Flash memory and fabrication method and operation method for the same

1. A flash memory, comprising two memory cells of vertical channels, wherein a lightly-doped N type silicon is used as a substrate; a P+ region is provided on each of the both ends of the silicon surface, and two channel regions perpendicular to the surface are provided between the P+ regions; an N+ region shared by the two channels is provided over the channels; a tunneling oxide layer, a polysilicon floating gate, a block oxide layer and a polysilicon control gate are sequentially provided on the outer side of each channel from inside to outside; and the polysilicon floating gate and the polysilicon control gate are isolated from the P+ region by a sidewall oxide layer.

2. A method for fabricating the flash memory according to claim 1 , comprising the following steps:

(1) performing shallow trench isolation with respect to a lightly-doped N-type bulk silicon substrate to form an active region, and performing an ion implantation to the silicon surface to form an N+ layer;

(2) sequentially depositing silicon dioxide and silicon nitride on the N+ layer, and anisotropicly etching the silicon dioxide and silicon nitride to form a double-layered hard mask;

(3) anisotropicly etching the silicon under protection of the double-layered hard mask to form an N+ region;

(4) isotropicly etching the silicon until the lower end of the channel to form a channel surface under the double-layered hard mask;

(5) performing an ion implantation at a predetermined angle with respect to the silicon surface on the outer sides of the channel to form two P+ regions;

(6) thermally growing firstly a sacrificial oxide layer, and removing the sacrificial oxide layer by wet etching; then further thermally growing an oxide layer; subsequently, depositing a layer of polysilicon, heavily doping the layer of polysilicon, and performing rapid thermal annealing thereto to activate the impurities;

(7) anisotropicly etching the polysilicon formed in step 6) so that the polysilicon directly under the double-layer hard mask is remained, to form a polysilicon floating gate; then etching the oxide layer on the outer side of the floating gate to expose the silicon substrate; and

(8) depositing an oxide layer to form a block oxide layer and a sidewall oxide layer; subsequently, depositing another layer of polysilicon; implanting impurities into the another layer of polysilicon at an angle of 0 degree, and performing RTA to activate the impurities; and then etching the another layer of polysilicon to form a control gate.

3. An operation method of the flash memory according to claim 1 , wherein each of the two memory cells is operated independently, and with respect to any one of the memory cells, in the case of programming, a positive bias is applied to the N+ region and the P+ region is grounded or a negative bias is applied to the P+ region while a positive bias is applied to the control gate, so that a part of the electrons enter into the polysilicon floating gate by passing through the tunneling oxide layer; in the case of erasing, a positive bias is applied to the N+ region and the P+ region is grounded or a negative bias is applied to the P+ region while a negative bias is applied to the control gate, so that a part of the holes enter into the polysilicon floating gate by passing through the tunneling oxide layer and recombine with the electrons therein; in the case of reading, a positive bias is applied to the N+ region and the P+ region is grounded or a negative bias is applied to the P+ region while a positive bias is applied to the control gate and is controlled to read the current from the N+ region without performing a false programming; during the process of performing the programming, erasing and reading with respect to one memory cell, the control gate of another memory cell is grounded while the bias of its P+ region is the same as that of the former cell.