The performance of mechanical action pipettes must be tested periodically to ensure accurate liquid delivery. The results of such testing then may be compared with pre-established tolerances, and out-of-tolerance conditions corrected. Tolerances that are too strict can cause a large number of so-called “false failures,” where a pipette in good working order produces test results that are out of tolerance. Tolerances that are too broad can degrade the quality of the laboratory’s analytical work.
Further complicating the situation is the fact that pipette performance is dependent upon factors beyond the mechanical condition of the pipette. Balancing this interplay of factors can make it a challenge to develop achievable tolerance limits for pipettes. This Lab Report reviews the key issues that should be considered when establishing tolerance limits for the working laboratory, and recommends a set of achievable tolerance limits for single and multi-channel pipettes of various sizes.
Many laboratories attempt to apply pipette manufacturers’ specifications as the tolerance limits for their in-lab testing program. However, a manufacturer’s specifications may not accurately reflect the performance attained in the working laboratory.
Users frequently find it difficult to reproduce a manufacturer’s performance claims, for a number of reasons:
Based on our experience with many different makes and models of pipettes, including single and multi-channels, Artel recommends the values in Table 1 as a starting point for achievable tolerance limits. These limits are based on a simple guideline: Two percent of full scale at all volume settings.5 For example, the systematic error for a 100 μL variable-volume pipette is ±2.0 μL (2%) at the 100 μL setting, and ±2.0 μL (4%) at the 50 μL setting.
This type of generalized tolerance limit has been employed successfully in a number of other fields, such as humidity measurement and syringe calibration, where a fixed percentage of full scale reading is the customary means for specifying performance. The ISO 8655-2 standard for pipette testing1 also uses a percentage of full scale approach.
The allowable percentage that ISO specifies is fairly constant for medium and large volume pipettes, but increases for the smallest volume pipettes. This deviation is necessary when using gravimetric methods in order to accommodate the increased error inherent in gravimetry at smaller volumes.6 If a more exact method such as ratiometric photometry is used, then a constant percentage of full scale may be applied to even the smallest volume pipettes.
The tolerance limits recommended here are based on what is typically achievable by a reasonably skilled operator. These recommendations do not take into account the more stringent data quality requirements of a particularly demanding analytical method. In such circumstances, laboratories should evaluate the results of past testing to fine-tune the initial tolerance limits relative to the requirements of the method. The following examples illustrate solutions for common problems encountered when establishing pipette tolerances.
An analytical method requires dispensing a 100 μL sample with ±3% systematic error. The laboratory has been using a 200 μL pipette set to 100 μL for this purpose. Table 1 shows the recommended tolerance to be 4%, which is too liberal for the method. The simplest and most reliable solution is to replace the 200 μL pipette with a 100 μL pipette. This pipette, when used at its full scale setting, can be tested against a 2% tolerance.
An analytical method requires ±1% systematic error at a volume of 1,000 μL. This is greater trueness than for any pipette in Table 1. Pipette performance data are examined to determine whether this degree of trueness can be attained. It is found that two particular operators are regularly attaining the desired level of performance when using a particular make and model of pipette, while other operators are not. The superior pipette is specified in the procedure, and the highly skilled operators are used as benchmarks against which others may be trained. The tolerance limit for this pipette can then be tightened to ±1% without causing a large number of false failures.
Begin by choosing either the “Relative Error” or “Absolute Error” tolerance limit values. These tolerance limits reflect what is reasonably achievable in a working laboratory. They presume that the pipette is calibrated and functioning properly, is used with good quality tips, and is tested by a reasonably skilled operator. When these criteria are met, most makes and models of pipettes should test within these tolerance limits unless they are mechanically defective.
For fixed-volume pipettes, the nominal value is the fixed volume. For variable-volume pipettes, the nominal value is the largest user-selectable volume setting; e.g., a 10-100 μL pipette has a nominal volume of 100 μL.
The absolute error for the nominal volume applies to every selectable pipette volume; e.g., a 100 μL nominal volume yields limits of ±2.0 μL systematic error (mean value) and less than 1.0 μL random error (measured as a standard deviation) for all volumes. The relative error varies throughout the pipette range; e.g., for a 10-100 μL pipette at 100 μL the relative systematic error is ±2.0%. However, at 10 μL the relative systematic error is ±20.0%.