is charged with the time constant ζ=Rsp×Csp=3.5 ns, i.e. after a period of approx. 3.5 ns the capacitor voltage Vc at the node 8 has reached approx. 63% of the voltage Vplate at the plate connection 4. Thus, a voltage at the plate connection 4 of Vplate=2.3V results, for instance, in a capacitor voltage at the node 8 of Vc=1.5V (high state of the DRAM memory device).
If, however, the polymer memory device 5 is in a highly resistive state, there results a time constant of ζ=Rsp×Csp=3500 ns. Within a charging time of 3.5 ns, 0.1% of the voltage Vplate is achieved. Thus, a voltage at the plate connection 4 of Vplate=2.3V results, for instance, in a capacitor voltage at the node 8 of Vc=0.0023V (low state of the DRAM memory device). This way, the original value of the DRAM memory device or the information originally stored in the DRAM memory device 3, 9, respectively, is restored in the DRAM memory.
In addition to the above described operating modes for the transfer of information between the non-volatile polymer memory 5 and the volatile DRAM memory device 3, 9, a regular operation is, of course, also possible with the semiconductor memory according to the invention, wherein the DRAM memory devices 3, 9 can be used as a conventional DRAM system memory. With this regular DRAM operation, the control gate 7 of the additional transistor 6 is in a low state or has a negative voltage. To minimize leaking currents, either the pn-junction of the additional transistor 6 or the polymer memory device 5 may be connected with the node 8, as with the circuits illustrated in
In the following, the operating mode for deleting the polymer memory devices 5 of a semiconductor memory according to the present invention will be described. The polymer memory cells 5 may be deleted both in the idle cycle of the DRAM memory if the DRAM memory is not accessed, and during the writing back of the information stored in the polymer memory (in parallel to this in another memory block), or during the reading out of the corresponding values from the polymer memory devices 5, respectively. In so doing, all the polymer memory devices 5 connected with the corresponding word line 1 are deleted. According to a further preferred embodiment of the present invention, these processes may also take place in parallel in a semiconductor memory.
For deleting the polymer memory devices 5, the word line 1 is first of all opened, which effects a charge sharing or a charge equalization, respectively, between the precharged bit line 2 and the capacitor 9. Then, the voltage at the bit line 2 or at the capacitor 9, respectively, increases or decreases and assumes, depending on the respectively stored value, finally a value of approx. 0.9 or approx. 1.1V in the manner that has already been described above. Prior to the evaluation by means of a sense amplifier, the word line 1 is closed and the control gate is opened by the word line transistor of the DRAM memory device 3. If a voltage of −4.5V is applied to the plate connection 4, the voltage Vplate is higher than the deleting voltage VDel of the polymer memory devices 5, this causing all polymer memory cells 5 of the corresponding word line 1 to be deleted. Subsequently, after the decreasing of the voltage at the plate connection 4 and with a closed control gate, the value that has now been amplified by the sense amplifier is again written in the capacitor 9 in that the word line transistor 3 of the DRAM memory device is opened again.
The voltages or logic levels, respectively, required for the above described operating modes for an inventive semiconductor memory according to the first preferred embodiment with an additional transistor are listed in the following Table.
In accordance with a second preferred embodiment of the inventive semiconductor memory, the polymer memory devices 5 are each coupled with the zener diode 10.
For better understanding of the circuits illustrated in
As is illustrated in
word line WLi
passing plate line PPlatei
passing word line PWLi
plate line Platei
plate line Platei+1
passing word line PWLi+1
passing plate line PPlatei+1
word line WLi+1
These electric lines are crossed orthogonally by bit lines BLj and BLj+1. The DRAM memory device comprises a word line transistor or selection transistor 3, respectively, and a capacitor 9. With respect to the bit line plane BLj, the selection transistor 3 and the capacitor 9 of the DRAM memory device are each controlled via the word lines WLi or WLi+1, respectively, and the bit line BLj. To this end, the word lines WLi or WLi+1, respectively, are each connected with the gates of the selection transistors 3 while the bit line BLj is connected with the source/drain path of the selection transistors 3.
The non-volatile memory device comprises a polymer memory device 5 and a zener diode 10. On the one side, the polymer memory device 5 is connected with the zener diode 10, and on the other side with the plate line Platei or Platei+1, respectively. The zener diode 10 is connected with the capacitor 9 of the DRAM memory device via an appropriate node.
With respect to the bit line plane BLj+1, the selection transistor 3 and the capacitor 9 of the DRAM memory device is controlled via the passing word lines PWLi or PWLi+1, respectively, and the bit line BLj+1. To this end, the passing word lines PWLi or PWLi+1, respectively, are each connected with the gates of the selection transistors 3 while the bit line BLj+1 is connected with the source/drain path of the selection transistors 3. The polymer memory device 5 is on the one side again connected with the zener diode 10, but on the other side with the passing plate line PPlatei or PPlatei+1, respectively. Accordingly, the control gate of the additional transistor 6 is contacted and controlled via the passing control gate line PCG. The next bit line plane BLj+2 has the same structure as the bit line plane BLj. This structure recurs as many times as desired, which is indicated by the orders i and i+1 or j, j+1 and j+2, respectively.
In the following, the operating mode of the circuits illustrated in
By applying a negative voltage at the plate connection 4 of, for instance, Vplate=−3.5V, it is effected that a voltage of approx. Vc−Vplate=5V is generated between the capacitor voltage at the node 8 to the plate connection 4. The capacitor voltage Vc of the DRAM memory device 9 is approx. 1.3V to 1.8V (high state of the DRAM memory device). This voltage lies above the writing voltage and is further sufficient to operate the zener diode 10 (at a setpoint of 2V) in the breakthrough and to thus apply a voltage of 3V at the polymer memory device 5. By the charge of the capacitor 9, information is written into the polymer memory device 5 in that the polymer memory resistance is changed, i.e. the polymer memory device 5 becomes lowly resistive.
If the capacitor 9 does not include any charge (low state of the DRAM memory device), the voltage difference Vc−Vplate=3.5V is not sufficient to apply the breakthrough voltage and the writing voltage. Then, the polymer memory 5 remains highly resistive and is therefore not written. This transfer of the information contained in the DRAM memory device 3 to the polymer memory devices 5 may again be effected in parallel on the entire chip on which there are a number of circuits illustrated in
The operating mode or the method, respectively, for loading the respectively stored value from the non-volatile polymer memory 5 to the DRAM memory device 3 of the embodiment illustrated in
In addition to the above-described operating modes for the transfer of information between the non-volatile polymer memory 5 and the volatile DRAM memory device 3, a regular DRAM operation is also possible with the inventive semiconductor memory even in the embodiment with a zener diode 10, in which the DRAM memory devices 3 are used as a conventional DRAM system memory. With this regular DRAM operation, the zener diode 10 blocks and a regular DRAM operation is possible. By connecting either the pn-junction of the zener diode 10 or the polymer memory 5 with the node 8, leaking currents can additionally be minimized. The plate connection 4 is kept at a voltage of Vplate=0V.
In the following, the operating mode for deleting the polymer memory devices 5 is described with the circuits illustrated in
For deleting the polymer memory devices 5, the word line 1 is first of all opened, which effects a charge sharing or a charge equalization, respectively, between the precharged bit line 2 and the capacitor 9. Then, the voltage at the bit line 2 or at the capacitor 9, respectively, increases or decreases and assumes, depending on the respectively stored value, finally a value of approx. 0.9 or approx. 1.1V in the manner that has been described above. Prior to the evaluation by means of a sense amplifier, the word line 1 is closed. If a voltage of −5V is applied to the plate connection 4, the voltage at the polymer memory device 5 is composed of Vplate−Vdiode−0.9V or of Vplate−Vdiode−1.1V, respectively. In both cases, the voltage is thus higher than the deleting voltage VDel of the polymer memory device 5, this causing all polymer memory cells 5 of the corresponding word line 1 to be deleted.
Subsequently, after the decreasing of the voltage at the plate connection 4 and with a closed control gate 7, the value that has now been amplified by the sense amplifier is again written in the capacitor 9 of the DRAM memory device in that the word line transistor 3 is opened again. Alternatively, after the reading out of the voltage of the capacitor 9 to the bit line 2 (charge equalization), the capacitor 9 can be (completely) discharged, for instance, via the plate connection 4, this making it possible to decrease the voltage for the writing process at the plate connection to approx. Vplate=3.7V.
In the following Table, the voltages or logic levels, respectively, required for the above-described operating modes are listed for an inventive semiconductor memory according to the second preferred embodiment with a zener diode.
In the following, the structure and the layout for an inventive semiconductor memory is described by means of FIGS. 5 to 10.
The structure illustrated in
word line WL
passing word line PWL
passing word line PWL
word line WL
In a second plane above the first plane, the following electric lines are arranged:
plate line Plate
passing plate line PPlate
passing plate line PPlate
plate line Plate
As results from
Below the adjacent passing word lines PWL, two respective trench capacitors TK are formed, at the upper end of which an L-shaped buried strap BS each is arranged as conducting transistor contact means. The trench capacitors TK are separated from each other by a shallow trench isolation layer STI. An N+ doped diffusion region each is provided at the sides of the L-shaped buried straps BS that are opposite to the shallow trench isolation layer STI. Between the buried strap BS and the passing word line PWL, a respective thick oxide layer Ox is formed to electrically isolate both the two L-shaped buried straps BS and the adjacent passing word lines PWL from each other.
Below the word lines WL, there is formed a respective gate oxide GO, whereby the N+ diffusion region below the tungsten vias W of the bit line BL and the adjacent N+ diffusion region that is adjacent to the L-shaped buried strap BS above the trench capacitor TK are electrically coupled with each other. This way, the N+ diffusion region below the tungsten vias W of the bit line BL forms, together with the N+ diffusion region adjacent to the L-shaped buried strap BS above the trench capacitor TK, the above-mentioned selection transistor of the DRAM memory device, which is controlled via the word line WL.
Between the passing word lines PWL and the trench capacitors TK, the forming of a transistor is prevented by a respective thick oxide layer Ox, and the capacitor for the DRAM memory device is generated instead. Below the thick oxide layer Ox, the L-shaped buried straps BS form a conducting transistor connection between the trench capacitor TK and the N+ doped source/drain region of the selection transistor.
Between the plane with the plate lines Plate and the plane with the word lines WL, a polymer memory device P is positioned. A P+ doped polysilicon contact that connects the N+ doped source/drain region via a suitable contact layer K with the polymer memory device P is adjacent to the N+ doped source/drain region of the selection transistor below the word line WL. This P+ doped polycontact below the polymer memory device P forms, together with the N+ doped source/drain region between the word line WL and the passing word line PWL, a zener diode.
Above the polymer memory device P, there is positioned the plate line Plate that is connected with the polymer memory device P also via a suitable contact layer K. The polymer memory device P may comprise a layer succession as has been described initially in conjunction with polymer memory devices. Accordingly, the material for the contact layer K above and below the polymer memory device P is, as a rule, different and depends on the type of polymer memory device P used. Adjacent to the plate line Plate, the passing plate line PPlate is arranged, which also serves the contacting of the polymer memory device P, as is described in relation with
FIGS. 6 to 9 each show a schematic representation of the layout for a semiconductor memory according to a preferred embodiment of the present invention.
At particular crosspoints with the bit line BL, a respective shallow trench isolation layer (not illustrated) may be provided between the word line WL and the plate line thereabove, as well as between the passing word line PWL and the passing plate line thereabove, so as to form either a transistor T or a capacitor C. A word line WL or a passing word line PWL forms, at a respective crosspoint with a bit line BL, only a transistor if no shallow trench isolation layer is positioned therebetween. If, at the crosspoint with a bit line BL, a shallow trench isolation layer STI is positioned between a word line WL and a passing word line PWL, no transistor is generated, but a capacitor C. This way, respective capacitors C of a DRAM memory device may be formed at particular crosspoints between the bit line BL with the word lines WL and the passing word lines PWL. Between the word line WL and the passing word line PWL, respective polymer memory devices P are formed in the spacing at the crosspoints with the bit line BL.
The dashed ovals illustrate memory cells which comprise both a volatile memory device and a non-volatile memory device. These memory cells each comprise a bit line contact BK, a capacitor C of a DRAM memory device, a selection transistor T that is generated at the crosspoint with the word line WL, a polymer memory device P, and a zener diode.
word line WLi
control gate line CGi
plate line Platei
word line WLi+1
control gate line CGi+1
plate line Platei+1
These electric lines are crossed orthogonally by bit lines BLj and BLj+1. The volatile DRAM memory device comprises a word line transistor or selection transistor 3, respectively, and a capacitor 9, which are correspondingly controlled via the word lines WL and the bit lines BL. To this end, the word lines WLi+1 are connected with the gates of the selection transistors 3 while the bit lines are connected with the source/drain paths of the selection transistors 3.
The non-volatile memory device comprises a polymer memory device 5 and an additional transistor 6. On the one side, the polymer memory device 5 is connected with the additional transistor 6, and on the other side with the plate line. The control gate of the additional transistor 6 is contacted and controlled via the control gate line CG. Via a node, the additional transistor 6 is connected with the capacitor 9 of the DRAM memory device. This structure recurs as many times as desired, which is indicated by the orders i and i+1.
With the folded bit line concept illustrated in
In each of
word line WL
control gate line CG
plate line Plate
plate line Plate
control gate line CG
word line WL
These electric lines are crossed orthogonally by bit lines BL. On the bit line BL, bit line contacts BK for contacting the corresponding bit line BL are arranged. At the crosspoints of the two plate lines Plate with the bit lines BL, respective polymer memory devices P are formed. Between the word lines WL and the control gate lines CG that are arranged in parallel to each other, respective capacitors C of a DRAM memory device are formed on the bit lines BL.
By the dashed ovals, memory cells are illustrated that comprise both a volatile memory device and a non-volatile memory device. These memory cells each comprise a bit line contact BK, a capacitor C of a DRAM memory device, a selection transistor T that is generated at the crosspoint with the word line WL, a control transistor T arranged at the crosspoints of word line WL and the control gate line CG, and a polymer memory device P.
The open bit line concept illustrated in
plate line Plate
word line WL
control gate line CG
plate line Plate
word line WL
These electric lines are crossed orthogonally by bit lines BL. Between the plate line Plate and the word line WL that runs parallel thereto, the bit lines BL are each provided with bit line contacts BK for contacting the corresponding bit line BL. At the crosspoints of the two plate lines Plate with the bit lines BL, respective polymer memory devices P are formed. Between the word line WL and the parallel control gate line CG, respective capacitors C of a DRAM memory device are formed on the bit lines BL.
The dashed ovals illustrate memory cells that comprise both a volatile memory device and a non-volatile memory device. These memory cells each comprise a bit line contact BK, a capacitor C of a DRAM memory device, a selection transistor T, a control transistor T, and a polymer memory device P.
passing plate line PPlate
passing control gate line PCG
word line WL
passing word line PWL
control gate line CG
plate line Plate
passing plate line PPlate
passing control gate line PCG
word line WL
passing word line PWL
control gate line CG, etc.
These electric lines are crossed orthogonally by bit lines BL. The dashed line S2 running through the upper bit line indicates the section plane of the lateral view of
At the crosspoints of the two passing word lines PWL with the bit lines BL, either capacitors C of a DRAM memory device or transistors T are formed, wherein these are the above-described word line transistors or selection transistors, respectively, of the DRAM memory device. At particular crosspoints of the two control gate lines CG and at the crosspoints of the two passing control gate lines PCG with the bit lines BL, transistors T are formed, wherein these are the above-described additional transistors of the polymer memory devices.
The dashed ovals illustrate memory cells that comprise both a volatile memory device and a non-volatile memory device. These memory cells each comprise a bit line contact BK, a capacitor C of a DRAM memory device, a selection transistor T, an additional control transistor T, and a polymer memory device P.
passing control gate line PCG
word line WL
passing word line PWL
control gate line CG
passing control gate line PCG
word line WL
passing word line PWL
control gate line CG, etc.
In a second plane above the first plane, the following electric lines are arranged:
plate line Plate
passing plate line PPlate
Below the passing word line PWL there is formed a trench capacitor TK, wherein a U-shaped buried strap BS is arranged as a conducting transistor contact means in the upper portion of the trench capacitor TK. On both sides of the U-formed buried strap BS, an N+ doped diffusion region each is provided. Between the buried strap BS and the passing word line PWL there is formed a thick oxide layer Ox to electrically isolate the two N+ doped diffusion regions adjacent to the buried strap BS from each other.
Below the word line WL, a gate oxide GO is formed, whereby the N+ diffusion region below the tungsten via W of the bit line BL is electrically coupled to the adjacent N+ doped diffusion region that is adjacent to the buried strap BS above the trench capacitor TK. This way, the N+ doped diffusion region below the tungsten via W of the bit line BL forms, together with the N+ doped diffusion region that is adjacent to the buried strap BS above the trench capacitor TK, the abovementioned selection transistor of the DRAM memory device which is controlled via the word line WL.
Between the passing word line PWL and the trench capacitor TK, the forming of a transistor is prevented by a thick oxide layer Ox, and the capacitor for the DRAM memory device is formed instead. Below the thick oxide layer Ox, the U-shaped buried strap BS produces a conducting transistor connection between the trench capacitor TK and the N+ doped source/drain region of the selection transistor.
A further N+ doped region below the control gate line CG is connected via a tungsten contact WK and suitable contact layers K with a polymer memory device P that is arranged between the plane with the plate line Plate and the plane with the control gate line CG. Above the polymer memory device P there is positioned the plate line Plate that is connected with the polymer memory device P also via a suitable upper contact layer K. The polymer memory device P may comprise several layers, as has been described initially in conjunction with polymer memory devices. Accordingly, the material for the contact layer K above and below the polymer memory device P is, as a rule, different and depends from the type of polymer memory device P used.
Below the control gate line CG, a further gate oxide layer GO is formed, whereby the N+ doped diffusion region below the tungsten contact WK of the polymer memory device P is electrically coupled with the adjacent N+ diffusion region that is adjacent to the buried strap BS above the trench capacitor TK. This way, the N+ diffusion region below the tungsten contact WK of the polymer memory device P forms, together with the N+ diffusion region that is adjacent to the buried strap BS above the trench capacitor TK, the above-mentioned additional transistor of the polymer memory device P, which is controlled via the control gate line CG.
Below the passing control gate line PCG there is positioned a shallow trench isolation layer STI. The shallow trench isolation layer STI prevents an electric coupling between the N+ doped diffusion region below the tungsten vias W and the N+ doped diffusion region that is adjacent to the U-shaped buried strap BS. This way, the N+ doped source/drain region of the selection transistor is electrically isolated from the N+ doped source/drain region of the additional transistor and thus from the tungsten contact WK of the polymer memory device P.
Number | Date | Country | Kind |
---|---|---|---|
10 2004 052 586.2 | Oct 2004 | DE | national |
10 2005 045 312.0 | Sep 2005 | DE | national |