CN102930101B - Computing method for meal gate surface appearance - Google Patents

Abstract

The invention discloses a computing method for metal gate surface appearance and belongs to the technical field of integrated circuit manufacturing and electronic design automation. The method includes acquiring map information and measure parameter information of the metal gate surface appearance, then establishing a parameter prediction model of the metal gate surface appearance, and performing a prediction for the surface appearance on the back of a designed metal gate structure corrugated metal pipe (CMP) in a chip design process through the prediction model to obtain dish-shaped appearance of the metal gate and erosion appearance of dielectrics among metal gates. Accordingly, area which is probable to generate problems after the CMP processing step in the design can be found out, the area is provided for designers to improve, the corresponding problems occurring in the manufacturing process are avoided, and thereby the yield of products is improved.

Description

A kind of computing method of meal gate surface appearance

Technical field

The present invention relates to IC manufacturing and EDA Technique field, particularly a kind of computing method of meal gate surface appearance.

Background technology

Along with integrated circuit fabrication process node enters 32nm and following, traditional CMOS(ComplementaryMetal Oxide Semiconductor) device size reduce Technology Ways face technical bottleneck: on the one hand, ever-reduced gate oxide thickness causes increasing leakage current, add device power consumption, reduce device performance and reliability; On the other hand, gate oxide thickness also close to the limit, only the gate oxide of which floor silicon atom thickness remaining face continue thinning just without silicon can condition.In order to ensure that performance is improved while device dimensions shrink, guarantee that electricity leakage power dissipation within the acceptable range, the new device structure of high k/ metal gate (high k/metalgate) arises at the historic moment simultaneously.

In high k/ metal gate process, usually need twice chemically mechanical polishing (chemical mechanicalpolishing, hereinafter referred to as CMP) processing step.Be ground floor dielectric (ILD0) CMP together, for removing the oxide above grid, for redundancy polysilicon gate etching step is below prepared; Another road is metal gate CMP, carries out, remove unnecessary metal after the deposits such as electrode metal (as Al), simultaneously for follow-up processing step provides a smooth surface.The result of metal gate CMP step has great impact to subsequent technique, excessive polishing can cause damaging source and drain areas during contact etching, and directional polish can cause the short circuit between device, integrated circuit (IC) chip both can be made to work, affect chip yield.

Meanwhile, in the metal gate CMP step of high k/ metal gate device structure, different grid length, grid width and the spacing between grid and grid all can affect the surface topography after CMP, produce the defects such as metal dish and erosion medium resistance simultaneously.In addition, in the integrated circuit (IC) design of complexity, although limited strict design rule, but the geometric parameter of grid still can change in certain scope, thus the surface topography of silicon chip after affecting CMP in the fabrication process, make there is different height at the diverse location of chip, as photoetching, etching etc., harmful effect is caused to subsequent process steps thus affects chip yield.

Summary of the invention

Embodiments provide a kind of computing method of meal gate surface appearance, can predict the surface topography after designed metal-gate structures CMP in chip design process, and for correct may be shifted to an earlier date by produced problem, effectively can improve chip yield.

Embodiments provide a kind of computing method of meal gate surface appearance, comprising: the structural parameters information obtaining described meal gate surface appearance; According to described structural parameters information, obtain the layout information of described meal gate surface appearance; Obtain the measurement parameter information of described meal gate surface appearance; According to described layout information and described measurement parameter information, set up meal gate surface appearance parametric prediction model; The surface topography parameters of described meal gate surface appearance after cmp is calculated according to described forecast model.

Further, described method also comprises: the structural parameters information of the described meal gate surface appearance of described acquisition, comprising: obtain the grid length of described metal gate, the various combination between grid width and grid spacing.

Further, described method also comprises: the measurement parameter information of the described meal gate surface appearance of described acquisition, comprising: obtain the dish-shaped information of described metal gate, erosion medium resistance information and/or dielectric thickness information.

Further, described method also comprises: describedly set up meal gate surface appearance parametric prediction model, comprising: set up the pressure-plotting model of meal gate surface appearance, material removing rate model and/or model of fit.

Further, described method also comprises: the described model of fit setting up metal gate surface appearance, comprising:

According to described layout information and described measurement parameter information, obtain matching information; Wherein, described matching information is adopt the analogue value and the minimum factor value of actual measured value root mean square.

Further, described method also comprises: the measurement parameter information of the described meal gate surface appearance of described acquisition, comprising:

Measured by afm scan and/or scanning electron microscope section measuring method obtain described measurement parameter information.

The embodiment of the present invention is by obtaining layout information and the measurement parameter information of meal gate surface appearance, and then set up meal gate surface appearance parametric prediction model, can be predicted the surface topography after designed metal-gate structures CMP in chip design process by forecast model, and for shifting to an earlier date correct by produced problem, thus chip yield can be improve.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the computing method of a kind of meal gate surface appearance in the embodiment of the present invention;

Fig. 2 is a kind of meal gate surface appearance structural parameters schematic diagram in the embodiment of the present invention.

Embodiment

The embodiment of the present invention is by obtaining layout information and the measurement parameter information of meal gate surface appearance, and then set up meal gate surface appearance parametric prediction model, can be predicted the surface topography after designed metal-gate structures CMP in chip design process by forecast model, obtain the erosion topography of medium between the dish-shaped pattern of metal gate and metal gate grid, thus the region that may go wrong after CMP step in design can be found out, the regions such as value as dish-shaped in metal gate is excessive, be supplied to deviser to be improved, thus avoid and occur corresponding problem in the fabrication process, thus improve the yield of product.

As shown in Figure 1, embodiments provide a kind of computing method of meal gate surface appearance, described method specifically comprises:

Step 110: the structural parameters information obtaining described meal gate surface appearance,

Step 120: according to described structural parameters information, obtain the layout information of described meal gate surface appearance;

Step 130: the measurement parameter information obtaining described meal gate surface appearance;

Step 140: according to described layout information and described measurement parameter information, set up meal gate surface appearance parametric prediction model.

Wherein, the structural parameters information that step 110 obtains described meal gate surface appearance is specially: obtain concrete structural parameters information according to meal gate surface appearance, specifically comprise: metal gate grid are long, the structural parameters information such as metal gate grid width and metal distance between gates, specifically, described structural parameters information can be the combination of metal gate grid length and grid width, also can be the combination of metal gate grid length and grid spacing, it can also be the combination of metal gate grid width and grid spacing, further, described structural parameters information can also be that metal gate grid are long, grid width, the parameter information that these structural parameters of grid spacing are derived, such as density, tolerance etc.

Wherein, step 120, according to described structural parameters information, obtains the layout information of described meal gate surface appearance.Wherein, the rule obtaining the layout information of described meal gate surface appearance can be for: the scope of design of layout information is contained allowed whole region by (1); (2) in whole region, select appropriate point, wherein appropriate point can adopt logical point selective rule for grid width, grid length and grid spacing.For grid width, if the allowed band of grid width is 0.04um ~ 5um in design rule, then 0.04um can be chosen, 0.06um, 0.08um, 0.1um, 0.15um, 0.2um, 0.35um, 0.5um, 0.8um, 1.0um, 1,5um, 2um, 2,5um, 3um, 3,5um, 4um, 5um etc. series of points.(3) ensure that layout information is comparatively even in gamut, that is, the selection of layout information will cover fine line region and wide line region; Or long line region and short-term region; Or wide-space region and thin space region.(4) selected point quantity and the expansion of resolution chart quantity, the cost of flow is balanced and relation between time needed for measuring, namely compromised between development time and model accuracy and chosen, and then selected the array mode of metal gate grid length, grid width, grid spacing point.In sum, according to described structural parameters information, the layout information obtaining described meal gate surface appearance needs to consider above-mentioned rule, and certain the application does not limit and adopts above-mentioned rule simultaneously, can selectivity choose according to actual needs.Lift the acquisition that an example illustrates layout information below, table 1 lists each parameter combinations example in resolution chart, and Fig. 2 lists the layout information of this resolution chart., it should be noted that and do not represent the situation that the combination shown in table 1 and Fig. 2 is only possible and only have these situations, also may comprise the layout information of other parameter, parameter combinations or resolution chart.

Table 1: each parameter combinations example in resolution chart

	Grid width (um)	Grid long (um)	Grid spacing (um)
				#1	0.03	0.05	0.03
#2	0.08	0.20	0.05
				#3	0.20	0.50	0.20
#4	1.00	2.00	0.80 … … … …

The measurement parameter information that the step 130 of the embodiment of the present invention obtains described meal gate surface appearance is carried out after test layout drawing, generally speaking, flow technique is carried out after test domain is completed, in flow after corresponding steps completes, survey instrument can be adopted to carry out measurements acquisition measurement parameter information to the surface topography of domain.Specifically, measurement parameter information comprises metal gate dish information, erosion medium resistance information and/or dielectric thickness information etc., measuring method is mainly through atomic force microscope (Atomic ForceMicroscope, hereinafter referred to as AFM) scanning survey and/or scanning electron microscope (scanning electronmicroscope, hereinafter referred to as SEM) measuring method of cutting into slices obtain measurement parameter information.Wherein, AFM scanning can obtain the relative pattern of chip surface, and SEM can obtain the absolute thickness of chip appointed area; Also directly can be obtained the erosion medium resistance amount in the metal gate dish amount of each structure, each structure by AFM, obtain the dielectric thickness without graphics field by SEM; Meanwhile, SEM all can also be adopted to measure metal gate thickness in each structure and dielectric thickness and the dielectric thickness without graphics field, then calculate amount in need according to the definition of metal dish and erosion medium resistance; In addition, again can by measure other height or thickness value in order to calculate.

After acquisition layout information and measurement parameter information, enter step 140 according to described layout information and described measurement parameter information, set up meal gate surface appearance parametric prediction model.Wherein, the foundation of forecast model is based on certain physical principle, and the parametric prediction model adopted comprises pressure-plotting model, material removing rate model and/or model of fit.Specifically, according to described layout information and described measurement parameter information, the flow process setting up meal gate surface appearance parametric prediction model is as follows:

Step 1401: based on the physical chemical mechanism of CMP, determines adopted forecast model formula.Forecast model comprises pressure-plotting model, material removing rate model and/or model of fit.Further, pressure-plotting model can be the bench height model (also can be described as step height model) of pressure distribution, and concrete formula is:

$P^{\mathrm{up}}=\frac{P}{\mathrm{\ρ}}$

P ^down＝0，for h≥hc

$P^{\mathrm{up}}=P+\frac{E}{D}(1-\mathrm{\ρ})h$

$P^{\mathrm{down}}=P-\frac{E}{D}\mathrm{\ρh},$ for h≤hc

Wherein, P ^upthe pressure of elevated regions; P ^downthe pressure of sunk area; h _cbe height threshold, the difference of elevated regions and sunk area height is greater than h _ctime sunk area be not stressed; ρ is the ratio of elevated regions in whole region, or claims density; E is elastic modulus; D is the height of grinding pad rough peak when not stressing.

Material removal model can be Preston model, and concrete formula is:

RR=K _pPV；

Wherein RR is material removing rate, K _pbe that Preston constant is relevant to technique, P is pressure, and V is relative velocity.

Chip surface morphology is characterized according to forecast model formula.According to the forecast model formula adopted in previous step, the description surface pattern variablees such as metal dish and erosion medium resistance are expressed as the function of domain structure and other technique correlation factors in this step.As adopted pressure distribution formula above, the concrete formula of the metal obtained dish and erosion medium resistance is:

Metal dish amount D _mfollowing calculating formula can be adopted to calculate:

D _M=D _ss(1-e ^-t/τ)

Erosion medium resistance amount E _oXfollowing calculating formula can be adopted to calculate:

E _OX=Y ₁t+Y ₂D _ss(e ^-t/τ-1)

Wherein,

$Y_1=\frac{r_M{r}_{\mathrm{OX}}}{r_M(1-\mathrm{\ρ})+r_{\mathrm{OX}}\mathrm{\ρ}},$

$Y_2=\frac{r_{\mathrm{OX}}\mathrm{\ρ}}{r_M(1-\mathrm{\ρ})+r_{\mathrm{OX}}\mathrm{\ρ}},$

$\mathrm{\τ}=\frac{d_{\max }(1-\mathrm{\ρ})}{r_M(1-\mathrm{\ρ})+r_{\mathrm{OX}}\mathrm{\ρ}},$

$D_{\mathrm{SS}}=\frac{d_{\max }(r_M-r_{\mathrm{OX}})(1-\mathrm{\ρ})}{r_M(1-\mathrm{\ρ})+r_{\mathrm{OX}}\mathrm{\ρ}},$

d _max=A×(W _eff) ^α×(S _eff) ^β

Wherein, t is the CMP process processing time, r _mfor metal gate metal removal rate, r _oxfor dielectric layer clearance, ρ is metal gate density, W _efffor equivalent metal grid width, S _efffor equivalent metal grid spacing, d _maxfor the maximum dish amount of metal gate metal, A, α, β are fitting parameter.

Wherein,

$W_{\mathrm{eff}}=\frac{\underset{i}{\mathrm{\Σ}}{W}_i{p}_i}{\underset{i}{\mathrm{\Σ}}{W}_i}$

$S_{\mathrm{eff}}=\frac{W_{\mathrm{eff}}(1-\mathrm{\ρ})}{\mathrm{\ρ}}$

W _ifor metal grid width, p _ifor weight factor.

Step 1402: the actual metal dish that the morphology characterization formula in integrating step 1401 and measurement obtain and erosion medium resistance value, the fitting parameter in formula is obtained by approximating method, wherein matching information for adopting the analogue value and the minimum factor value of actual measured value root mean square, namely can obtain the fitting parameters such as A, α, β by approximating method.

After establishing meal gate surface appearance parametric prediction model, enter step 150 and calculate the surface topography parameters of described meal gate surface appearance after cmp according to described forecast model.Specifically, after the model obtaining description surface pattern and domain structure relation under special process, as determined r in model of fit _m, r _oxetc. technique related coefficient, for specific domain structure, only need input as layout informations such as grid length, grid width and grid spacing, as the density metal ρ in model of fit, equivalent metal grid width W _eff, equivalent metal grid interval S _effthe surface topography after specific domain structure CMP is obtained as information such as metal gate dish and erosion medium resistance Deng namely measurable.

Utilize the forecast model that the embodiment of the present invention is set up, can predict the surface topography after designed metal-gate structures CMP in chip design process, obtain the information such as corresponding metal gate dish, erosion medium resistance, thus the region that may go wrong is found in the design phase, correct in advance, avoid corresponding problem to occur in the fabrication process, thus improve product yield, accelerate time to market (TTM) and reduce cost.And a modeling process only being needed for identical technique, can to predict the different designs based on this technique and without the need to repeating modeling.

Above-described concrete implementation step; explanation has to a certain degree been carried out to object of the present invention, technical scheme and beneficial effect; be understood that; the above method is not limited in the present invention; within the spirit and principles in the present invention all; any amendment of making, equivalent replacement or improvement etc., all should be included within protection scope of the present invention.

Claims (2)

1. computing method for meal gate surface appearance, is characterized in that:

Obtain the structural parameters information of described meal gate surface appearance;

According to described structural parameters information, obtain the layout information of described meal gate surface appearance;

Obtain the measurement parameter information of described meal gate surface appearance;

According to described layout information and described measurement parameter information, set up meal gate surface appearance parametric prediction model;

The surface topography parameters of described meal gate surface appearance after cmp is calculated according to described forecast model;

Wherein, the structural parameters information of the described meal gate surface appearance of described acquisition, comprising: obtain the grid length of described metal gate, the various combination between grid width and grid spacing;

Wherein, the measurement parameter information of the described meal gate surface appearance of described acquisition, comprising: obtain the dish-shaped information of described metal gate, erosion medium resistance information and/or dielectric thickness information;

Wherein, describedly set up meal gate surface appearance parametric prediction model, comprising: set up the pressure-plotting model of meal gate surface appearance, material removing rate model and/or model of fit;

Wherein, the described model of fit setting up metal gate surface appearance, comprising:

According to described layout information and described measurement parameter information, obtain matching information; Wherein, described matching information is adopt the analogue value and the minimum factor value of actual measured value root mean square.

2. computing method as claimed in claim 1, it is characterized in that, the measurement parameter information of the described meal gate surface appearance of described acquisition, comprising:

Measured by afm scan and/or scanning electron microscope section measuring method obtain described measurement parameter information.