Which Particle Counter from Particle Measuring Systems is Best for You for the Measurement of Ultrapure Water (UPW) and Ultra Clean Fluids

Topics Covered

Background

Monitoring for Alarming Only

Cosmic Rays

Trend Analysis

Conclusion

Background

The purity of water and chemicals used in the semiconductor industry has improved dramatically, creating near particle-free environments. The ITRS roadmap for the future shows continued decreasing line widths and a corresponding decrease in the critical particle size.

The increasing cleanliness of liquids and the decreasing critical particle size has caused a corresponding increase in particle counter capability and cost. It is therefore important to balance cost against features. The issue when purchasing a particle counter for ultrapure water or ultra clean fluids is whether the objective is to monitor only for alarming of extreme out-of-spec conditions or to monitor both for alarming as well as trend analysis.

Today, particles are typically monitored at 0.05 microns in UPW and 0.065 microns in liquid chemicals. There are relatively few particles and it is not uncommon to encounter particle concentrations of these sizes at less than 1 particle/mL or even less than 0.2 particles/mL. The high sensitivity particle monitors required to detect these very small particles have a significant drawback in that they only examine a very small percentage of the total fluid flowing through their sample cells. The smaller the amount sampled, the less data you collect, and the longer it takes to gather sufficient data to be statistically significant.

Therefore, if the purpose of monitoring is for alarming purposes only, a lower flow particle counter such (e.g.: 0.25 mL/min) is adequate. But if the desire is to monitor data trends, then a higher flow counter (e.g.: 3.75 mL/min) is required.

Monitoring for Alarming Only

Instruments with small sample volumes can be used to detect extreme out-of-spec problems. For example, a particle counter with a flow rate of 100 mL/min that only examines 0.25% of this flow results in a sampled volume rate (often referred to as sample volume) of 0.25 mL/min. Some products in the market examine as low as 0.1 mL/min. This sampling rate affects the length of time required to gather results and therefore the amount of sample monitored over a specific period.

Assuming the particle concentration of the UPW is 0.5 particles/mL >0.05 microns, what sample size is needed to collect meaningful data and be is representative of the liquid tested? Sample size and analysis time are inextricably linked. If the analysis time is too short and very few particles are counted, then the sample-to-sample repeatability will be very poor. If the analysis time is too long, then the analyst will be blind to changes in the process between sample intervals. These concerns can be addressed with a simple examination of expected purity of the liquid chemical, the sample volume of the particle counter, and counting statistics. Counting statistics predicts the repeatability when a certain number of raw events are measured. The precision expected is the square root of the number of raw counts divided by the total number of raw counts.

Precision Percentage = √ (# raw counts) / # raw counts)

For example, if a 20% relative standard deviation is desired, then 20 counts must be measured during the sample interval (square root of 20 is 4.47 and divided by 20 = 22%).

The conditions cited above for concentration and sample volume (0.5 cts/mL > 0.05 µm and 0.25 mL/min) predict that we would need 160 minutes to detect 20 particles. It should be noted that only 40 mL of fluid would be examined during this entire time. Since the time interval is too long to be practical, this sensor would be more suitable for alarming, rather than trending.

Cosmic Rays

High energy electromagnetic radiation or high energy particles can affect the results. While some of the effect may be due to natural background radiation caused by nuclear transition, all sources of background are typically referred to as cosmic ray events. Cosmic rays impact the detector about once per minute and are indistinguishable from particles due to the nature of the optical particle counter (i.e.: examining light scattering produced from particles).

Particle counters can employ hardware and software correction schemes to correct for false counts. This is especially important for sensors with low sample volumes since the number of cosmic ray events can exceed the true particle concentrations in UPW. The frequency of cosmic ray events tends to be constant for a given location.

Trend Analysis

Collecting repeatability data for trend analysis requires more raw counts per minute than low sample volume particle counters can provide. Counters with medium sample volumes of 1 to 10 mL/min can meet this need. For a sample volume of 3.75 mL/min the time to measure 40 mL of fluid (for an average of 20 particles) is only 10 minutes. This is contrasted to a low flow alarming device (“Sensor B” which requires 160 minutes to produce the same data (see the table below).

Table 1 highlights the differences between the technologies available today. The ideal liquid monitoring strategy is to combine low flow alarm devices with high flow monitors for trend analysis.

Table 1. Sample volume rate differences between particle counters affects the time needed to measure a specific volume of fluid. If the particle concentration is 0.5 particles/mL, only 20 particles will normally be detected in a 40 mL volume.

Name

Purpose

Flow

Sample Volume

Time to Measure 1L of Fluid

Time to Measure 40 ML of Fluid

Sensor A

Trending

High Flow

3.75 mL/min

4.4 hours

10.6 minutes

Sensor B

Alarming

Low Flow

0.25 mL/min

28 Days

2 hours
40 minutes

Sensor C

Alarming

Low Flow

0.1 mL/min

6.9 Days

6 hours
40 minutes

Conclusion

Before purchasing a particle counter it is important to understand the goal of particle monitoring. If the goal is to provide trend analysis and to be alerted in cases of out-of-spec conditions, then the particle monitor must not only have high sensitivity, but also a high sample volume. However, if the need is simply an alarm when the liquid conditions are extremely out-of spec, it is more economical to purchase a particle counter with a smaller sample volume.

Source: Particle Measuring Systems

For more information on this source please visit Particle Measuring Systems

 

Date Added: Oct 31, 2007 | Updated: Jun 11, 2013
Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit