(Download this paper for this formula.) (163.5 KB)

Download the Performance Calculator separately (this file is the instructions for use)

Introduction

Frequently particle counter users are faced with trying to determine which particle counting system will provide the best statistical data for monitoring their DI water system. When comparing the specifications of different particle counters, many of the specifications appear similar making it difficult to ascertain which system is superior and most appropriate for monitoring the liquid system in question.

Particle Measuring Systems has addressed the need for a gauging device by developing a simple tool named the Performance calculator. The Performance Calculator allows users to enter the relevant specifications of different particle counting systems. Once the information is entered, the Performance Calculator produces measures that indicate a particle counter's ability to produce good statistical data allowing users to improve control over the liquid system saving the time and costs associated with responding to false alarm conditions.

Description

The Performance Calculator allows the user the ability to input different parameters of up to three particle counters. Once each particle counter's parameters are entered, the Performance Calculator will produce information on how the particle counter's sensitivity and sample volume affect the data. This benefit is illustrated through looking at the number of alarms that may be generated on a given system.

There are three main variables that influence the quality of data for liquid particle counters. These are sample rate, sensitivity, and the Particle Size Distribution (PSD) within the liquid being sampled. The first, sample rate, influences the particle counter's ability to obtain a meaningful description of the sample. This is explained through simple population statistics where the larger the sample the better the description of the population. The benefit of increasing the sample volume for particle counters is realized when trying to establish different control limits for the liquid system. Because a large sample volume provides a better description, the user is able to set more stringent controls on the system. In particular, this is realized through inter-sample variability. As the sample volume increases, inter-sample variability decreases which in turn reduces the probability that false alarms will be generated on the system making the liquid particle counter easier to control.

To help illustrate the importance of sample volume, the Particle Measuring System HSLIS M50 and Ultra DI® particle counters are compared. The specifications of each instrument are found below. (Download this paper for all tables and figures.) (163.5 KB)

Each particle counter has equivalent sensitivity, but the sample rate is significantly different (sample rate is a product of flow rate and view volume). The importance of the sampling rate is illustrated in the following two figures. As seen in the first figure, particle counters that have a low sample rate will take a longer period of time to reach a statistically significant sample. For instance, if it takes 20 mL to obtain a statistically significant sample, it will take the HSLIS M50 particle counter 80 minutes versus 5.33 minutes for the Ultra DI50 particle counter to obtain a statistically valid sample. The increased sample rate results in a 15x time saving!

The second figure shows how a larger sample population reduces inter-sample variability. As the figure illustrates, the particle counter with the larger sample volume, in this case the Ultra DI50, does not create as many alarm conditions and provides an overall better description of the population due to decreased statistical variation. This allows the user to define tighter control limits without the worry of having to babysit a system that creates false alarms.

The remaining variables that affect data quality are sensitivity and sample PSD. Sensitivity is the smallest size particle that can be detected by a particle counter. Particle Size Distribution, as the name implies, is the manner particles are dispersed within the sample. These two variables are interrelated for determining which particle counter is appropriate for monitoring the system. This relationship is explained by first understanding the different types of PSDs. Examples of different PSDs are found below in Figure 3.

Knollenberg and Veal, in Optical Particle Monitors, Counters and Spectrometers: Performance Characterization, Comparison and Use, discovered that particles had a certain distribution within DI water systems that could be approximated by the power law expression of:

(Download this paper for this formula.) (163.5 KB)

Using this expression, Knollenberg and Veal found that a typical distribution for DI water systems could be estimated by N = 0.0006D-3.

Applying Knollenberg and Veal's work, Mitchell, in Statistical Analysis of Particle Instruments for Liquid-borne Particles: Understanding the Impact of Size Sensitivity and Sample Volume, found this expression could also be used for other ultrapure liquids to approximate the PSD. Applying this to particle counters, more sensitive instruments will count more particles in a given period with sample volume remaining equal. For instance, if comparing between 0.1 and 0.2 mm sensitive particle counters with equivalent sample volumes that are sampling from a D-3 particle distribution, the user could expect to find 8 times the number of particles with the more sensitive instrument!

(Download this paper for all tables and figures.) (163.5 KB)

The importance of the number of particles the instrument detects helps determine which particle counter is most appropriate for the liquid being sampled. For instance, if sampling a liquid with a D-3 distribution with a concentration of 1,000 particles/ml ³ 0.2 mm and greater, an instrument with a maximum concentration of 5,000 particles/ml may be overwhelmed if the sensitivity is 0.1 mm or 0.05 mm. On the other hand if the concentration is very low, such as 1 particle/ml ³ 0.2 mm, the user would be able to specify a more sensitive instrument to obtain a better description of the PSD without overwhelming the particle counter.

Explanation

First presented at Watertech-Portand (3), the Performance Calculator was created from basic particle counting statistics (2) to illustrate the potential number of false alarms that could be generated per day. As seen in figure 4, the Performance Calculator also produces Upper Control Limits (UCL) for each defined particle counting system to show how sensitivity and sample volume combine to reduce variability and improve control over a system.

Within this spreadsheet, the user defines the particle counting systems to be compared by inputting different criteria found within the active cells (orange). Once the information has been entered, the Performance Calculator will issue results (blue) showing how the instruments compare.

Summary

Developed to assist in the selection of particle counters for ultrapure liquid applications, the Performance Calculator provides a means to benchmark different particle counters in regards to sensitivity, sample volume and a liquid system's Particle Size Distribution. In addition, the Performance Calculator indicates which particle counting system will provide the best control over a process in terms of reducing control limits and inter-sample variability. By controlling the statistical variation, the user will be able to reduce the time and associated costs of responding to false alarm conditions.

(Download the Performance Calculator) (30.5 KB).

(Download this paper for all tables and figures.) (163.5 KB)

Ultra DI® and LiQuilaz® are registered trademarks of Particle Measuring Systems, Inc.

Contact us if you need more information or have questions.

Published: 2000