The demand for the delivery of ultra-high purity (UHP) gas to the process tool has resulted in the implementation of ultraclean gas distribution systems within the semiconductor industry. The gas distribution systems are designed and constructed to transport the UHP gas to the point-of-use (POU) without any compromise in the gas purity. The selection of components for use within the gas distribution system plays a pivotal role in the delivery of the gas to the process tool at the appropriate purity level. The components typically encountered within a distribution system include valves, regulators, mass flow controllers and filters. The materials of construction of the latter components are chosen so as to minimize any possible sources of microcontamination. The POU filter plays a critical role in the delivery of UHP gas, as the filter is typically the last component in contact with the process gas prior to delivery to the wafer surface.

In order to meet the requirements of the gas delivery system, the POU filter must provide a means of removal of any particulate contamination added to the gas stream during transport to the process tool. In addition, the POU filter must not outgas any chemical contaminant into the gas stream.

The tendency of organic materials to outgas over a prolonged period of time has restricted the use of polymeric filters in a number of applications. Stainless steel filters were developed to allow POU filtration without any measurable outgassing.

Pall stainless steel POU filters are suitable for high temperature applications, such as, diffusion, CVD and EPI. The all stainless steel construction also allows the filter to be baked at elevated temperatures and allows for rapid desorption of atmospheric contaminants.

While the above factors influenced the development of all stainless steel filters, the use of polymeric filters continues to be widespread. The selection of a POU filter for a given application will be strongly influenced by compatibility and temperature considerations. In cases where polymeric and all-stainless filters are acceptable, the filter performance will be influenced by materials selection, filter design, manufacturing environment and post assembly cleaning.

The following is a comparison of the performance of the Pall 4400 Ultramet-L^® Gaskleen^® Filter Assembly (all stainless steel) and Gaskleen^® IV Series Assembly (all fluoropolymer filter). The following results outline the merits of both technologies.

Particulate Cleanliness & Efficiency

The test setup employed in the cleanliness testing of the gas filter assemblies is described in Figure I. A -40 degree F dew point air source was used to eliminate any variations that may be caused by humidity effects. The air source was passed through the coalescing element to remove any liquid droplets that may be present in the gas stream. All piping downstream of the coalescer was composed of electropolished 316L stainless steel. Where possible, VCR fittings were used.

FIGURE I: Cleanliness and Efficiency Test Setup

The concentration of particles downstream of the test filter was monitored by a Condensation Nucleus Counter (CNC, Model 3025 from TSI, Minneapolis, MN.) sampling at a flow rate of 0.01 SCFM for particles larger than 0.003µm in size. An isokinetic sampler was employed to properly sample the particles present in the effluent gas stream.

Initially the system’s background cleanliness was determined at an inlet pressure of 30 psi - for 3 intervals of five minutes steady flow and ten minutes pulsed flow. The pulsed flow was achieved by use of a solenoid valve upstream of the test filter and resulted in 30 pulses per minute at 30 psi. The standard cleanliness test was conducted at a 30 psi inlet pressure and 50 slpm air flow for the Ultramet-L 4400 filter assembly and at a 60 psi inlet pressure and 110 slpm air flow for the Gaskleen IV filter assembly. The particulate cleanliness of the POU filter assemblies are shown in Table I.

TABLE I
The particle counts observed during particulate cleanliness and efficiency testing.

	Particle Contribution counts/ft3	Downstream Counts counts/ft3	Efficiency
Ultramet-L® Gaskleen® 4400 Series Assembly GLFF4400VMM4	<1	1	109
Gaskleen® IV Series Assembly SGLFPF6402VMM4	<1	1	109

The efficiency of the POU filter assemblies was determined by challenging the filter with an aerosol produced from a 0.04% aqueous sodium chloride solution. The atomized NaCl_aq provided a particulate challenge with the particle size distribution outlined in Figure II. The aerosol had higher particle concentration in the most penetrating particle size range. The challenge level was 1.2 x 10⁹ counts/ft³ for particles ≥3 nm and the downstream counts were monitored for 90 minutes at a steady test flow. The air flows in question were 50 slpm and 110 slpm for the 4400 Ultramet-L filter assembly and the Gaskleen IV filter assembly, respectively. The results of the efficiency testing are shown in Table I. As no counts were detected downstream of the POU filter assemblies evaluated, a retention efficiency of 10⁹ for particles ≥3nm including the most penetrating particle size was demonstrated.


FIGURE II
Typical Particle Size Distribution for Polydisperse 0.04% NaCl(aq) Challenge Aerosol.

A number of studies^(1,2) have revealed that all metal filters (4400 Ultramet-L Gaskleen Filter Assembly) do not exhibit an appreciable level of outgassing. In addition, the design of the 4400 Ultramet-L Gaskleen Filter Assembly ensures that the assembly exhibits a high degree of transparency to moisture⁽³⁾. Transparency ensures that significant levels of moisture are not adsorbed by the filter and released slowly in to the gas stream over a prolonged time period.

The tendency for fluoropolymers to outgas, and the associated temperature constraints has limited the use of fluoropolymer filters in gas distribution systems. The transparency of a preconditioned fluoropolymer filter to moisture was investigated in the present study.

The Gaskleen^® IV filter assembly is preconditioned during the manufacturing process to ensure a moisture contribution of < 10 ppb above the background level. For the purposes of this test a Gaskleen IV filter assembly (fluorocarbon filter element within a stainless steel housing) was uncapped and exposed to 42% relative humidity for 24 hours (no flow). At the end of the exposure period the moisture level desorbed from the assembly was monitored with an Alphadew (Stephens Analytical, Montreal, Canada) hygrometer. The hygrometer relies on the detection of moisture by a silicon chip sensor. The sensor responds to the partial pressure of water vapor in the sample gas and gives a moisture concentration readout.

The test stand employed to monitor the moisture levels downstream of the test component is illustrated in Figure III.

FIGURE III
Setup for moisture testing.

The test setup is based on the recommended setup outlined in SEMATECH Semaspec # 90120397B-STD - "Test Method for the Determination of Moisture Contribution by Gas Distribution System Components". The carrier gas employed was purified nitrogen. The moisture levels downstream of the Gaskleen IV filter assembly are shown in Figure IV. The flowrate of the purified nitrogen carrier gas was set at 5 slpm for the purpose of this testing. The maximum recommended flowrate for the Gaskleen IV filter is 400 slpm.

The baseline for the test system at ambient temperature was measured as 15 ppb. After 30 minutes the temperature of the system was increased to 70 degrees C and the moisture levels were monitored for an additional 60 minutes. The moisture levels remained 20 ppb throughout this time period. The spoolpiece was removed from the test stand and subsequently reinstalled to determine the moisture contribution due to the installation process. The spoolpiece refers to a length of electropolished 316L stainless steel with 1/4" VCR fittings. The moisture level downstream of the spoolpiece was monitored for 1 hour at 22ºC and 1 hour at 70 degrees C. The moisture level remained 20 ppb at both temperatures.

The Gaskleen® IV filter assembly which had previously been exposed to 42% relative humidity for 24 hours was installed on the test stand. The filter assembly was subjected to the 10 minute purge (as recommended) at 0.5 scfm prior to commencing moisture measurements. The moisture level downstream of the Gaskleen IV filter assembly reached a maximum of 35 ppb above background within 20 minutes of installation. The moisture levels decayed to 15 ppb above baseline within 180 minutes of installation at ambient temperature. The filter assembly was subsequently heated to 70 degrees C and the moisture level rapidly increased to 70 ppb above baseline. The moisture subsequently decayed to 15 ppb above baseline within 180 minutes of increasing the temperature.

FIGURE IV
Moisture levels detected downstream of Gaskleen IV Filter Assembly.


The moisture levels observed indicate that the Gaskleen IV filter assembly was transparent to the moisture. The rapid desorption at 70 degrees C is followed by a slower release of moisture from the polymer, as moisture permeates to the surface. The final moisture level may be acceptable in a number of processes.

A number of semiconductor processes, including diffusion, CVD and EPI are performed at temperatures above the maximum recommended temperature for the fluoropolymer filter. In addition, fluoropolymer filters are typically sealed to the filter housing by means of an elastomeric seal - it should be noted that the Gaskleen IV filter assembly does not contain any elastomeric seal. The correct selection of the elastomeric seal material is instrumental in ensuring long term compatibility of the filter assembly with the process gas.

The practice of heat tracing gas lines to ensure rapid desorption of moisture has resulted in the use of all metal filters. The baking of the gas lines including the filter assemblies is typically performed at temperatures in excess of 200 degrees C.

The critical factors which influence the selection of a filter assembly for use in a UHP gas distribution system will typically be temperature and compatibility considerations. The all metal filter was originally developed to ensure minimal outgassing of chemical impurities into the gas stream and allowed for high temperature operation.

In cases where fluoropolymer and all metal filters are suitable for use, the outgassing and transparency characteristics of the filter are determined by the material selection, filter design, manufacturing conditions and post assembly conditioning. In a number of cases the fluoropolymer filter will meet the required specification.