Back To: Home



What is analytical chemistry?
December 2010
by Peter T. Kissinger  |  Email the author

We analytical chemists continue to have a variety of concerns about what we do, and those concerns vary with the context of where we do it. The argument stems from confusion over the worthiness of providing a service and whether that can ever fit our definition of science. Our profession is not alone in this regard. Physicians, lawyers, accountants and investment bankers provide services for which they are well-compensated and held in high regard (in their own minds) by that pecuniary measure of service quality.  
On the other hand, the last three professions are also the objects of frequent derision. There are many lawyer jokes, but few analytical chemistry jokes. The compensation in a field is proportional to the number of jokes devoted to it. Please create some for analytical chemistry! 
The professorial class at medical, law and business schools is not particularly concerned that the professions underlying their teaching provide essential services to society at large. After all, people who wear suits deserve employment so that builders of sailboats and golf courses also have employment.  
Definitions of analytical chemistry vary. "Analytical chemistry is what analytical chemists do" is commonly attributed to my mentor, the late Prof. Charles N. Reilley of the University of North Carolina (UNC) at Chapel Hill. "Analytical chemistry is the strategy and tactics of chemical measurements" is attributed to Prof. Larry R. Faulkner, who rose to leadership positions at the University of Illinois and later led the University of Texas.  
My own definition has centered on "getting quality data about chemical systems to support quality decisions," with the implication that analytical chemists can distinguish data sufficient for the purpose or not. Don't all chemists (scientists) do this? No, they do not. Too many simply accept what is delivered by an instrument print-out or what is sent to them by an in-house or contract research laboratory or academic core laboratory. These are the days of digital displays and outputs on spreadsheets. These are the days where instruments are presumed to be calibrated, but are often not shown to be calibrated. 
The very best analytical chemists are attentive to the problem addressed, the sample considered, and the basic science of the measurement, the quality of the data and the appropriateness of the conclusion drawn from it. These people are too rare. Data is regularly handed off to others who lack appropriate skepticism about what it means. The only perfect numbers are those from counting small numbers of things. Everything measured has a decimal point.
Let's combine the three definitions. Analytical chemistry is what analytical chemists do with strategies and tactics of chemical measurements to get quality data to support quality decisions. But we must dig deeper. Strategy involves grasping the underlying principles of metrology and developing entirely new ones as needed. Implied is the choice of an approach, including a tool that fits the requirements of the conclusion to be made.  
Tactics involves sampling, calibration and validation to the purpose. Quality data implies validation and understanding the selectivity, accuracy and precision of the information to be reported. The underlying basis of a measurement is a critical component of analytical chemistry. Academic analytical chemists and scientists in the very best instrument companies want to know how things work and what limitations result.  
I recently attended talks at Purdue by Profs. Jim Jorgenson (UNC-Chapel Hill) on separations and Gary Hieftje (Indiana University) on spectroscopy in electrical discharges. Both speakers represented the very best traditions of analytical chemistry—understanding the basic science behind the signal responsible for the data reported. Both have had a large impact on their respective fields. You know it when you see it and it feels good. Analytical chemists in this category enlighten us by deepening our understanding of nature. That is what I call science, and it is not to be confused with an industrial-quality control lab. Both are no doubt important, both suggest challenging problems to be solved. One is not better than the other, but the differences should be clear.
Nearly every experimental scientist and engineer focused on biology, environment, pharmacy, food, agriculture, medicine, forensics, synthetic chemistry, materials and energy does analytical chemistry. It's that important! Some suggest that because it is so popular, it must be both easy and unimportant. It is neither. It is very hard to get good numbers. Just try to get them twice. Too few of us seriously consider how an error is propagated through a measurement scheme.
I have my favorites. I've seen five digits reported for a method calibrated with a standard reported to be 95 percent pure, more or less. I see pH meters that read to three digits past the decimal point, when the calibrators vary plenty in the second decimal place and were received in the lab three years ago. I see data reported for chromatographic methods far outside the range of calibrators establishing the slope. I've seen glucose meters reporting a few extra digits when the accuracy on a good day is no better than 15 percent over a rather limited range.   
Let's face it, we really don't yet understand all the complications of what goes on in the source of a mass spectrometer for a complex mixture, when sample-to-sample, those things we are not determining vary in concentration by orders of magnitude in ways we cannot know. Why pretend that we do understand?  
Suspend belief in instrument outputs now and then. Be curious about tables of numbers. Don't trust. Verify. Go out there and get good numbers that we all can trust before we make a decision about how the universe works, if Mr. Jones should go to jail or if the latest cancer drug should be approved.
Do we think about how good the numbers really need to be to support a decision? While 7.385 ng seems better than 7.4 ng or 7 ng, would one really be preferred over the other? Given a standard deviation of +/-1 ng, the last of these pleases me most. The first looks better on a PowerPoint slide. That's fashion, not science. That's adding noise, not enlightenment.  
Peter T. Kissinger is chairman emeritus of BASi, chairman of Prosolia in Indianapolis and a professor of chemistry at Purdue University.



1000 N West Street, Suite 1200,
Wilmington, Delaware, 19801
Ph: 888-781-0328 |  Fax: 705-528-0270
© Copyright 2020 Old River Publications LLC. All righs reserved.  |  Web site managed and designed by OffWhite.