At a Glance
	A new cylinder package has been developed for carbon monoxide that reduces iron contamination by three orders of magnitude from 870 ppb to sub-ppb levels (0.6 ppb). This significant reduction in iron contamination has been achieved in exchange for a sub-ppb increase in aluminum, zinc and copper contamination from the package. A brass valve was found to be superior to a stainless steel DISS valve in carbon monoxide service. This is because significant levels of iron (50 ppb) and nickel (1 ppb) were found to be extracted from the DISS valve.

Carbon monoxide (CO) is currently used in the semiconductor industry to modify the rate and uniformity of dry etch processing. As with all process gases, the purity of the CO has an impact on the process yield and quality. Previously, we showed that the cylinder package used for CO in dry etch processing directly effects the level of metal contamination on the wafer.¹ Results from this work show that iron surface contamination can differ by orders of magnitude depending on the choice of cylinder and valve materials.

Typically, the exact chemical composition of metallic impurities in high-purity gases is unknown. In the case of CO, it is well established that iron and nickel carbonyl compounds can form via gas-solid reactions in high-pressure environments such as that found inside a gas cylinder.² It is believed that these reactions contribute to the metals' contamination in compressed CO since cylinder package materials commonly contain iron and nickel, and these metals are common contaminants in compressed CO.

In order to determine the level of metal contamination associated with different CO packages, several cylinders were sampled by a hydrolysis sampling method and the metals analyzed for by flame emission spectroscopy, ICP-OES, GFAA and ICP-MS.

Experimental

Two cylinders for each of three CO cylinder packages were compared for metal contamination. One, a representative of commonly used packages in the industry, was a 4130 carbon steel cylinder outfitted with a brass packed valve. The second was a 6061 aluminum cylinder outfitted with a 316 stainless steel diameter index safety system (DISS) diaphragm valve. The third was a new Ultraline package developed by Matheson Gas Products that minimizes the use of steel and nickel wetted parts, which included a 6061 aluminum cylinder outfitted with a valve identical to the one employed on the carbon steel package. The elemental composition of the components used in each package is listed in Table 1.

1. This hydrolysis sampling apparatus was used to test cylinders and generate the data shown in Tables 1 and 2.

Test cylinders were sampled using the hydrolysis sampling apparatus illustrated in Figure 1. Cylinders were connected to the apparatus under a purge of 0.01 µm filtered nitrogen. A Kel-F gasket was used for cylinders with a DISS valve. The cylinder was opened, and the initial 50 g of gas was discarded. A total of 3.5 kg of CO was then sampled through 18 megohm water at 1 liter/min. Prior to actually performing the sampling, a blank was collected. The 18 megohm water was sparged using 0.01 µm filtered nitrogen, and the water was allowed to sit in the bubbler for about 8 hrs. The purpose of this blank is to account for any metals that may be leached from the bubbler or transmitted through the apparatus. Although it has not been demonstrated that this sampling method is quantitative for carbonyl sampling, by assuming that the amount of metals captured is proportional to their concentration, the method is useful for comparison purposes.

Analysis was performed by flame emission spectroscopy for the analysis of potassium and sodium, ICP-OES for calcium, GFAA for iron and ICP-MS for all other analyzed elements. The bubbler and blank solutions were acidified to ensure dissolution of the metals by minimizing their adsorption on the bubbler walls. The samples were then analyzed directly along with standards containing each analyzed element.

Results

In this report, the discussion of the metal impurities is limited to iron, nickel, aluminum, zinc and copper, because these metals are the main package constituents and the ones that appeared to measurably vary in the three packages (Table 2).

For iron contamination, it is seen from the data in Table 2 that the use of an aluminum cylinder and a brass valve in the Ultraline package reduced the levels by an average factor of 1400 over the carbon steel package and 78 over the Ultraline DISS package. The reduced iron contamination in the gas correlates well with the iron content in the Ultraline package materials listed in Table 1. The carbon steel package contained the largest amount of iron, the Ultraline DISS package contained less iron and the Ultraline package contained the least iron.

For nickel contamination, approximately equal levels were found in the Ultraline and the carbon steel package; these contained no stainless steel. However, six times the nickel levels was found in the CO from the Ultraline DISS package, in comparison to the other packages. This is attributed to the stainless steel valve (shown to be 12% nickel in Table 1) employed in the package.

While allowing significant reductions of iron and nickel contamination, the data from the Ultraline package suggest a small enhancement in the aluminum (~0.4 ppb). However, this increase is not supported by the aluminum levels found associated with the Ultraline DISS package, which employs the same aluminum cylinder. The average aluminum levels attributable to the carbon steel and Ultraline DISS package were 0.03 ppb.

The zinc levels for the Ultraline package were higher by a factor of two in comparison to the Ultraline DISS package and by seven in comparison to the carbon steel package. In the packages, the zinc arises as a constituent of brass. However, no correlation was found between the brass content of the packages and the zinc levels found as a CO contaminant.

The highest and lowest copper levels found for the three packages were 0.06 and 0.01 ppb, respectively. The data do not suggest any correlation of copper contamination and package materials.

The main mechanism for conveying iron and nickel into compressed CO is believed to be formation of the metal carbonyls. However, no direct reactions of CO with aluminum or copper were found in the literature. Zinc carbonyls were not found to even be referenced in the literature. It is probable that the mechanism for introducing the metals aluminum, zinc and copper into compressed CO is other than by carbonyl formation.

Conclusion

A new cylinder package has been developed for CO that reduces iron contamination by three orders of magnitude from 870 ppb to sub-ppb levels (0.6 ppb). This significant reduction in iron contamination has been achieved in exchange for a sub-ppb increase in aluminum, zinc and copper contamination from the package. A brass valve was found to be superior to a stainless steel DISS valve in CO service. This is because significant iron (50 ppb) and nickel (1 ppb) were found to be extracted from the DISS valve.

References

1. G. Cooper, P. Andersen, C. Radens, V. Houlding, Semiconductor International, pp. 301-302, July 1997.

2. D. Shriver, P. Atkins, C. Langford, Inorganic Chemistry, W.H. Freeman and Co., 1990.

3. American Society for Metals, Metals Handbook, 1985.

Peter C. Andersen has a bachelor's degree in chemistry from the University of California at Santa Barbara and is currently a research chemist in the analytical R&D division of Matheson Gas Products.

Gerald Cooper has an master's degree in theoretical physical chemistry and is the lab manager of the Advanced Technology Center at Matheson Gas Products in Longmont, Colo.

Virginia H. Houlding is the technical director of Gas Operations at Matheson Gas Products and has a doctorate in physical chemistry from the University of Southern California.

The authors can be reached at:
Phone (303) 678-0700
Fax (303) 442-0711

Table 1. Typical Elemental Composition of Cylinder Package Materials3
Composition of wetted surfaces (% wt.)
	Al	C	Cr	Cu	Fe	Mg	Mn	Mo	Ni	P	Pb	S	Si	Zn
Carbon steel package
4130 carbon steel cylinder		0.3	0.95		98		0.5	0.2		0.04		0.04	0.23
C37700 brass valve				59							2			39
Ultraline DISS package
6061 aluminum cylinder	98		0.2			1	0.28						0.6
316L SS valve		0.03	17		65		2	2.5	12	0.5		0.03	1
Ultraline package
6061 aluminum cylinder	98		0.2			1	0.28						0.6
C37700 brass valve				59							2			39

Table 2. Elemental Impurities in Various CO Packages
Carbon steel package				Ultraline DISS package			Ultraline package
	Cylinder 1	Cylinder 2	Average	Cylinder 1	Cylinder 2	Average	Cylinder 1	Cylinder 2	Average
Element	(ppbw)	(ppbw)	(ppbw)	(ppbw)	(ppbw)	(ppbw)	(ppbw)	(ppbw)	(ppbw)
Al	0.037	0.029	0.033	0.044	0.019	0.032	0.123	0.610	0.367
B	0.076	4.25	2.16	0.044	0.05	0.047	0.678	5.96	3.32
Ca	0.712	0.766	0.739	0.012	0.013	0.013	0.015	0.265	0.140
Cd	0.023	0.023	0.023	0.013	0.015	0.014	0.017	0.021	0.019
Co	0.006	0.014	0.010	0.005	0.007	0.006	0.006	0.007	0.007
Cr	0.012	0.023	0.018	0.007	0.011	0.009	0.013	0.008	0.010
Cu	0.019	0.019	0.019	0.011	0.013	0.012	0.071	0.042	0.057
Fe	265	1472	869	50.2	47.4	48.8	0.840	0.403	0.622
Ga	0.004	0.005	0.005	0.003	0.003	0.003	0.009	0.011	0.010
Ge	0.015	0.015	0.015	0.009	0.010	0.010	0.011	0.027	0.019
In	0.006	0.006	0.006	0.003	0.004	0.004	0.008	0.010	0.009
K	0.160	0.193	0.177	0.014	0.016	0.015	0.063	0.686	0.375
Li	0.189	0.190	0.190	0.110	0.124	0.117	0.014	0.018	0.016
Mg	0.015	0.015	0.015	0.018	0.010	0.014	0.012	0.011	0.012
Mn	0.009	0.009	0.009	0.005	0.006	0.006	0.007	0.008	0.008
Mo	0.017	0.017	0.017	0.010	0.011	0.011	0.011	0.014	0.013
Na	0.263	0.881	0.572	0.016	0.018	0.017	3.45	7.14	5.30
Ni	0.077	0.147	0.112	0.808	1.17	0.989	0.054	0.316	0.185
Pb	0.019	0.019	0.019	0.011	0.012	0.012	0.016	0.020	0.018
Pt	0.018	0.018	0.018	0.010	0.012	0.011	0.008	0.010	0.090
Sb	0.011	0.011	0.011	0.007	0.008	0.008	0.022	0.027	0.025
Sn	0.012	0.012	0.012	0.007	0.008	0.008	0.012	0.014	0.013
Zn	0.025	0.015	0.020	0.109	0.010	0.060	0.058	0.233	0.146
Total metal impurities			873			50.2			10.8

Talkback

We would love your feedback!

Post a comment

» VIEW ALL TALKBACK THREADS

Related Content

• Topics
• Author

Related Content

 

By This Author

There are no other articles written by this author.

SPONSORED LINKS