The four numbers above are the gold nanoparticle challenge test results for JIUNYUAN's advanced semiconductor-grade PP membrane filter cartridge (5 nm rating, 10-inch, 2-222 flat end cap, new filter), third-party certified by the Industrial Technology Research Institute (ITRI). The numbers look good on their own, but for process engineers and buyers, what is actually useful is understanding them: under what conditions 99.41% was measured, how much passes through when it is converted to penetration, and whether it can be compared directly with the “5 nm” rating on other filters. This article breaks the data down.
- Why is nanometer-level retention tested with gold nanoparticles?
- Under what conditions was this data measured?
- What does 99.41% mean? Converting to penetration and LRV
- One rating step apart: how much does 5 nm particle penetration differ?
- Why is 20 nm retention slightly lower than 5 nm?
- Ask these 6 questions before comparing retention numbers
- Where does a 5 nm PP membrane filter fit?
- FAQ
Why is nanometer-level retention tested with gold nanoparticles?
For micron-level filters, retention efficiency can be measured by directly counting upstream and downstream particles with a liquid particle counter. But the smaller the particle, the weaker the light it scatters — in the Rayleigh scattering regime, scattered light intensity is roughly proportional to the sixth power of particle diameter, so a 10-fold reduction in particle size makes the signal one million times weaker. At single-digit nanometers, direct counting with optical methods is already very difficult, and a different way to make the particles “visible” is needed.
Gold nanoparticles (abbreviated GNP or AuNP) are currently a commonly used challenge particle for nanometer-level retention testing, for four reasons:
For the full range of test methods from micron to nanometer level, see “Liquid Filter Efficiency: From Conventional to Nanometer Test Methods.”
Under what conditions was this data measured?
A retention number must always be read together with its test conditions. The conditions for this data set are as follows:
| Item | Condition |
|---|---|
| Filter under test | Nanometer-grade advanced PP membrane filter cartridge, 5 nm rating |
| Filter format | 10-inch cartridge, 2-222 flat end cap |
| Filter media | PP |
| Filter condition | New filter (freshly manufactured) |
| Challenge particles | 5 nm and 20 nm gold nanoparticles, each size tested separately |
| Upstream concentration | Gold concentration fed into the system: 500 ppt |
| Sampling | Three sample bottles each upstream and downstream; mean and standard deviation calculated |
| Data verification | Third-party certified by ITRI |
What does 99.41% mean? Converting to penetration and LRV
Once retention is above 99%, the percentages all look about the same. Switching to penetration and LRV (Log Reduction Value) is what brings the differences to light.
Retention = (1 − downstream concentration ÷ upstream concentration) × 100%
Penetration = 100% − retention
LRV = log₁₀(upstream concentration ÷ downstream concentration) = −log₁₀(penetration)
(Use penetration as a decimal when calculating LRV, e.g. enter 0.59% as 0.0059.)
| Challenge size | Retention (mean) | Std. dev. | Penetration | LRV | Downstream gold (estimated) |
|---|---|---|---|---|---|
| 5 nm | 99.41% | 0.18% | 0.59% | 2.23 | approx. 2.95 ppt |
| 20 nm | 99.08% | 0.04% | 0.92% | 2.04 | approx. 4.6 ppt |
Downstream gold concentrations are back-calculated from the 500 ppt upstream concentration and the mean retention. They are shown only to indicate the order of magnitude and are not raw measured values from the report.
Put more intuitively: out of every 1,000 gold particles of 5 nm, about 6 pass through the filter. Lining up common retention figures side by side shows what LRV means:
| Retention | Penetration | LRV | Particles passing per 1,000 |
|---|---|---|---|
| 90% | 10% | 1 | 100 |
| 95% | 5% | 1.3 | 50 |
| 99% | 1% | 2 | 10 |
| 99.41% | 0.59% | 2.23 | approx. 6 |
| 99.9% | 0.1% | 3 | 1 |
One rating step apart: how much does 5 nm particle penetration differ?
The product page for JIUNYUAN's Semiconductor-grade PP Membrane Filter Cartridge also lists a separate data set measured under ultrapure water conditions: the 10 nm-rated filter has 95% retention of 10 nm GNP and 90% retention of 5 nm GNP. Putting the two ratings from the same PP membrane product line side by side shows how many of the same 5 nm gold particles pass through each:
The two data sets come from different tests, and the conditions may not be exactly the same. They are used here only to illustrate the trend.
Against the same 5 nm gold particles, the 10 nm-rated filter lets 10% through, while the 5 nm-rated filter lets only 0.59% through: a difference in penetration of about 17 times.
This comparison also shows something else: retention ratings have no uniform definition. The 10 nm-rated filter above has 95% retention for 10 nm particles, not 99.9%. The retention level each manufacturer uses to define its “X nm” rating, and the particle used to test it, may both differ. So when comparing two filters, do not look only at the rating; look at retention or LRV at the same challenge particle size. For how to choose a rating, see “Liquid Filter Rating Selection: Matching Process Needs.”
Why is 20 nm retention slightly lower than 5 nm?
Intuitively, larger particles are easier to capture, so retention at 20 nm should be higher than at 5 nm. Yet this data shows 99.08% versus 99.41%. On closer inspection, several points are worth noting:
- The gap is clearly larger than the sampling variation. The standard deviations of the two triplicate sample sets are only 0.18% and 0.04%, which shows good repeatability. The 0.33 percentage-point gap is clearly larger than the standard deviations, so it more likely comes from a systematic difference between the two tests than from sampling luck.
- ICP-MS measures “total gold.” The instrument cannot tell whether gold is present as nanoparticles or as dissolved gold ions. Gold nanoparticles are made by reducing gold salts. If trace dissolved gold remains in the challenge solution, that portion passes through no matter how small the membrane pores are, and its content is not necessarily the same in the 5 nm and 20 nm standard solutions.
- Above 99% retention, a very small background can swing the result. Based on 500 ppt upstream, if one challenge solution carries only about 1.6 ppt more gold that cannot be sieved out, that is enough to cause a 0.33 percentage-point gap.
- The two tests were prepared and run separately. The surface stabilizers in the standard solutions for different particle sizes, and the adsorption between particles and the membrane surface, may both differ.
Ask these 6 questions before comparing retention numbers
Whenever you receive a nanometer-level retention report, whichever vendor's filter it covers, you can check it against the list below:
- What is the challenge particle? Gold, PSL (polystyrene latex microspheres), and SiO₂ have different surface properties and different adsorption behavior on the membrane surface, so their numbers are not directly interchangeable.
- What is the challenge particle size? “The retention of a 5 nm-rated filter” and “a filter's retention of 5 nm particles” are two different things.
- What is the upstream concentration? How far is the measured downstream value from the detection limit? Was the dissolved background of the challenge solution itself measured or subtracted?
- What fluid and flow rate were used? Results in ultrapure water are not necessarily the same as results in the actual process chemical.
- Was it a new filter or a flushed one? How many samples were taken per set? Was a standard deviation reported?
- Who ran the test? Third-party data and in-house data carry different weight. Does the report fully disclose particle specifications, chemical matrix, flow rate, and sampling method?
JIUNYUAN's Filter Retention Testing service can challenge filters with 5–60 nm gold nanoparticles, PSL, and SiO₂ standard particles in the chemical matrix and at the flow rate specified by the customer, taking three sample bottles each upstream and downstream. The report lists the mean, standard deviation, and full test conditions.
Where does a 5 nm PP membrane filter fit?
This filter is positioned for point-of-use (POU) filtration of front-end semiconductor chemicals, and pre-filtration of raw materials. PP's advantages are cost and environmental friendliness. Its temperature resistance extends to about 80 °C, and its chemical compatibility covers the two main categories below:
Pure water and various dilute acids and alkalis
Organic solvent process chemicals
FAQ
Why isn't a 5 nm filter verified with a liquid particle counter?
Optical counting relies on light scattered by particles, and scattered light intensity drops sharply as particle size decreases, so at single-digit nanometers direct counting is already very difficult. Gold nanoparticles combined with ICP-MS take a different approach: retention is calculated indirectly by “measuring the concentration of elemental gold,” with quantification at the ppt level.
99.41% is new-filter data. Will it change after a period of use?
This data was measured on a freshly manufactured new filter. Retention in service is affected by factors such as loading, chemicals, and differential pressure, so this data cannot directly represent performance after use. To confirm, a used filter can be tested with the same method.
Can retention be tested with our own process chemicals?
Yes. JIUNYUAN's laboratory introduces standard particles in the chemical matrix and at the flow rate specified by the customer, and agrees on the challenge concentration based on the process conditions. When submitting a filter for testing, please also provide its nominal rating, the process chemicals actually used, and the target flow rate, which serve as the basis for selecting the challenge particles and test conditions.
Both rated 5 nm: how do I choose between PP and PTFE filters?
Start with the chemical and the temperature. For ultrapure water, dilute acids and alkalis, photoresist, and common organic solvents at temperatures up to 80 °C, PP has a cost advantage. For strongly oxidizing acids or high-temperature chemicals, use PTFE. For a full comparison of the four main materials, see “Liquid Filter Materials Compared: PES, PTFE, PP, UPE.”
Is a finer rating always better?
No. The finer the rating, the higher the differential pressure at the same flow rate and the lower the usable flow rate, and the filter may also clog faster. Work backward from the particle size the process needs to capture and the yield risk, and choose the rating that is “just enough.”
Sources
- JIUNYUAN, “Advanced Semiconductor-grade PP Membrane Filter Cartridge” retention test data (third-party certified by ITRI)
- JIUNYUAN, “Semiconductor-grade PP Membrane Filter Cartridge” product page (GNP 10 nm/5 nm measured in ultrapure water)
- JIUNYUAN, “Filter Retention Testing” test methods and instrument specifications