METHOD OF PLASMA VAPOUR DEPOSITION

Abstract

A method induces plasma vapour deposition of metal into a recess in a workpiece. The method achieves re-sputtering of the metal at the base of the recess with a sputter gas by utilising a mixture of Ar and He and/or Ne as the sputter gas with a ratio of He and/or Ne:Ar of at least about 10:1.

Description

BACKGROUND

This invention relates to a method of plasma vapour deposition of metal into a recess in a workpiece in a manner to achieve re-sputtering of the metal at the base of the recess onto the sidewalls.

It is known to use ionised metal sputtering techniques involving a high powered unbalanced magnetron discharge source. The metal ions are attracted to the base of a recess formed in a workpiece, such as a semi-conductor wafer, using a DC bias induced by applying RF power to the wafer pedestal. This gives an improved bottom and sidewall coverage in the recess. Further improvement is known to be achievable by re-sputtering the metal already deposited in the base onto the lower and mid parts of the recess sidewall. To achieve re-sputtering DC bias voltages in the range of 50 to 500 volts are required. The higher the DC bias the greater re-sputtering and there is an improvement in the sidewall coverage.

Even so the current results are not satisfactory for all purposes.

SUMMARY

From one aspect the invention consists in a method of plasma vapour deposition in metal into a recess in a work piece in a manner to achieve re-sputtering of the metal at the base of the recess onto the side walls by sputtering a metal target with a sputter gas characterised in that the sputter gas is a mixture of Ar and He when the ratio He:Ar is at least about 10:1.

Preferably the ratio is about 20:1.

The flow rate of Ar<10 sccm and the flow rate He>100 sccm. Thus for example the He flow rate may be about 140 sccm and the Ar may be about 7 sccm.

In any of the above cases the metal may be copper.

In a further embodiment the method may be characterised in that the mix of He and Ar is such that the target current density is at least about 0.035 A/cm² and preferably about 0.037 A/cm².

Although the invention has been defined above it is to be understood it includes any inventive combination of the features set out above or in the following description.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention may be performed in various ways and specific embodiments will now be described by way of example with reference to the accompanying drawings in which:

FIG. 1 is a cut away view of a DC magnetron ionised deposition sputtering system;

FIG. 2 is a table showing a secondary electron emission as a function of bombarding energy and gas types for ions incident on Mo and W targets. This figure is taken from Glow Discharge Processes, Chapmen, Wiley and Sons 1980;

FIG. 3 illustrates the variation in target current and voltage with respective flow rates by adding He to the Ar sputter process;

FIG. 4 shows the sidewall coverage in a 5:1 trench structure for a standard Ar process;

FIG. 5 shows the equivalent view to FIG. 4 for an optimised He/Ar process;

FIG. 6 is an equivalent view for an He/Ar process with a reduced He flow;

FIG. 6( a), (b), (c) show the effects of He/Ar flow on deposition rate (tooling factor); stress and within wafer non uniformity respectively; and

FIG. 7 is an equivalent view to FIG. 5 with another He recipe.

DETAILED DESCRIPTION

In an unbalanced DC Magnetron Ionised sputtering deposition system, metal and gas ionisation is achieved through the collision of metal and gas atoms with secondary electrons produced at the target surface. The secondary electrons are accelerated by the negative bias supply to the target surface and achieve high energies, typically 200 to 1000 eV. These high energy secondary electrons collide with metal and gas atoms or molecules and produce ions. Accordingly if the amount of secondary electrons emitted from the target could be increased then there would be an increase in the ionisation fraction that is the ratio of ions to neutrals from any species.

As can be seen from FIG. 2, the secondary electron characteristics for various sputter gasses vary depending on the sputter gasses used. Thus He and Ne have higher levels of secondary electron emission as compared to heavier gasses such as Ar and Kr. The Applicants therefore appreciated that sputtering with He or Ne instead of Ar was likely to result in more secondary electrons emitted from the target and hence ionisation fraction.

This effect has been investigated by the Applicants by progressively introducing increased levels of helium flow and reducing the argon flow using a copper target and it would be seen that as the He is increased and the Ar is reduced the current increases due to formation of secondary electrons at the target surface. The presence of a low level of argon was discovered to be necessary to maintain film density quality.

Experiments

The Applicants then carried out an experiment using the apparatus of FIG. 1 with the process condition shown in Table 1 below.

	TABLE 1
	Parameter	Std process	He/Ar process
	DC Power (kW)	16	16
	Upper coil (A)	15	15
	Lower coil (A)	7	10.5
	Ar (sccm)	100	7
	He (sccm)	0	140
	Pressure (mT)	5.97	5.02
	RF bias power (W)	335	500
	Platen DC bias (V)	245	275

This compared a standard argon only process run on the apparatus of FIG. 1 and an He/Ar process where the Ar flow was 7 sccm whereas the He flow was 140 sccm. The resultant coverage for the first process is shown in FIG. 4, whilst for the second process the coverage is illustrated in FIG. 5. The resultant step coverage is summarised in Table 2 below.

It will be seen that the He/Ar process shows a 5% absolute, 30% proportional, increase in sidewall coverage over the Ar only process value.

It can be seen from Table 1 that the platen DC bias is increased by only 30V (12%) with the He/Ar process, despite the RF power applied to the platen having been increased by 165 W (50%) compared with the Ar only (standard) process. This shows that the He/Ar process has a higher level of ionisation, since the positive ions in the plasma will tend to reduce the negative DC bias achieved at the platen.

In FIG. 6 the same process was run but this time with a 75 He sccm flow and the results are summarised in Table 3 below.

TABLE 2
140 He/7 Ar Step Coverage using SEM
	Measurement	Thickness	Coverage
	Position	(nm)	(%)
	Field	360	100
	Sidewall - Top	141	39
	Sidewall - Middle	75	21
	Sidewall - Bottom	75	21
	Base	120	33

TABLE 3
75 He/7 Ar Step Coverage using SEM
	Measurement	Thickness	Coverage
	Position	(nm)	(%)
	Field	360	100
	Sidewall - Top	56	16
	Sidewall - Middle	35	10
	Sidewall - Bottom	58	16
	Base	136	38

From this it will be seen that reducing the He flow results in reduced sidewall coverage both as compared to the 140 He/7Ar process and the Ar only process. This is because in this experiment the secondary electrons produced in the plasma were reduced due to the lower He flow.

Experiments have shown that <10 Sccm Ar is required to maintain the plasma and allow the sputtering of the Cu target to occur. Conversly He flows >100 Sccm are required to help maintain the plasma and provide additional secondary electrons to enhance the sputtering effect at the bottom of the structure and so improve sidewall coverage.

In further experiments the deposition rate stress and within wafer non uniformity were monitored with varying flows and the results are shown respectively in FIGS. 6A to C. It will be seen that although the deposition rate drops with the addition of He the 7Ar/140He process is still high enough from production purposes. This combination has a particularly good stress value and an exceptable uniformity. It will be appreciated that reduced film stress helps to prevent delamination of the Cu film from the underlying materials.

FIG. 7 illustrates a reduced helium flow with an increased argon flow and again the step coverage is reduced as is shown in Table 4 below.

TABLE 2
140 He/7 Ar Step Coverage using SEM
	Measurement	Thickness	Coverage
	Position	(nm)	(%)
	Field	360	100
	Sidewall - Top	141	39
	Sidewall - Middle	75	21
	Sidewall - Bottom	75	21
	Base	120	33

TABLE 4
75 He/25 Ar Step Coverage using SEM
	Measurement	Thickness	Coverage
	Position	(nm)	(%)
	Field	372	100
	Sidewall - Top	63	17
	Sidewall - Middle	59	16
	Sidewall - Bottom	56	15
	Base	108	30

Accordingly in summary it is seen that considerable improvements can be achieved by having helium flow rates above about 100 sccm and argon flow rates below about 10 sccm. The 7Ar/140He resulted in a particularly preferred process for the reasons indicated above.

It may in fact be more generally applicable to speak about the partial pressures of the gases as these should relatively remain relative constant in terms of the performance achieved, whilst the actual flow rates may vary from chamber to chamber. Table 5 below sets out the equivalent partial pressures for the experimental flow rates.

TABLE 5
Ar	He	TOTAL	Ar Partial	He Partial
Flow	Flow	Pressure	Pressure	Pressure
(sccm)	(sccm)	(mT)	(mT)	(mT)
100	0	6.0	6.0	0.0
7	140	6.2	0.3	5.9
7	75	3.5	0.3	3.2
25	75	4.7	1.2	3.5

The metal used in the experiments was copper. The invention process would be similar for Titanium, Tantalum, Gold or Ruthenium.

Claims

1. A method of plasma vapour deposition of metal into a recess in a workpiece in a manner to achieve re-sputtering of the metal at the base of the recess on to the sidewalls by sputtering a metal target with a sputter gas characterised in that the sputter gas is a mixture of Ar and He and/or Ne wherein the ratio He and/or Ne:Ar is at least about 10:1.

2. The method as claimed in claim 1, wherein the ratio is about 20:1.

3. The method as claimed in claim 1, wherein the flow rate of Ar<10 sccm, and the flow rate of He and/or Ne>100 sccm.

4. The method as claimed in claim 3, wherein the He flow rate is about 140 sccm and the Ar flow rate is about 7 sccm.

5. The method as claimed in claim 1, wherein the metal is selected from Copper, Titanium, Tantalum, Gold or Ruthenium.

6. The method as claimed in claim 1, wherein the mix of He and/or Ne and Ar is such that the target current density is 0.035 A/cm2.