For me, all of this started not with pipettes, but with automatic burettes. We were in fact not allowed to use pipettes in that laboratory, in favor of class ‘A’ volumetric glassware. The sample preparation deliveries supported HPLC test methods, and I can’t say I blamed the scientific director for disallowing their use, due to the criticality of the process. The analytical chemistry department we supported could not do with much less than the manufacturers stated tolerances for the auto burettes, and we had a very difficult time trying to achieve those specifications. It wasn’t the product’s incapability of meeting what the OEM stated on the spec sheet; we simply weren’t able to realize these specs with the lack of measurement science that we incorporated in the calibration.
We were using the very balances that I described in the analytical balance editorial. Just in case you’re jogging your memory, they were quite capable of supporting the calibration due to the time and money we invested on their operating environment and qualifications. We also took pains to measure the water density with an accurate temperature probe. But alas, we were stumped as to why we were rarely meeting the accuracy required.
We contacted the OEM regarding this issue, and got some very interesting information.
As you might have guessed, the integrity of the gravimetric devices used in this science is of utmost importance, as is the determination of the density of the water used as a delivery medium. There were no surprises when we perused that part of the technical manual sent to us from the Swiss. We were enlightened to read that we were not taking into consideration the laboratory air temperature, relative humidity, density of air, station pressure, degassing of the water (purified, of course), gas traps on the holding vessel, evaporation, etc. Many of the aforementioned contribute to what’s known as the Z-Factor: a critical correction applied to this calibration. If we’re concerned with a tight uncertainty budget, maintaining many of these environmental conditions to a degree of stability is also critical. After a few trials and tribulations, we were able to achieve our target specifications. The calibration SOP went from a few pages to a full-blown document once we incorporated all the additional environmental measurements, calculations and supporting documentation. At that point, we had learned some valuable lessons, and felt rather triumphant when we successfully incorporated the new science into our knowledge base, only I had no idea how much this would help me in the future.
When I took on the task of setting up a calibration program in a biotech environment, I naturally inherited the oversight of pipette calibrations for multiple departments. ‘Critical pipettes were sent off to a CCL (Contract Calibration Laboratory), and non-critical pipettes were ‘calibrated’ by a vendor that came onsite with an analytical balance and a few tackle boxes full of parts and tools, all sporting their company logo on baseball caps.
When I finally got the chance to investigate the pipette program, I quickly figured out that I had my work cut out for me to incorporate some much-needed improvements. The ‘pipette guys’, all two of them, were turning out over half of our entire roster of non-critical pipettes a day. Many of them were 50µl or less, all returned cleaned, repaired and/or adjusted to ‘manufacturer specs’.
I also beat David Blaine’s underwater stunt once, and it’s a pity you missed it.
First of all, an analytical balance needs to stabilize to local environmental conditions for at least the time it should have took them to service a good number of pipettes. The more temperature equilibration needed, the more time the balance will need to stabilize. I wondered what environmental conditions the balance was exposed to in transit. Did it sit in a service vehicle in the cold all night? Second, a full-blown qualification, such as the test stated in USP 41 should be executed to verify the analytical balance is operational and properly installed to support the necessary weight ranges – in its exact place of use. They were placing a few higher standard weight denominations on the balance to ‘check it’, before the balance ever had a chance to equilibrate. They did their ‘magic’ on a flimsy conference table, which consequently supported the balance. To accurately weigh 100mg (the uncorrected equivalent of 100µl) in those conditions would have been a feat. It was highly unlikely that the most critical component, the gravimetric device, was even accurate for the intended use as installed. They did at least take the density of water into account, but all the other measurement errors could easily match or exceed (false negative or false positive) the manufacturer tolerances of many of the pipettes.
Back to the auto burettes, we were measuring much higher volumes than those typically aspirated using pipettes, so the Z-Factor, alongside everything else, becomes much more relevant for smaller volumes.
While the assessment of the CCL used at the time showed results that were less extreme as the onsite contractors, I still didn’t have a great level of confidence with them either. I began the search for a viable alternative, and placed a number of recommended CCL’s under scrutiny. One of which made the grade. The technical director and I developed a brilliant working relationship over the years. The laboratory is to this day one of the few accredited to perform pipette calibrations in the US to ISO 17025. I was amazed at the laboratories infrastructure and environmental control specifications, which were clearly conducive to paving the way for quality work. It’s notable to mention that the quality of the gravimetric capabilities alone in such a lab greatly exceeds most of those found in a typical analytical laboratory. Even in the balance ‘Mecca’ I described in the earlier piece, I would not have trusted accuracy greater than +/- 0.1% below 20mg (the uncorrected equivalent of 20µl), and that’s weighing solids, not water, which would evaporate rapidly without necessary precautions. There were still too many limitations compared to that of a properly commissioned mass laboratory. Unfortunately, volumes this small are commonplace for pipettes.
Performing a ‘round robin’ of the pipettes freshly calibrated by what quickly became the former contractors and the accredited lab yielded some very divergent results in many cases, and while we paid more for the quality calibrations, it was easily justified in the end.
For the less critical pipettes, qualified onsite contractors where hired, who used a spectrophotometric means of calibration. This application circumvented the maintaining and determination of many of the environmental conditions that contribute to the Z-Factor, as well as gravimetric measurements, removing the analytical balance from the equation.
Many companies have purchased these ‘spectrophotometers for pipettes’, and use them for periodic checks between calibration intervals to avoid OOT conditions by pulling pipettes from service before they get too close to the specified limit of error. These instruments are also dynamite for training end-users how to properly pipette, which, where lacking, is yet another significant source of error.
For such a simple concept, the reality is indeed not so simple at all.