US 5,912,487 AGrant
Split Gate Flash EEPROM Memory Cell Structure
Issue Date:1999-06-15
•10 Claims
•6 Drawing Sheets
Abstract
A MOSFET device on a lightly doped semiconductor substrate comprises forming a dielectric layer on the substrate, a floating gate layer over the dielectric layer, a sacrificial layer on the floating gate layer, and a split-gate channel mask patterned with openings over the sacrificial layer.
Metadata
Assignee
- United Microelectronics Corp.
Inventor
- Gary Hong
Application Information
Application Number:US 9126090
Filing Date:1997-08-18
Priority Date:1994-11-23
Art Unit:285
Classifications
IPC:
H01L 29788
Field of Search:
257261;314;315;316;344;336;408
Patent Drawings (6 sheets)
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to semiconductor memories and more particularly to flash EEPROM cells and the method of manufacture thereof.
2. Description of Related Art
FIGS. 1A and 1B show a prior art device as described in Samachisa et al "A 128K Flash EEPROM Using Double-Poly-silicon Technology", IEEE Solid State Circuit Vol SC-22, No. 5, pp 676-683 (Oct. 1987). FIG. 1A shows a cross section of a flash EEPROM cell 9 and FIG. 1B shows a plan view of the flash EEPROM cell 9 of FIG. 1A. Referring to FIG. 1A, EEPROM cell 9 includes two spaced apart n+ regions in the upper surface of substrate 10. Above the substrate 10 and those two n+ regions is a floating gate FG near the right hand one of the two n+ regions and above floating gate FG is the control gate CG which traverses the floating gate FG and bridges between the ends of the two n+ regions.
The drawback of the device of FIGS. 1A and 1B is that the channel length L1 of the floating gate FG and isolation transistor length L2 depend on the photolithography alignment between polysilicon 1 layer (FG) and polysilicon 2 layer (CG), even though the total channel length L=L1+L2 is fixed. In summary, the channel length is not easily controlled, because it depends upon alignment.
FIG. 2 shows a split gate prior art memory device (see U.S. Pat. No. 4,868,629) with a substrate 50 with n++ regions 54a and 54b in the surface of the substrate 50. Between the n++ regions are floating gate 52 and photoresist 53-1. The split-gate device uses photoresist 53 to form an offset channel for an isolation transistor. Therefore alignment of photoresist 53-1 with floating gate 52 determines the channel length of the isolation transistor and the total channel length of the split-gate device. The problem is the use of photoresist (53-1) to define the offset for a split-gate structure, which depends upon the alignment of the photolithography process.
Several prior art patents describing split-gate, flash EEPROM processes and structures are as follows:
U.S. Pat. No. 5,280,446 of Ma et al for "Flash EPROM Memory Circuit Having Source Side Programming;"
U.S. Pat. No. 5,274,588 of Manzur et al for "Split-Gate for an EEPROM"; and
U.S. Pat. No. 5,194,925 of Ajika et al for "Electrically Programmable Non-Volatile Semiconductor Memory Device."
SUMMARY OF THE INVENTION
In accordance with this invention, a method is provided for fabricating an MOSFET device on a lightly doped semiconductor substrate comprises
forming a first dielectric layer on the substrate,
forming a floating gate layer over the first dielectric layer,
forming a sacrificial layer on the floating gate layer,
forming a split-gate channel mask with a pattern of openings over the sacrificial layer,
etching the sacrificial layer to remove material beneath the openings in the mask and etching the floating gate layer to remove material beneath the openings in the mask to form a self-aligned channel mask for ion implanting the source/drain regions of the device,
isotropically overetching the floating gate layer to form a floating gate therefrom,
ion implanting a doped source region and a doped drain region in the substrate defining the channel of the device,
removing the channel mask,
removing the remainder of the sacrificial layer,
forming a drain side mask over the floating gate and the source region,
ion implanting a drain side N- region in the substrate,
forming a blanket interpolysilicon dielectric layer over the device,
forming a blanket deposit of a control gate layer over the interpolysilicon layer,
forming a control gate mask with a pattern of openings on the device,
patterning the control gate layer by etching away material from the control gate layer through the openings in the control gate mask, and
removing the control gate mask.
Preferably, the first dielectric layer comprises PSG having a thickness within a range between about 200 .ANG. and about 1000 .ANG.; the interpolysilicon dielectric layer comprises ONO; the floating gate layer comprises a polysilicon layer having a thickness within a range between about 500 .ANG. and about 2,000 .ANG.; the control gate layer comprises a polysilicon layer having a thickness within a range between about 1,000 .ANG. and about 5,000 .ANG.; the source/drain implant comprises implantation of arsenic ions with a concentration of about 5E15/cm.sup.2 at an energy of about 50 keV; and the drain side implant comprises implantation of phosphorus ions with a concentration of about 1E14/cm.sup.2 at an energy of about 50 keV.
In accordance with another aspect of this invention, a method of fabricating an MOSFET device on a lightly doped semiconductor substrate comprises
forming a tunnel oxide dielectric layer on the substrate,
forming a polysilicon 1 floating gate layer over the tunnel oxide dielectric layer,
forming a PSG sacrificial layer on the floating gate layer,
forming a split-gate channel mask over the sacrificial layer,
etching the sacrificial layer to remove material beneath the openings in the mask and etching the floating gate layer to remove material beneath the openings in the mask to form a self-aligned channel mask for ion implanting the source/drain regions of the device,
isotropically overetching the floating gate layer to form a floating gate therefrom,
ion implanting a doped source region and a doped drain region in the substrate in a self-aligned process to define the channel of the device,
removing the split-gate channel mask,
removing the remainder of the sacrificial layer,
forming a drain side mask over the floating gate and the source region,
ion implanting a drain side N- region in the substrate,
forming a blanket ONO interpolysilicon dielectric layer over the device,
forming a blanket deposit of a polysilicon 2 control gate layer over the interpolysilicon layer,
forming a control gate mask on the device,
patterning the control gate layer by etching away material from the control gate layer through the openings in the control gate mask, and
removing the control gate mask.
Preferably, the floating gate layer comprises a polysilicon layer having a thickness within a range between about 500 .ANG. and about 2,000 .ANG.; the control gate layer comprises a polysilicon layer having a thickness within a range between about 1,000 .ANG. and about 5,000 .ANG.; the source/drain implant comprises implantation of arsenic ions with a concentration of about 5E15/cm.sup.2 at an energy of about 50 keV; and the drain side implant comprises implantation of phosphorus ions with a concentration of about 1E14/cm.sup.2 at an energy of about 50 keV.
In accordance with still another aspect of this invention, a MOSFET device on a lightly doped semiconductor substrate comprises
a tunnel oxide layer on the substrate,
a floating gate over the first tunnel oxide layer,
an ion implanted doped source region and a doped drain region in the substrate defining the channel of the device self-aligned with the floating gate,
an ion implanted drain side N- region in the substrate,
a blanket ONO interpolysilicon dielectric layer over the device,
a control gate extending over the interpolysilicon layer,
a control gate mask on the device, and
the control gate layer having been patterned by etching away portions of the control gate layer unprotected by the control gate mask.
Preferably, the floating gate layer comprises a polysilicon layer having a thickness within a range between about 500 .ANG. and about 2,000 .ANG.; control gate layer comprises a polysilicon layer having a thickness within a range between about 1,000 .ANG. and about 5,000 .ANG.; the source/drain regions are doped with arsenic ions implanted with a concentration of about 5E15/cm.sup.2 at an energy of about 50 keV; the drain side implant comprises phosphorus ions implanted with a concentration of about 1E14/cm.sup.2 at an energy of about 50 keV.
It is also preferred that, the device can operate in accordance with the table as follows:
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing and other aspects and advantages of this invention are explained and described below with reference to the accompanying drawings, in which:
FIGS. 1A and 1B show a prior art flash EEPROM device.
FIG. 2 shows a split gate prior art memory device.
FIG. 3A-3J show a cross-sectional view of a fragment of a split-gate flash EEPROM device illustrating the process flow during the manufacture of a device in accordance with this invention.
DESCRIPTION OF THE PREFERRED EMBODIMENT
This invention provides a structure and a process sequence for producing a flash EEPROM memory device.
FIG. 3A-3J show a cross-sectional view of a fragment of a split-gate flash EEPROM device 20 illustrating the process flow during the manufacture of a device 20 in accordance with this invention.
Referring to FIG. 3J, device 20 is formed on a silicon substrate 21 in which are formed N+ source/drain (S/D) regions 26/27, and N- implant region 48 (for separating the drain N+ region from the floating gate 24 in a split-gate configuration, as explained below.) Between the N+ source region 26 and the N- region 48 is the channel 28. On the surface of the substrate 20 above the channel 28 and N- region 48 is a tunnel oxide layer 22, above which is formed a polysilicon 1 floating gate 24. Above floating gate 24, exposed tunnel oxide layer 22 is formed a conformal interpolysilicon dielectric layer 44 preferably composed of an ONO thin film as described below. Extending across dielectric layer 44 is a polysilicon 2 word line layer formed into the configuration of a control gate 50 formed on the device 20.
PROCESS
Front end process
In the early stages of fabrication in accordance with the method of this invention, a P- doped silicon substrate 21 is produced by a conventional doping process.
Active Masking/Etching FIG. 3A shows a cross-sectional view of a fragment of a device 20 with a P- substrate 21.
Tunnel Oxidation
By the conventional process of tunnel oxidation the surface of substrate 21 is covered with a blanket, first dielectric, tunnel oxide layer 22. Layer 22 has a thickness within a range between about 60 .ANG. and about 100 .ANG. in accordance with the conventional process of growing the tunnel oxide (silicon dioxide), which is thermal oxidation in dry O.sub.2 and N.sub.2 (e.g. at a temperature of about 850.degree. C. for about 10 minutes.)
Polysilicon 1 Deposition and Doping
Then, still referring to FIG. 3A, a blanket polysilicon 1 layer 24 (which is to be patterned into a floating gate as shown in FIG. 3D) is deposited to a range of thicknesses between about 500 .ANG. and about 2,000 .ANG. by the conventional LPCVD (low pressure chemical vapor deposition) process. Polycide is not a substitute for polysilicon layer 24 in this case.
Subsequently, the polysilicon 1 layer 24 is doped in a blanket application of arsenic (As) by diffusion or by ion implantation of dopant ions 34. In the case of ion implantation, the arsenic dopant 34 is applied with a dose within a range between about 1.times.10.sup.14 /cm.sup.2 and about 1.times.10.sup.16 /cm.sup.2 within a range of energies between about 30 keV and about 100 keV. An alternative way of doping the polysilicon 1 layer 24 is a conventional diffusion process in which phosphoryl chloride POCl.sub.3 is diffused into the layer 24 at a temperature within the range between about 800.degree. C. and about 900.degree. C. for a time within the range between about 10 minutes and about 30 minutes.
Formation of Sacrificial (PSG) Layer
A sacrificial, phosphosilicate glass (PSG) layer 38 is formed over the surface of polysilicon 1 layer 24. Layer 38 is deposited to a preferred range of thicknesses within a range between about 200 .ANG. and about 1,000 .ANG. by a conventional process of forming a PSG layer, such as use of phosphine (PH.sub.3) gas as the source of phosphorus in a CVD silicon dioxide (SiO.sub.2) formation process to form phosphosilicate glass. The source of the SiO.sub.2 in the process is SiH.sub.4 and O.sub.2 input gases.
Total Channel (Split-Gate) Masking
Referring to FIG. 3C a photoresist mask 40 was formed over the layer 24 and mask 40 as seen in FIGS. 3C, above the polysilicon 1 (floating gate) layer 24.
Mask 40 extends across the entire width of the portion of the substrate 21 to be formed into the channel 28 of the device 20. The channel 28 to be formed can be seen in FIGS. 3D-3J as described below.
Etching of Sacrificial (PSG) Layer
Referring again to FIG. 3C, the sacrificial, PSG, layer 38 is etched down towards the polysilicon 1 layer 24 except in the areas protected by mask 40. The process used to etch layer 38 is similar to dry SiO.sub.2 etching, e.g. using fluorine F.sub.2 or chlorine (Cl.sub.2) based gases.
Anisotropic Etching of Polysilicon 1 Layer
Referring again to FIG. 3C, the polysilicon 1 layer 24 is etched anisotropically except in the areas protected by mask 40 straight down to the tunnel oxide layer 22. The process used to etch polysilicon 1 layer 24 is a plasma etch using a chlorine (Cl.sub.2) based gas.
Isotropic Overetching of Polysilicon 1 Layer
Referring to FIG. 3D, the polysilicon 1 layer 24 is overetched isotropically (etching equally in all directions) to form an undercut 25 of the mask 40 and PSG layer 38. That is to say that an isotropic etching process is employed removing material from the exposed surfaces on the side of the layer 24 (except for the surfaces protected by mask 40.) The overetching results in an undercut 25 of the floating gate 24 beneath the PSG layer 38 and the mask 40 to shape the floating gate 24 narrower than the channel 28. This isotropic etching is performed for a limited time until the required width of the floating gate has been reached, which is a function of factors well known by those skilled in the art. The width of undercutting the polysilicon 1 layer 24 is determined by the duration of the isotropic etching process.
EXAMPLE
The plasma etch used to control the width of polysilicon 1 layer 24 to provide an undercut 25 of polysilicon width of about 0.2 .mu.m/side can be the following process parameters:
The width of mask 40 is within the range between about 0.4 .mu.m and about 1.4 .mu.m and the width of floating gate 24 is within the range between about 0.2 .mu.m and about 1 .mu.m.
N+ Source/Drain Ion Implant
Referring again to FIG. 3D, after the isotropic etching process, arsenic (As) dopant ions 36 are ion implanted through tunnel oxide layer 22 into the P-substrate 21 to form N+ S/D (source/drain) regions 26, 27. N+ regions 26, 27 lie below the surfaces of tunnel oxide 22 in the P-substrate 21 on either side of the photoresist mask 40 (as well as the remainder of PSG layer 38.) The N+ S/D regions 26, 27 define the initial boundaries of channel 28 of the device 20 which channel 28 will be narrowed further by formation of the drain size N- region as shown in FIG. 3F. The implantation of As ions 36 is performed (preferably at an angle of 0.degree. directionally with respect to vertical) with at an energy of about 50 keV. A dose of As ions 36 of 5E15cm.sup.-2 is preferred.
Photoresist Strip and PSG Etch
Referring to FIG. 3E, the next step in the process is for the photoresist mask 40 to be stripped by conventional means.
Next, as shown by FIG. 3E, the remainder of sacrificial, PSG, layer 38 on top of floating gate 24 has been stripped away in a dilute aqueous solution of hydrogen fluoride acid etchant fluid, leaving the upper surface of polysilicon floating gate 24 completely exposed. The etch rate of PSG in HF acid is much faster than that of regular SiO.sub.2 such that the PSG layer 38 can be etched away without causing significant loss of SiO.sub.2 layer 22. Alternatively, PSG layer 38 can be replaced by a silicon nitride (Si.sub.3 N.sub.4) layer. The wet nitride etching process will not attack SiO.sub.2 either.
Formation of Drain Side Mask N-
Referring to FIG. 3F a drain side photoresist mask 41 was formed over the source side including N+ region 26 of floating gate 24 with the drain side of polysilicon 1 (floating gate) layer 24 and the drain region 27 and the undoped portion 48 of substrate 21 exposed at the right end of the channel 28 of the device 20.
Implant in Drain Side Region
Referring again to FIG. 3F, N- drain side region 48 in P- substrate 21 is ion implanted with N- phosphorus (P) dopant 43. The result of the overetching of the polysilicon 1 layer 24 below the PSG layer 38 has allowed the results which are as follows:
(1) N- ion implantation for the tunneling point area with self-alignment; and
(2) formation of a split gate structure.
Region 48 is adjacent to N+ drain region 27 and is formed in the P-substrate 21 below the exposed surfaces of tunnel oxide 22.
The implantation of phosphorus (P) ions 43 is performed within the range of energy is between about 30 keV and about 100 keV with a dose between about 1E13cm.sup.-2 and about 5E14cm.sup.-2.
Preferably, the ion implantation of ions 43 is performed at an energy of about 50 keV, and a dose of phosphorus (P) ions 43 of about 1E14cm.sup.-2 is preferred.
Interpolysilicon Dielectric (ONO, Oxide or ON)
Referring to FIGS. 3G, the product of FIG. 3F is shown after the formation of a conformal, dielectric interpolysilicon thin film sandwich 44 (e. g. ONO or silicon dioxide.) In the case of use of ONO film the effective thickness is preferably about 200 .ANG., applied to the exposed surfaces of the device of FIG. 3F by means of a conventional process. A range of thickness between about 100 .ANG. and about 300 .ANG. is usually used. ONO comprises a stack of layers in the sequence O:N:O (Silicon Oxide:Silicon Nitride:Silicon Oxide.)
Control Gate Deposition and Doping
Then, referring to FIG. 3H, the product of FIG. 3G is shown after a blanket, polysilicon 2, control gate layer 50 has been deposited upon thin film (ONO) sandwich 44 to a preferred range of thicknesses between about 1,000 .ANG. and about 5,000 .ANG. by the conventional LPCVD (low pressure chemical vapor deposition) process. In this case, polycide is a substitute for polysilicon 2 in control gate layer 50.
Subsequently, a control gate layer 50 composed of polysilicon 2 is usually doped in a blanket application of dopant of a POCl3 doping process within a temperature range between about 850.degree. C. and about 950.degree. C. for between about 10 minutes and about 30 minutes.
Control Gate (Word Line) Masking and Etching
Referring to FIGS. 3I a photoresist mask 52 was formed over the polysilicon 2 layer 50. Mask 52 was patterned in the form of word lines and control gates in the conventional manner, which will be well understood by those skilled in the art. Then the polysilicon 2 layer 50 is etched accordingly using a conventional etchant and the mask 52 to pattern the polysilicon 2 layer 50.
Back end process
Conventional back end process steps follow including depositions of a glass (BPSG) layer, flow, contact formation, metallization and passivation.
Operation of the flash memory cell in accordance with this invention can be summarized as follows:
To program the cell the process is as follows: an electron is pulled out of the floating gate 24 through the area of tunnel oxide layer 22 above the N- region 48 to the N-(N+) drain 27 (because the relatively high voltage is between V.sub.D (5 V) and V.sub.G (-10 V).)
To erase the cell, the process is as follows: electrons are pulled from the drain 27, source 26 and substrate 21 through the tunnel oxide layer 22 above all three regions into the floating gate 24.
Both the program and the erase functions employ Fowler-Nordheim tunneling.
In Example 1, the program and erase operations are controlled through the thin tunnel oxide layer 22 under the floating gate 24 by Fowler-Nordheim tunneling.
Near the source side (region 26) the offset created by the same overetch of polysilicon 1 layer 24 forms the channel for the isolation transistor.
This invention is adapted for use in Flash EEPROM memory devices.
While this invention has been described in terms of the above specific embodiment(s), those skilled in the art will recognize that the invention can be practiced with modifications within the spirit and scope of the appended claims, i.e. that changes can be made in form and detail, without departing from the spirit and scope of the invention. Accordingly all such changes come within the purview of the present invention and the invention encompasses the subject matter of the claims which follow.
Claims
Having thus described the invention, what is claimed as new and desirable to be secured by Letters Patent is as follows:
1. A nonvolatile memory device including asymmetric floating gate memory transistors, the nonvolatile memory device comprising: a tunnel dielectric layer on a semiconductor substrate having a surface region doped to a first conductivity type; a floating gate on said tunnel dielectric layer, said floating gate having a first and a second edge; a drain region doped to a second conductivity type, said drain region comprising a relatively high doping level portion and a relatively low doping level portion, said relatively high doping level portion spaced from said first edge of said floating gate by a first distance and an inner edge of said relatively low doping level portion aligned with said first edge of said floating gate, said inner edge separated from said first floating gate only by said tunnel dielectric layer; a source region doped to said second conductivity type, said source region having an inner source edge, wherein said inner source edge is spaced from said second edge of said floating gate by the first distance, said inner source edge separated from said floating gate only by said tunnel dielectric layer, wherein said source region and said relatively high doping portion of said drain are doped with arsenic ions implanted with a concentration of about 5E15/cm.sup.2 at an energy of about 50 KeV, and said relatively low doping level portion of said drain comprises phosphorous ions implanted with a concentration of about 1E14/cm.sup.2 at an energy of about 50 KeV; a channel region defined in said substrate between said inner source edge and said inner edge of said relatively low doping level portion of said drain; an interpolysilicon dielectric layer over said floating gates and over said channel region; and a control gate on said interpolysilicon dielectric layer and extending over said channel region said floating gate and said relatively low doping level portion of said drain region, said control gate separated from said channel region by said interpolysilicon dielectric layer over an isolation transistor portion of said channel region between said source and said floating gate, said control gate separated from said relatively low doping level portion of said drain region by said interpolysilicon dielectric layer.
2. The nonvolatile memory of claim 1, wherein cell programming is accomplished by applying a negative voltage to said control gate and a positive voltage to said drain region so that electrons travel from said floating gate to said drain region.
3. The nonvolatile memory of claim 1, wherein cell erase is accomplished by transporting electrons from said source region, said semiconductor substrate and said drain region into said floating gate.
4. The nonvolatile memory of claim 3, wherein cell programming is accomplished by applying a negative voltage to said control gate and a positive voltage to said drain region so that electrons travel from said floating gate to said drain region.
5. The nonvolatile memory of claim 3, wherein cell programming and cell erase proceed through Fowler-Nordheim tunneling.
6. A nonvolatile memory device in accordance with claim 1 wherein said floating gate comprises a first polysilicon layer having a thickness within a range between about 500 .ANG. and about 2,000 .ANG..
7. A nonvolatile memory device in accordance with claim 6 wherein said control gate comprises a second polysilicon layer having a thickness within a range between about 1,000 .ANG. and about 5,000 .ANG..
8. A nonvolatile memory device in accordance with claim 1 wherein said floating gate has a thickness within a range between about 500 .ANG. and about 2000 .ANG..
9. A nonvolatile memory device in accordance with claim 8 wherein said control gate has a thickness within a range between about 1000 .ANG. and about 5000 .ANG..
10. A nonvolatile memory device in accordance with claim 1 wherein said nonvolatile memory device operates in accordance with the table as follows: V.sub.G V.sub.D V.sub.S V.sub.sub ______________________________________ Erase 18 0 0 0 Program -10 5 float 0 Read 5 1 0 0 ______________________________________
Patent Citations (17)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US4267558 | 1981-05-01 | Guterman | |
| US4302766 | 1981-11-01 | Guterman et al. | |
| US4868629 | 1989-09-01 | Eitan | |
| US5065201 | 1991-11-01 | Yamauchi | |
| US5153684 | 1992-10-01 | Shoji et al. | |
| US5172200 | 1992-12-01 | Muragishi et al. | |
| US5194925 | 1993-03-01 | Ajika et al. | |
| US5216268 | 1993-06-01 | Chen et al. | |
| US5274588 | 1993-12-01 | Manzur et al. | |
| US5280446 | 1994-01-01 | Ma et al. | |
| US5286665 | 1994-02-01 | Muragishi et al. | |
| US5317179 | 1994-05-01 | Chen et al. | |
| US5379253 | 1995-01-01 | Bergemont | |
| US5446298 | 1995-08-01 | Kojima | |
| US5483487 | 1996-01-01 | Sung-Mu | |
| US5504358 | 1996-04-01 | Hong | |
| US5569946 | 1996-10-01 | Hong |
Non-Patent Literature (1)
- Samachiga et al., "128L Flash EEPROM Using Double-Poly-Silicon Technology," IEEE Solid State Circuit Vol. SC-22, No. 5, Oct. 1987, pp. 676-683.