US 20070004135 A1
A memory charge storage device has regions of sacrificial material overlying a substrate (12). For each memory cell a first doped region (20) and a second doped region (24) are formed within the substrate and on opposite sides of one (16) of the regions of sacrificial material. A discrete charge storage layer (28) overlies the substrate and is between the regions of sacrificial material. In one form a control electrode (34) is formed per memory cell overlying the substrate with an underlying substrate diffusion and laterally adjacent one of the regions of sacrificial material. A third substrate diffusion (60) is positioned between the two control electrodes. In another form two control electrodes are formed per memory cell with a substrate diffusion underlying each control electrode. In both forms a select electrode (64) overlies and is between both of the two control electrodes.
1. A method for forming a storage device structure comprising:
providing a substrate of a first conductivity type;
forming overlying regions of sacrificial material;
forming a first doped region of a second conductivity type opposite the first conductivity type and a second doped region of the second conductivity type within the substrate and on opposite sides of one of the regions of sacrificial material;
after the steps of forming the overlying regions and forming the first doped region, forming a discrete charge storage layer overlying the substrate and between the regions of the sacrificial material;
forming a first conductive control electrode and a second conductive control electrode overlying the substrate and on opposite sides of the one of the regions of sacrificial material; and
forming a conductive select electrode overlying and between both of the first conductive control electrode and the second conductive control electrode.
2. The method of
forming a third doped region of the second conductivity type between the first doped region and the second doped region and between the first conductive control electrode and the second conductive control electrode.
3. The method of
forming the conductive select electrode also over the third doped region.
4. The method of
forming disposable sidewall spacers adjacent the overlying regions prior to forming the first doped region and the second doped region; and
removing the disposable sidewall spacers after forming the first doped region and the second doped region.
5. The method of
using nanoclusters to form the discrete charge storage layer.
6. The method of
using nitride as the sacrificial material.
7. The method of
8. A method for forming a storage device structure comprising providing a substrate;
forming a plurality of regions of sacrificial material overlying the substrate;
forming a disposable sidewall spacer on each of the regions of sacrificial material to decrease distance separating the regions of sacrificial material;
forming a plurality of diffusions in portions of the substrate that are not covered by either a disposable sidewall spacer or the regions of sacrificial material;
removing the disposable sidewall spacer on each of the regions of sacrificial material;
depositing a discrete charge storage layer and patterning the discrete storage layer to form a plurality of discrete charge storage materials that overlie each of the plurality of diffusions;
forming a plurality of control gates, each of the plurality of control gates overlying one of the plurality of discrete charge storage materials;
forming a gate dielectric layer in contact with a portion of the plurality of control gates; and
forming a select gate overlying each of the plurality of control gates and laterally between each of the plurality of control gates, wherein adjacent memory cells share the select gate.
9. The method of
forming another plurality of diffusions in the substrate, between the regions of sacrificial material and underlying the select gate, each of the another plurality of diffusions functioning as current electrode.
10. The method of
biasing the storage device structure to permit each of the plurality of diffusions to function as a source for a first of two storage bits in a single memory cell and as a drain for a second of the two storage bits.
11. The method of
forming removable sidewall spacers around each of the plurality of control gates prior to forming the another plurality of diffusions in the substrate, the removable sidewall spacers substantially centering the another plurality of diffusions in the substrate between the plurality of control gates; and
removing the removable sidewall spacers.
12. A storage device structure comprising:
a first plurality of diffusions formed in the substrate;
a discrete charge storage layer overlying each of the first plurality of diffusions;
a control gate overlying each discrete charge storage layer, and
a select gate overlying and between each control gate, wherein each of the first plurality of diffusions functions as a current electrode for the storage device structure.
13. The storage device structure of
a dielectric layer overlying and electrically isolating each control gate and discrete charge storage layer from each other;
14. The storage device structure of
a second plurality of diffusions formed in the substrate, each positioned between adjacent control gates, each of the second plurality of diffusions underlying the select gate.
15. The storage device structure of
16. The storage device structure of
17. The storage device structure of
18. The storage device structure of
This application is related to the following patent applications:
U.S. Patent application titled, “Source Side Injection Storage Device with Control Gates Adjacent to Shared Source/Drain and Method Therefor,” by Hong et al., having docket number SC14220TP, filed concurrently herewith, and assigned to the assignee hereof; and
U.S. patent application titled, “Source Side Injection Storage Device with Spacer Gates and Method Therefor,” by Hong et al., having docket number SC14169TP, filed concurrently herewith, and assigned to the assignee hereof.
This invention relates to non-volatile memories, and more particularly to storage devices in the non-volatile memories that use source side injection.
Source side injection (SSI) has been found to have benefits over regular hot carrier injection (HCI) used in the programming of non-volatile memories (NVMs). Programming by SSI is able to be performed at significantly lower power than programming by regular (HCI). This is particularly important in uses such as cell phones in which battery operation is very important. One of the disadvantages of SSI is that the storage devices require more area on the integrated circuit which increases cost. The design of the individual memory cells for SSI generally includes a transition in the gate structure over the channel which requires more area.
One of the techniques in the attempt to reduce the impact of the increased storage-device size has been the use of a virtual ground array (VGA) architecture. VGA has been known to require relatively small area compared to other architectures while increasing other difficulties such as read disturb. This has nonetheless been a popular architecture for low cost NVMs. Further reductions in space in the storage cell would further reduce size and thus cost.
The foregoing and further and more specific objects and advantages of the invention will become readily apparent to those skilled in the art from the following detailed description of a preferred embodiment thereof taken in conjunction with the following drawings:
In one aspect a storage device has a control gate that is shared by two memory cells and the drain for both memory cells is a first doped region directly under the control gate. The control gate, in the channel direction, completely covers this doped region. The source, in a second doped region, for a given memory cell is disposed away from the shared control gate of the given memory cell. The second doped region is shared by an adjacent memory cell that has a different control gate. This structure provides for reduced area while retaining the ability to perform programming by SSI. This is better understood by reference to the drawings and the following description.
In operation, a memory cell is defined as the structure between one doped region under a control gate and an adjacent doped region between control gates. Thus for example, one memory cell comprises doped region 22, doped region 60, and the portion of control gate 34 between doped regions 22 and 60, the portion of select gate 64 between doped regions 22 and 60. The channel is the portion of substrate 12 between doped regions 22 and 60 along the top surface of substrate 12. Oxide 47 functions as the gap dielectric where the electrons in the channel region gain energy for injection during programming where select gate 64 is closer to the channel than control gate 34. To program this memory cell, select gate 64 is biased to a voltage of about 2 to 3 volts, doped region 60, which functions as a source, is grounded, control gate 34 is biased to a voltage of about 5 to 6 volts, and doped region 22, which functions a drain, is biased to about 5 volts. This establishes a current flow through the channel. Electrons come from the source, doped region 60, and are injected into storage layer 42 at the edge of control gate 34 which is closest to the source. Thus, source side injection is achieved. This is continued until program layer 42 has captured sufficient electrons for providing enough bias to significantly impede current in the channel during a read operation. During a read, both select gate 64 and control gate 34 are biased sufficiently to cause measurable current flow through the channel in the absence of being programmed. In the programmed condition, the accumulation of electrons in storage layer 42 prevents the channel from inverting in that the region of the channel immediately under where the electrons accumulated. This impedes current between doped regions 60 and 22 so that the difference in current flow between being programmed and not programmed is significant and can be measured. Erase is performed by biasing control gate 34 to about minus 6 volts and substrate 12 to about plus 6 volts.
Similar operation is applied to the other memory cells shown in
In the case of storage device 10′ of
As an example, for programming the bit at storage layer 42′ between doped regions 22′ and 24′, doped region 22′ functions as the drain and doped region 24′ functions as the source. Control gate 36′ is positively biased to invert the channel thereunder, select gate 64′ is positively biased to invert the channel between control gates 34′ and 36′, control gate 34′ is positively biased to attract electrons into storage layer 42′, and doped region 22′ is positively biased to induce channel current that generates hot carriers. Similarly for the bit represented at storage layer 44′ that is between doped regions 22′ and 24′, programming is achieved by reversing the biases at doped regions 22′ and 24′ and at control gates 34′ and 36′ while keeping substrate 12 and select gate 64′ the same.
Reading is achieved by using doped region 22′ as the drain for reading the bits in storage layer 42′ and using doped region 24′ as the drain while reading the bits at storage layer 44′. Programming is achieved in the same way as for storage device 10 of
Various changes and modifications to the embodiments herein chosen for purposes of illustration will readily occur to those skilled in the art. For example, these embodiments have been shown using a bulk silicon substrate but another substrate type, such as semiconductor on insulator (SOI) or SOI hybrid, could also be used. Also, hot carrier injection (HCI) can also be used in conjunction with SSI to cause programming at interior portions of the storage layers. Thus for example programming could be at both the left and right side of storage layer 42 close to doped region 22. By injecting electrons in the central area of storage layer 42 using HCI and injecting electrons at the lateral outside edge of storage layer 42 using SSI, two bits of information may be programmed on each side of storage layer 42. Literally then, storage layer 42 could actually store four bits. To the extent that such modifications and variations do not depart from the spirit of the invention, they are intended to be included within the scope thereof which is assessed only by a fair interpretation of the following claims.