Validate and Calibrate

FROM THE OCTOBER ISSUE: Deploying best practices that include a regular schedule for equipment monitoring, as well as staff with eyes on the details, improves accuracy.
By Susanne Retka Schill | September 18, 2019

Precision in the ethanol laboratory is important to the bottom line. Being 100 percent accurate is unrealistic, but consistently hitting the target close to the bullseye is important when tiny samples from enormous tanks are used to monitor fermentation or end-product quality assurance.

Most plants have well-run labs, says Tera Stoughtenger, technical services manager at Lallemand Biofuels & Distilled Spirits, but there are times when labs might be understaffed or overworked. She cites one plant that started up without appropriate staff for the necessary sampling. The pH probe in the liquefaction tank was indicating high levels, and with the probe in a control loop, sulfuric acid was automatically added. Because of the missed sampling, the probe’s results weren’t being crosschecked with a lab test and the operators were unaware the pH probe was producing a false high. Not only could that result in unneeded chemical addition, Stoughtenger says, but if the pH gets too low, interference with enzyme action or the inhibition of yeast could reduce ethanol yield.

Another potentially costly mishap occurred at a plant where Stoughtenger worked years ago. Samples were saved in a refrigerator, but the labels used were similar to those used on standards. A standard is a solution with known values used to verify that the high performance liquid chromatography (HPLC) instrument is reading correctly. The night shift operators had run a sample to monitor fermentation and when the HPLC results seemed off, called the lab manager. “The operators were told to do a standard check, which they had been trained to do, but they grabbed the wrong solution and didn’t look at the values before continuing on,” she recalls. “There were quite a few values completely off. They went into panic mode, thinking the ferms didn’t get the correct ingredients, and started adding things.”

The lesson learned is to make sure things are labeled properly, Stoughtenger says. “And make sure operators are up to speed. You need to go through the SOPs (standard operating procedures) in the lab more than once. Do refreshers every quarter, if you can, because there’s so much going on in the plant and operators do change positions.” She also recommends striving for simplicity in SOPs. “Make sure they aren’t too lengthy, are easy to follow and easily repeatable.” Reliable results require everyone running tests to follow the same procedures.

Best Practices
Andrew Hawkins, director of laboratory services at Phibro Ethanol Performance Group, encourages ethanol plant laboratories to adopt the best practices developed by other technical industries, such as pharmaceuticals, by implementing a metrology program. “Metrology means keeping your instruments in good working order, so you get the best, most-accurate and lowest-error data possible,” he says. “Some plants I’ve been to keep a log book for every instrument, which shows when they’ve calibrated it and when they’ve validated it. And then, there’s other plants that say, ‘I’m supposed to calibrate that?’ about some of the less obvious equipment.”

While the need for validating and calibrating HPLCs and pH probes and meters is commonly understood, Hawkins says all lab instruments need proper attention. “When I started in microbiology, I had to calibrate micropipettes. Some people recognize they’re supposed to be calibrated, or at least validated, every six months or so. Others have them in a drawer, pull them out, use them, and have never done anything.” Micropipettes measure small quantities of liquids, anywhere from 1 or 10 micro liters to 100 or 1,000 micro liters. (An average drop is 50 to 100 micro liters.) “They are designed for a specific volume when you’re making dilutions so you can quantify the chemicals in that solution,” Hawkins explains, adding that the nature of working with corn mash in a laboratory means anything and everything can get fouled by the sticky mash.

“It’s not uncommon to have stuff get dirty,” he says. “You might get rushed. There’s lots of stuff to analyze in the lab. Sometimes days are crazy. Maybe you don’t clean your equipment or maybe someone spills some mash.” A mash spill could easily get into the mechanism of a balance, throwing off the weight. Even moving the instrument on the counter can throw off a highly sensitive balance. Balances are primarily used to weigh samples of corn mash before and after being dried to determine the percentage of solids, and maintaining consistent solids loading is a main control point for an ethanol plant. Hawkins recommends the balance be validated weekly by using a standard set of weights. “Put a 1-gram weight on and verify the balance reads 1 gram, or within the tolerance.”

“Keeping track of how your instruments are working can alert you when there’s an issue,” Hawkins says. “The hardest part is getting into a routine and being disciplined enough to keep doing it.”  The challenge is that many only start investigating their instruments when there’s a problem, he says. “For example, they report some data and make a change in the plant and go, ‘Oh, that went south fast.’ And then they start doing an investigation to figure out why, and they might track it back to some bad lab data and say, ‘I need to validate and calibrate.’”

Phibro is one of several companies that organizes round robins for plants to compare their HPLC results, Hawkins says. “We send out quantified mixes of chemicals that have been made by an independent group and verified. We use that to cross validate instruments. We’ll run it here on our instruments and last year we sent it out to 51 plants. They’ll run it and tell us the numbers.” Everyone sees how their plant’s result compares. While some error and variation always will create some noise in the data, Hawkins says when comparing 20 or 50 plants, the outlier—the instrument that is substantially off—is easily identifiable.

Hawkins cautions about over calibration as well. “If you recalibrate when it’s random error, you can introduce bias and cause your machine to constantly read high or low.” He recommends calibrating HPLCs only as needed. In a recent survey of plants asking about their procedures, Hawkins found the highest response was for calibrating daily or three times a week. The second highest was monthly.  “There were a few folks who said they’ll calibrate when they make changes on the instrument, and that’s actually a pretty good guide,” Hawkins says. “We’ll tell people you should check your HPLC every day. Have a check standard, a mixture of known chemicals in known concentration that you buy, or very carefully make, and make sure that’s what your machine reads. If it’s plus or minus a small amount, that’s normal. But if it’s plus or minus, say 10 percent, and it’s constantly plus or minus that 10 percent, then it’s time to recalibrate.”

Coverage and Credibility
“It’s important to do validation,” says Caleb Ogden, technical services manager at LBDS. “If you can back up your numbers and results with validation and verification, you’ll have a lot better confidence in your results. Most plants are doing this, but lab coverage—having somebody there to get things done properly—is often a challenge.” Maintaining good records over time gives credibility to a lab’s program, he adds. “When somebody comes in to look at your data, they feel more comfortable buying into what you're providing, rather than questioning it. Lately, with margins as tight as they are, it’s not something you can ignore. It definitely is important to keep chemical costs down.”

Good lab results also help a plant’s operations, Stoughtenger says. “The main thing with the lab is you want to make sure it’s running accurately so that fermentation and solids are consistent. It makes your plant run more consistently. Then you can tell if something is off in the plant if you start to see values go off the rails, or your measurements look different. As long as it’s been known the lab equipment is working fine and the upkeep is there, you can focus on pinpointing the plant issues.”

Author: Susanne Retka Schill
Freelance Journalist
[email protected]