36


Accuracy depends on knowing the sample variables (shape, density, refractive index) and instrument variables (calibration, alignment, temperature). Good accuracy implies good sampling and sample preparation techniques have been used. Sometimes accuracy is important; sometimes it is not. Materials used in the coatings industry need to be characterised accurately. The large particles affect the film forming capability of the coating; the medium size particles affect the light scattering properties; and the small particles control the rheology. In quality and process control applications, relative changes from batch-to-batch are much more important than accuracy. In these cases, reproducibility is the main specification.


Relative numbers are acceptable unless they have to be compared with other techniques or absolute requirements. Then accuracy becomes paramount.


Accuracy has often been defined by the historical use of an instrument in a particular field. Although not really a definition, its practicality, however, cannot be ignored. New instrumentation should agree or, at least, correlate with the historical results. But if this argument is carried too far, then bad measurements are perpetuated. Most instruments claim accuracy when tested with spherical standards. There are very few reliable standards. There are, however, reference materials for checking precision, reproducibility and resolution. While useful, these are not absolute standards, and, as such, should not be confused with them.


Recommendation:


Even if you are only interested in relative changes, test an instrument with reference materials just to verify the precision, resolution and reproducibility claims.


Precision: Instrument precision is a measure of the variance in repeated measurements on the same sample. Precision limits resolution, reproducibility and accuracy. Precision is a useful criterion by which to assess instruments even if the accuracy cannot be determined. The precision of a measurement may be +/- 1%, yet the absolute accuracy might be much worse. It is common to have good precision but poor accuracy.


Reproducibility: Reproducibility is a measure of the variance from sample-to- sample or instrument-to-instrument or operator-to-operator. If you only have one instrument and one operator, then questions of reproducibility may not be of much interest. But if you have several plant operations, with several operators, all using the same manufacturer’s model, then check reproducibility. If it is much worse than the basic precision of any one instrument, then look for the source of the error. Is it preparation differences, or variations from one instrument to the other?


Variations in instrument performance are much greater than most novices would guess. These can occur because of a change in production technique, detector response, software, or a combination of all three.


Recommendation:


Always perform round-robin tests using the same sample; this can reduce or eliminate sample variations. Send an exact set of common operator instructions with the sample to minimize operator variations. The results should quantify instrument-to-instrument variations.


Resolution: Resolution has two quite distinct definitions in particle sizing. The first definition concerns the minimum detectable differences between different runs. It answers the question, ‘Can the differences between two samples be resolved?’ This definition is closely related to the precision of the measurement.


The second definition concerns the minimum detectable differences between features of the size distribution in one run. The simplest example is the ratio between two peaks in a bimodal distribution. If the minimum ratio is 2-to-1, then the resolution is rather low. If it is 1.1-to-1, then it is rather high. Ensemble averaging instruments, all forms of light scattering and diffraction in particular, are medium to low-resolution instruments.


Beyond a certain point resolution is not determined by the number of channels in a SPC, nor by the number of reported size classes, nor by the resolution of the output devices (CRT, printer) used to format the results. Yet, many manufacturers specifications would have you believe that resolution is defined in one of these ways. Resolution is, fundamentally, a function of the basic signal-to-noise ratio of the instrument. Reporting more than the fundamental resolution is like magnifying the noise: more numbers are obtained, but they are meaningless.


Above one micron it is quite common for ground material to exhibit very broad distributions. In this case resolution is seemingly not very important.


Do not be fooled by this common assertion. If the fundamental resolution of an instrument is undetermined, then how does one know if the broad distribution is really hiding practical and, possibly, significant information? Are those long tails real? After all, low-resolution instruments often smear out the distribution producing unrealistically long tails.


Recommendation: Test resolution by mixing narrowly distributed and previously measured samples - the reference standards.


Accuracy, precision, resolution, and reproducibility are functions of the size range. Errors are greatest at the extremes. If possible, do not purchase an instrument for measurements at the extremes. A common mistake is to check an instrument in its midrange and then proceed to use it at one or another of the extremes.


Be skeptical of claims of accuracy and precision. if these really refer only to the average size. If it is not clear from the manufacturer's literature, then ask for


clarification. The average of any distribution is least subject to variation. Even instruments with poor resolution and instrument-to-instrument reproducibility may yield results with 1% or 2% precision in the average for any one instrument. Higher moments, such as the measure of width or skewness and the tails of the distribution, are more sensitive to uncertainties. So pay particular attention to the variance in some of these more sensitive statistics when evaluating instrumentation.


Support: Support is defined here as good technical support. Is the manufacturer familiar with your particular problem? Can they suggest sample preparation techniques? To support you after the sale, does the manufacturer offer adequate training, good technical manuals, and experts available to help you interpret results?


The instrument manufacturer should have a laboratory with other instruments available with which to validate the usefulness of the proposed instrument. Sample preparation techniques are often the key to good measurements, and the manufacturer should guide you in this aspect of particle sizing. A continuing program of development by the manufacturer will ensure the user that the instrument will not become obsolete in the near future.


Recommendation:


Judge the level of support you will need. Question instrument manufacturers on how they will provide support. Ask for references to verify any claims that are made.


Ease-of-Use: There is nothing more subjective than the concept of ease-of-use. In one limit it means automated sample preparation, automated instrument control, and automated data analysis and printout - all unattended.


Some manufacturers strive for this under the banner of the ‘one button’ instrument.


Other users think that an instrument is incomplete without a complete data archiving, retrieval, and data base management system. These objectives are hardly ‘one button’. They require a rudimentary knowledge of desktop computer operation.


Recommendation:


If ease-of-use is important to your application then be sure to watch measurements being made before you purchase. Make sure that the entire process - sample prep, measurement, data analysis, and cleanup - is demonstrated.


Versatility: Versatility is here defined as the ability to measure a wide variety of samples and sizes under a variety of sample preparation conditions. For example, the electrozone technique requires a conducting liquid, which is most often water with an electrolyte (salt) added. For many applications this condition is not restrictive; for others it is. Electron microscopes cannot be used on samples that sublime under a vacuum. Some instruments work with almost any liquid; others do not. Either the technique may be limited, or its implementation by a particular manufacturer may be.


Recommendation: Try to estimate a realistic range of samples and the corresponding size ranges that you intend to measure. Experience shows that it is usually better to choose dedicated instruments that do a good job for their intended purpose rather than going for the ‘zero-to-infinity’ machines, which do a poor job on a variety of samples.


Life Cycle Costing: Instrument cost is the least and the most significant part of purchasing an instrument. If the instrument cannot perform the appointed tasks, it is no bargain at any price. If it can do the job properly, it may be a bargain at twice the price.


Particle sizing instruments vary in price from a few hundred dollars (pipettes, turbidimeters, simple microscopes) to a few hundred thousand dollars (electron microscopes complete with image analysis software). As of the publication of this article, most modern instruments range from $15,000 to $60,000 with the majority around $30,000. But the initial cost of an instrument is only part of its total cost.


The total price of an instrument is best judged in terms of the life cycle cost. This includes initial price, operating cost, and maintenance and repair costs. Every instrument needs some type of maintenance. It may be as simple as cleaning air filters once every 3 months. It may be as difficult as replacing mechanical parts or aligning an optical system. To some, these are not difficult tasks; to others they are. Every instrument will, sooner or later, require repairs. Any vendor who denies this is not worthy of further consideration.


Recommendations:


Ask the vendor for a list of users who have had the instrument for at least one year. Ask these users for their experience with maintenance and repairs. Ask the vendor what the typical problems have been, and what cures are necessary. Ask about maintenance. Compare the user and vendor responses.


Summary: The mix and priority of quantitative and qualitative specifications you use in making your decision will, to some extent, be determined by your intended use.


Although it may be dangerous to pigeonhole your intended use by putting it into one of the three categories shown in Table 1, it may also help you to focus on what factors are most important in solving your particle sizing problem.


Remember, many users do not fall into such neat categories. And, one person's research may be another person's quality assurance. But if you recognise a pattern in one of these categories that fits your needs, do not hesitate to use them to organise your thinking. Ultimately, you will make a better choice.


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44