Electrically-Flash-Erasable and Electrically-Programmable Memory Storage Devices With Self Aligned Tunnel Dielectric Area

The present invention relates to an electrically-flash-erasable and electrically-programmable memory device with thin tunnel dielectric area self aligned to the floating gate and the method of fabricating the same.

Since the invention of non-volatile memory cells having both electrically erasable and electrically programmable capabilities as disclosed in U.S. Pat. No. 4,115,914 issued to Harari on Sept. 26, 1978; U.S. Pat. No. 4,203,158 issued to Frohman-Bentchkowsky et al on May 13, 1980, the high volume commercial production of EEPROMs that employ the thin tunnel dielectrics for electrically erasing and electrically programming has gradually become a reality. These EEPROMs consist of one selection device in series with a memory device that has a floating gate over the channel and the tunnel dielectric area on the drain, and a stacked control gate over the floating gate. The programming of the memory device is achieved by applying a suitable potential across the drain and the control gate of the memory device to cause charge carriers to tunnel through the tunnel dielectric from the floating gate to the drain. The erasing of the memory device is achieved by applying a suitable potential across the control gate and the drain of the memory device to cause charge carriers to tunnel through the tunnel dielectric from the drain to the floating gate. In order to minimize the memory cell size, a flash EEPROM memory cell was introduced. The flash EEPROM cell is the hybrid of EPROM and EEPROM memory cell. The most common type of flash EEPROM is similar to EPROM in structure except the gate oxide underneath the floating gate is in the range of 70 A to 200 A. The programming of the flash EEPROM is achieved by applying a suitable potential at the drain and the control gate and grounding the source as in the EPROM to inject channel hot charge carriers into the floating gate. The erasing of the flash EEPROM is achieved by applying a suitable potential at the source and grounding the control gate to tunnel charge carriers from the floating gate to the source through the thin gate oxide. The key drawback of this flash EEPROM is the large thin gate oxide area under the whole floating gate. The yield and the reliability of the flash EEPROM are limited by the defect density of the large thin gate oxide area. Another drawback of this flash EEPROM is the low gate controlled junction breakdown at the source junction due to the thin gate oxide over the source junction.

It is therefore the principal object of this invention to provide an improved version of a flash EEPROM memory device. In the proposed version of the flash EEPROM memory device, the floating gate is over the channel area and insulated from the channel by 200 to 1000 A of gate oxide, and is also over the thin tunnel dielectric area at the source and insulated from the source by 70 A to 200 A of tunnel dielectric. Another improvement of the proposed version of the flash EEPROM memory device is that the tunnel dielectric area is small and self aligned to the floating gate.

It is another object of the present invention to provide a method for fabricating an electrically-flash-erasable and also electrically-programmable memory storage device having a floating gate structure separated from the channel by gate oxide and from the source by tunnel dielectric, a tunnel dielectric area self aligned to the floating gate, and a stacked control gate over the floating gate and insulated from the floating gate by interpoly dielectric.

These and other objects of the present invention are accomplished by a method of fabricating the flash EEPROM cells on a semiconductor substrate. The method includes the steps of etching the first layer of poly silicon into the floating gate, and sealing the sidewall of the floating gate with nitride to prevent the oxidation of the floating gate sidewall during the tunnel dielectric formation. The nitride is removed from all areas except those at the floating gate sidewall using the property of anisotropic etch. After implanting impurity to form the buried source drain region, oxide is removed from the tunnel dielectric area and the tunnel dielectric is formed. Then, the nitride is removed in hot phosphoric acid and a layer of add-on poly silicon is deposited and doped. The property of anisotropic poly silicon etch is used to remove all add-on poly silicon except those at the floating gate sidewall only. The original floating gate with the add-on poly silicon at the sidewall becomes the new floating gate. Oxide is then grown to seal the edge of the add-on poly silicon and the exposed tunnel dielectric area. The small area of tunnel dielectric under the add-on poly silicon becomes the new tunnel dielectric area which is self aligned to the original floating gate. This method of forming the small self aligned tunnel dielectric area can be incorporated into the non-self aligned and the self aligned flash EEPROM fabrication processes.

The novel features of the invention are set forth in the appended claims. The invention itself and other features as well as advantages thereof, can best be understood by reference to the detailed description which follows, read in conjunction with the accompanying drawings, wherein:

FIG. 1 is the top view and the cross section view along AA and BB of a pair of non-self aligned flash EEPROM memory cells constructed in accordance with the first preferred embodiment of the present invention; and

FIG. 2A and FIG. 2B are the cross section view along AA and BB illustrate the first poly silicon floating gate formation stage during the front end processing steps in accordance with the first preferred embodiment of the present invention; and

FIG. 3A and FIG. 3B are the cross section view along AA and BB illustrate the sealing nitride deposition after the first poly silicon floating gate formation in accordance with the first preferred embodiment of the present invention; and

FIG. 4A and FIG. 4B are the cross section view along AA and BB illustrate the formation of the sealing nitride at the sidewall of the first poly silicon floating gate in accordance with the first preferred embodiment of the present invention; and

FIG. 5A and FIG. 5B are the cross section view along AA and BB illustrate the impurity implant to form the buried source drain for the flash EEPROM in accordance with the first preferred embodiment of the present invention; and

FIG. 6A and FIG. 6B are the cross section view along AA and BB illustrate the oxide removal from the tunnel dielectric area in accordance with the first preferred embodiment of the present invention; and

FIG. 7A and FIG. 7B are the cross section view along AA and BB illustrate the tunnel dielectric formation in accordance with the first preferred embodiment of the present invention; and

FIG. 8A and FIG. 8B are the cross section view along AA and BB illustrate the add-on poly silicon deposition after the removal of the sealing nitride from the sidewall of the first poly silicon floating gate in accordance with the first preferred embodiment of the present invention; and

FIG. 9A and FIG. 9B are the cross section view along AA and BB illustrate the formation of the add-on poly silicon at the sidewall of the first poly silicon floating gate to form the new floating gate in accordance with the first preferred embodiment of the present invention; and

FIG. 10A and FIG. 10B are the cross section view along AA and BB illustrate the formation of the interpoly dielectric in accordance with the first preferred embodiment of the present invention; and

FIG. 11A and FIG. 11B are the cross section view along AA and BB illustrate the formation of the second poly control gate in accordance with the first preferred embodiment of the present invention; and

FIG. 12 is the top view and the cross section view along AA and BB of a pair of self aligned flash EEPROM memory cells constructed in accordance with the second preferred embodiment of the present invention; and

FIG. 13A and FIG. 13B are the cross section view along AA and BB illustrate the oxidation after the second poly silicon control gate formation during the front end processing steps in accordance with the second preferred embodiment of the present invention; and

FIG. 14A and FIG. 14B are the cross section view along AA and BB illustrate the interpoly nitride removal after the second poly silicon control gate oxidation in accordance with the second preferred embodiment of the present invention; and

FIG. 15A and FIG. 15B are the cross section view along AA and BB illustrate the sealing nitride deposition after the first poly silicon floating gate formation in accordance with the second preferred embodiment of the present invention; and

FIG. 16A and FIG. 16B are the cross section view along AA and BB illustrate the oxide removal from the tunnel dielectric area in accordance with the second preferred embodiment of the present invention; and

FIG. 17A and FIG. 17B are the cross section view along AA and BB illustrate the photoresist protecting the new floating gate that consists of the add-on poly silicon at the sidewall of the first poly silicon floating gate, during the removal of all add-on poly silicon at the sidewall of all other poly silicon patterns in accordance with the second preferred embodiment of the present invention; and

FIG. 18A and FIG. 18B are the cross section view along AA and BB illustrate the new floating gate that consists of the add-on poly silicon at the sidewall of the first poly silicon floating gate in accordance with the second preferred embodiment of the present invention.

The method of the present invention is utilized to fabricate flash EEPROMs as illustrated in the first preferred embodiment shown in FIG. 1 and the second preferred embodiment shown in FIG. 12.

FIG. 1 shows the first preferred embodiment for a pair of mirror-image non-self aligned flash EEPROMs. The flash EEPROM 1 is constructed on a P-type substrate 9 of monocrystalline silicon or the like. It includes a floating gate 30 disposed over the channel area and separated from the channel area by gate oxide 21, an add-on floating gate 31 shorted electrically to the floating gate 30 and disposed over the buried source and separated from the buried source by the tunnel dielectric 22, a control gate 50 disposed over the floating gate 30 and the add-on floating gate 31 and separated from these floating gate by interpoly dielectric 40 and 41, and buried N+ junction 10 serving as the buried source and the buried drain of the flash EEPROM. The buried source and the buried drain are disposed under the add-on floating gate 31 and disposed under the control gate 50 and separated from the control gate 50 by interpoly dielectric 41 and an oxide layer 23. The shallower junction 11 forms the source drain of the periphery transistors. The metal line 80 is connected to the drain of all flash EEPROMs at the same row through contact 70. The doped CVD oxide film 60 is underneath the metal to smooth out the sharp edges. The thick field oxide 20 serves as the isolation between different devices.

When a suitable potential is applied to the control gate 50 and to the buried drain while grounding the buried source, the hot channel electrons are injected into the floating gate 30 and 31. On the other hand, when a suitable potential is applied to the buried source while grounding the control gate 50, the electrons are tunneled from the add-on floating gate 31 to the buried source through tunnel dielectric 22. If there are excess electrons at the floating gate 30 and the add-on floating gate 31 of the flash EEPROM 1, the positive voltage required at the control gate 50 to conduct current through the flash EEPROM 1 becomes higher. On the other hand, if there are deficit in electrons at the floating gate 30 and the add-on floating gate 31 of the flash EEPROM 1, the positive voltage required at the control gate 50 to conduct current through the flash EEPROM 1 becomes less. Therefore, whether there are current or there are no current conducting through the flash EEPROM 1 at a given voltage at the control gate 50 can represent whether 1 or 0 is stored in the flash EEPROM 1.

FIG. 1 also shows how the flash EEPROMs are connected into a memory array. The diffusion line 220 is the common source line of the flash EEPROMs at two adjacent columns, the control gate 50 is the common control gate line for all flash EEPROMs at the same column, and the metal bit line 80 connects the drain of all flash EEPROMs at the same row through the contact 70. All flash EEPROMs can be placed as the mirror image of each other as shown in FIG. 1.

The fabrication of a memory array incorporating the paired flash EEPROMs shown in FIG. 1 will now be described in details. For the convenience of explanation, an N-channel flash EEPROM array will be used as an example. A P-type single crystal silicon will be used as the starting substrate. As in the conventional fabrication process for the N-channel EPROM, a CVD nitride thin film is deposited on the initial oxide grown on the P-type substrate. Then the photoresist patterns of the active area is formed on top of the nitride film, and the nitride outside of the photoresist patterns is etched off. P-type impurity Boron is then implanted into the area outside of the photoresist patterns to raise the substrate doping outside of the active area. This is done to raise the threshold voltage of the parasitic field devices to form better isolation between the active areas. The photoresist patterns are stripped off after the Boron implant, and the thick field oxide 20 is grown outside of the active area. During the field oxidation, the active areas 25 and 220 are still protected by nitride pattern and no oxide is grown. After the field oxide 20 is grown, the nitride and the initial oxide underneath the nitride in the active areas 25 and 220 are stripped off to expose the substrate in the active area only. Then a high quality gate oxide 21 of 200 A to 1000 A is grown in the active areas 25 and 220, and the photoresist patterns are formed to expose the flash EEPROMs area for 5E11 to 5E12 /cm2 dose of boron implant. After the photoresist patterns are stripped off, the first layer of poly silicon 30 is deposited and doped. An oxide layer 39 is then grown on the first layer of poly silicon 30. The photoresist patterns are formed to retain the floating gate 30 with the oxide 39, and remove the poly silicon from the area outside of the photoresist patterns. After the photoresist patterns are stripped of, the cross section along AA and BB are shown in FIG. 2A and FIG. 2B. A sealing nitride film 49 is then deposited and oxidized as shown in FIG. 3A and FIG. 3B. The anisotropic etch is used to remove all nitride except sealing nitride 48 at the sidewall of the floating gate 30 as shown in FIG. 4A and FIG. 4B. Then the photoresist patterns are formed to expose the flash EEPROMs area for the buried N+ implant 10 at the dose of 5E13 to 5E15 /cm2 as shown in FIG. 5A and FIG. 5B. After the photoresist patterns are stripped off, the new photoresist patterns 29 are formed to expose the tunnel dielectric area at the source of the flash EEPROM, and the oxide is etched off to expose the substrate 27 as shown in FIG. 6A and FIG. 6B. The photoresist patterns are again stripped off and the tunnel dielectric 28 is then formed as shown in FIG. 7A and FIG. 7B. The sealing nitride at the sidewall of the floating gate 30 is removed in hot phosphoric acid afterward. The add-on poly silicon layer 38 is deposited and doped as shown in FIG. 8A and FIG. 8B. The anisotropic poly silicon etch is used to remove all add-on poly silicon film except the add-on poly silicon 31 at the sidewall of the floating gate 30 to form the add-on floating gate 31 as shown in FIG. 9A and FIG. 9B. The tunnel dielectric area 22 under the add-on floating gate becomes the new small tunnel dielectric area. The oxide layer 39 on the floating gate 30 is then removed and the high quality oxide layer 40 is grown on the floating gate 30 and oxide 23 is also grown in the exposed tunnel dielectric area at the same time. The high quality nitride film 41 is deposited and oxidized afterward as shown in FIG. 10A and FIG. 10B. Then the photoresist patterns are formed to protect the flash EEPROMs area and the first poly gate area during the removal of the nitride 41 and oxide 23. The high quality gate oxide for the second poly silicon devices is grown and the impurity implant is done to adjust the threshold of the second poly silicon devices. The second layer of poly silicon is then deposited and doped. The photoresist patterns are used to pattern the control gate 50 and the second poly silicon gate patterns. The control gate formation after the resist removal is shown in FIG. 11A and FIG. 11B. The remaining fabrication process steps are similar to that of the conventional N-channel EPROMs. After the formation of second poly patterns, the shallow source drain junction 11 is implanted, doped CVD oxide film is deposited, contacts are open, and metal film is deposited and patterned.

FIG. 12 shows the second preferred embodiment for a pair of mirror-image self aligned flash EEPROMs. The flash EEPROM 1 is constructed on a P-type substrate 9 of monocrystalline silicon or the like. It includes a floating gate 30 disposed over the channel area and separated from the channel area by gate oxide 21, an add-on floating gate 31 shorted electrically to the floating gate 30 and disposed over the buried source and separated from the buried source by the tunnel dielectric 22, a control gate 50 disposed over the floating gate 30 and separated from the floating gate 30 by interpoly dielectric 40 and 41, and buried N+ junction 10 serving as the buried source and the buried drain of the flash EEPROM. The buried source and the buried drain are disposed under the add-on floating gate 31. The add-on floating gate 31 is disposed at the sidewall of the second gate 50 and insulated from the second gate 50. The shallower junction 11 forms the source drain of the periphery transistors. The metal line 80 is connected to the drain of all flash EEPROMs at the same row through contact 70. The doped CVD oxide film 60 is underneath the metal to smooth out the sharp edges. The thick field oxide 20 serves as the isolation between different devices.

When a suitable potential is applied to the control gate 50 and to the buried drain while grounding the buried source, the hot channel electrons are injected into the floating gate 30 and 31. On the other hand, when a suitable potential is applied to the buried source while grounding the control gate 50, the electrons are tunneled from the add-on floating gate 31 to the buried source through tunnel dielectric 22. If there are excess electrons at the floating gate 30 and the add-on floating gate 31 of the flash EEPROM 1, the positive voltage required at the control gate 50 to conduct current through the flash EEPROM 1 becomes higher. On the other hand, if there are deficit in electrons at the floating gate 30 and the add-on floating gate 31 of the flash EEPROM 1, the positive voltage required at the control gate 50 to conduct current through the flash EEPROM 1 becomes less. Therefore, whether there are current or there are no current conducting through the flash EEPROM 1 at a given voltage at the control gate 50 can represent whether 1 or 0 is stored in the flash EEPROM 1.

FIG. 12 shows how the flash EEPROMs are connected into a memory array. The diffusion line 220 is the common source line of the flash EEPROMs at two adjacent columns, the control gate 50 is the common control gate line for all flash EEPROMs at the same column, and the metal bit line 80 connects the drain of all flash EEPROMs at the same row through the contact 70. Flash EEPROMs are placed as the mirror image of each other in FIG. 12.

The fabrication of a memory array incorporating the paired flash EEPROMs shown in FIG. 12 can now be described in details. For the convenience of explanation, an N-channel flash EEPROM array will be used as an example. A P-type single crystal silicon will be used as the starting substrate. As in the conventional fabrication process for the N-channel EPROM, a CVD nitride thin film is deposited on the initial oxide grown on the P-type substrate. Then the photoresist patterns of the active area is formed on top of the nitride film, and the nitride outside of the photoresist patterns is etched off. P-type impurity Boron is then implanted into the area outside of the photoresist patterns to raise the substrate doping outside of the active area. This is done to raise the threshold voltage of the parasitic field devices to form better isolation between the active areas. The photoresist patterns are stripped off after the Boron implant, and the thick field oxide 20 is grown outside of the active area. During the field oxidation, the active areas 25 and 220 are still protected by nitride pattern and no oxide is grown. After the field oxide 20 is grown, the nitride and the initial oxide underneath the nitride in the active areas 25 and 220 are stripped off to expose the substrate in the active area only. Then a high quality gate oxide 21 of 200 A to 1000 A is grown in the active areas 25 and 220, and the photoresist patterns are formed to expose the flash EEPROMs area for 5E11 to 5E12/cm2 dose of boron implant. After the photoresist patterns are stripped off, the first layer of poly silicon 35 is deposited and doped. The photoresist patterns are then used to pattern the first poly silicon. After the photoresist patterns are stripped, the high quality oxide layer 45 is grown on the first poly silicon and the high quality nitride film 46 is deposited and oxidized afterward. Then the photoresist patterns are formed to protect the flash EEPROMs area and the first poly silicon area during the removal of the nitride 45 and oxide from the second poly silicon device area. The high quality gate oxide for the second poly silicon devices is grown and the impurity implant is done to adjust the threshold of the second poly silicon devices. The second layer of poly silicon is then deposited and doped. The photoresist patterns are used to pattern the control gate 50 and the second poly silicon gate patterns. After the photoresist is stripped off, the oxide layer 51 is grown on the second poly silicon control gate 50 and other second poly silicon patterns as shown in FIG. 13A and FIG. 13B. The nitride film 46 outside of the second poly silicon control gate is then etched off as shown in FIG. 14A and FIG. 14B. Then photoresist patterns are used to expose the flash EEPROMs only, and the first poly silicon outside of the second poly silicon control gate is etched off. After the photoresist patterns are stripped off, the sealing nitride film 59 is deposited and oxidized as shown in FIG. 15A and FIG. 15B. The anisotropic etch is used to remove all nitride except sealing nitride 58 at the sidewall of the floating gate 30. Then the photoresist patterns are formed to expose the flash EEPROMs area for the buried N+ implant 10 at the dose of 5E13 to 5E15/cm2. After the photoresist patterns are stripped off, the new photoresist patterns 69 are formed to expose the tunnel dielectric area at the source of the flash EEPROMs, and the oxide is etched off to expose the substrate 68 as shown in FIG. 16A and FIG. 16B. The photoresist patterns are again stripped off and the tunnel dielectric 28 is then formed. The sealing nitride at the sidewall of the floating gate 30 is then removed in hot phosphoric acid. The add-on poly silicon layer is deposited and doped. The anisotropic poly silicon etch is used to remove all add-on poly silicon film except the add-on poly silicon 31 at the sidewall of the floating gate 30 to form the add-on floating gate 31. The tunnel dielectric area 22 under the add-on floating gate 31 becomes the new small tunnel dielectric area. The photoresist patterns 67 are used to protect the flash EEPROM area during the removal of the add-on poly silicon from the sidewall of all other poly silicon patterns as shown in FIG. 17A and FIG. 17B. After the photoresist patterns are stripped off, the shallow source drain junction 11 is implanted and the flash EEPROM is shown in FIG. 18A and FIG. 18B. The remaining fabrication process steps are similar to that of the conventional N-channel EPROMs. After oxide is grown on the add-on poly silicon and the exposed tunnel dielectric area, doped CVD oxide film is deposited, contacts are open, and metal film is deposited and patterned.

The present invention has been set forth in the form of two preferred embodiments. Various modifications of the preferred embodiments disclosed herein may be made by those skilled in the art upon reference to this disclosure without departing from the scope and the spirit of the present invention. Moreover, such modifications are considered to be within the purview of the appended claims.