Abstract

The need to verify the performance of a volumetric Optical Particle Counter (OPC) is necessary in many applications. This paper describes the proper way to verify the performance of a Particle Measuring Systems' LiQuilaz-S series OPC by calculating the sizing accuracy. To do this, one must precisely collect "split" data and know the instrument resolution. The following instructions are for collecting data in Facility Net Software using LiQuilaz-S series volumetric spectrometers. The sizing calculator is available on the Particle Measuring Systems website.

Introduction

One of the most common questions regarding the performance of an OPC is whether it sizes and counts accurately. Counting accuracy refers to the instruments ability to count the correct concentration of particles in a solution; sizing accuracy refers to the instruments ability to correctly determine the size of a Polystyrene Latex Sphere (PSL). Accuracy has implications in qualifying chemical delivery systems, accepting incoming chemicals, or qualifying incoming parts. The need for OPCs to match and to qualify their performance is thus very important. In order to verify an OPC, you must use appropriate methods to ensure time and or money is not wasted. These methods include good laboratory technique, proper equipment set up, data collection, and using the Sizing Calculator as described here. Particle Measuring Systems uses sizing accuracy to verify the performance of a LiQuilaz-S OPC. LiQuilaz-S OPC Design By definition and design, the LiQuilaz spectrometer is a Volumetric instrument that counts 100% of the particles that pass through the sensor. All of the fluid passing through the capillary is analyzed because the capillary is fully illuminated by the laser beam as shown in Picture 1. Therefore, the LiQuilaz spectrometer eliminates the need to perform count standard tests.

Picture 1: Volumetric OPC Capillary Illumination

If, for example, there is something wrong or the particle counter is misaligned then it would not size accurately. Therefore, as long as the sensor is sizing accurately, it is counting accurately.

Size Standards

Polystyrene Latex Spheres (PSLs) are industry recognized size standards that are NIST traceable and used for instrument calibration. They are readily available from Duke Scientific in various sizes. Table 1 lists the recommended PSL sizes for verification.

Picture 2: Polystyrene Latex Spheres

Test Setup

To properly verify the performance of the OPC, you must correctly design the test and set up the test equipment. Picture 3 shows a typical particle injection system. Laboratory technique is critical. Individuals performing the sizing accuracy test must have experience measuring particles in ultra-pure liquids, preparing laboratory glassware, particle counter operation, and posses overall good laboratory techniques and practices. The following requirements must be met to determine the sizing accuracy of the OPC. - The OPC must be installed online with a particle injection system (e.g.:. PMS VSPI) capable of injecting PSLs into the water.

• The flow rate must be precisely set to the calibrated flow rate using a flow controller calibrated with a stop watch and graduated cylinder. Typical flow rates are 20, 50, or 80 mL/min. The flow rate should be verified for each instrument.
• The DI water being used in the test must be below 100 cts/mL for the most sensitive size channel of the OPC under test.
• The total concentration of PSLs injected into the system should be 3000 +/- 500 particles/mL.
• The DC light must be within specification (S02 <0.5V, S03/S05 <0.05V).
• If multiple units are under test, they are connected properly in parallel.

Picture 3: Particle Injection System

Data Collection

Data collection is equally important in properly verifying the performance of the OPC. Particle size channels must be defined with the exact particle size of interest. The exact particle size can be found on the PSL bottle as shown in picture 2. The exact particle size is listed as Diam: 0.350 µm. The exact size must be used, not the nominal size. The first channel will be the 1st channel of the particle counter. Next, enter the particle size channel that is the exact particle size on the PSL bottle. Enter an equal number of channels above and below the particle of interest using the channel spacing listed in table 1. Channel spacing is defined under the instrument configuration in Facility Net.

Table 1: Channel Spacing for Sizing Calculator

The following example in Figure 1 shows proper channel configuration for a LiQuilaz S02 using the nominal 0.35 µm PSL. The first channel of the LiQuilaz S02 is 0.2 µm. The exact PSL size is 0.350 µm. Channels are set up above and below 0.350 µm using 0.02 µm channel spacing. PSLs are injected showing a clearly defined peak around the particle size of 0.350 µm. One should notice with very clean water there is zero to relatively few particles in the smallest size channels.

Figure 1: LiQuilaz 0.350 µm PSL Data

It is critical that clean water is used during these tests. As one will see, collecting the split data requires summing the particles above and below the PSL size. If there is contamination in the samples, it will artificially increase the counts below the PSL size thus skewing the data towards a negative sizing error. Figure 2 shows high numbers of counts in the smallest particle channels due to contamination.

Figure 2: LiQuilaz 0.350 µm PSL Data with Contamination

Split Data and Sizing Accuracy

Next, "split" data must be collected and entered into the sizing calculator. "Splits" are defined as the ratio of total number of counts below and above the particle size of interest. In a perfect OPC, the "split" would be 50%, that is half the PSLs would be below 0.350 µm and half would be above 0.350 µm. In the real world, instrument resolution and sizing accuracy determine the actual "splits".

For example in Figure 1, the low "split" is calculated by summing the total number of differential counts between 0.210 and 0.330. The high split is calculated by summing the total number of differential counts 0.350 to 0.470. The low split is 2070.6 counts and the high split is 1137.0 counts. The split ratio in this example is 65% - 35%.

Once data is entered into the sizing calculator, the sizing accuracy of the OPC can be determined. The sizing calculator requires the user to input the exact PSL size, the total number of counts, and the number of counts above the PSL size. The calculator then calculates the sizing error. Instrument resolution is the average instrument resolution empirically determined over time.

Using the same example in Figure 1, one would enter 0.350 µm into the Mean PSL size cell. Enter 3207.6 into the Total cell and 1137.0 into the High cell. The sizing calculator in spreadsheet 1 shows that this OPC has a sizing error of -1.87%.

ISO standards suggest that instruments of this type should have sizing errors of +/-10% or less. If the sizing error is greater than 10% one should review their procedures. If it is determined that the sizing error is indeed greater than 10%, the LiQuilaz will require service and calibration by an authorized technician. It is important to note the size and/or shape of the peak is not indicative of sizing accuracy, nor are the actual splits.

Spreadsheet 1: Sizing Calculator

Instrument Model : LiQuilaz S02 Instrument Serial Number: Date:

Particle Sizing Accuracy Calculator

Sizing accuracy test should be performed at a calibration particle size (error at time of calibration is near zero). The exact PSL mean size should be used. ALL CALCULATOR INPUTS ARE IN YELLOW CELLS. Enter the exact PSL size into the calculator. After performing the split test, enter the total number of the counts observed into the Total cell and the total number of counts observed above the particle size of interest into the High cell. The program calculates Relative Size Error.

Channel 0.35 Inputs Mean PSL particle size (EXACT) 0.350 Channel Resolution 5% Total 3207.60 High 1137.00 Split Ratio (high size counts /total counts) 0.354 Outputs Standard Deviation 0.018 Cumulative proportion 0.65 Calculated PSL Size 0.343 Relative Size Error -1.87% Calculated Threshold Position 0.357

Counting Error has the same sign as the Relative Size Error A negative counting error means that the sensor is undercounting A positive counting error means that the sensor is overcounting

Counting Error

After the sizing error of the OPC is known, you can calculate the actual counting error for a real world sample. Calculating the counting error assumes a normal ambient distribution of particles with a distribution of D^-3 (where D is the particle diameter). Continuing with our previous example, the counting error for total cumulative counts would be -5.8% as shown in spreadsheet 2. This means that in this example, the OPC is counting 5.8% less total cumulative particles 0.35 µm and greater than a perfectly calibrated OPC would be counting.

Spreadsheet 2: Counting Error

Counting Error Estimate Particle Counter PSL Mean Size (µm) Calculated PSL Size (µm) PSD Relative Size Error Calculated Threshold (µm) Counting Error for Ambient Fluids

LiQuilazS02 0.350 0.343 -3 -1.87% 0.36 -5.81%

Conclusion

Verifying the sizing accuracy of a Liquid Particle Counter is a useful way to determine if it is working properly. If it is sizing correctly then it is counting correctly and ultimately will provide the best instrument to instrument matching as is often required by users. The techniques and methodology for determining the sizing accuracy as described here are the best methods to verify instrument performance. Particle Measuring Systems uses sizing accuracy to verify the performance of a LiQuilaz-S OPC.

App65a 200707 LiQuilaz® is a registered trademark of Particle Measuring Systems, Inc. Author: Roger Carlone, Particle Measuring Systems