Beware the Library Search My Child!

Apologies to Lewis Carroll for butchering the first line of his nonsense poem “Jabberwocky” (for a chemistry version of the poem called ”Beware the Physical Chem” click here: https://www.alphachisigma.org/Page.aspx?pid=536 ). But I do mean what I say in the title, beware of library searching. It is a frequently abused procedure in infrared spectroscopy, and if it is used improperly it can be dangerous.

Back in the day before personal computers (boy I’m giving my age away here) library searching was done by eye. Sadtler Research compiled thousands of paper copies of infrared spectra in green three ring binders, and the user had to flip through them comparing the sample to the reference visually. Thanks to modern computers the comparison job has been automated. Library search programs use algorithms instead of visual comparison to make decisions about match quality. A number, called a Hit Quality Index (HQI) is calculated for each comparison. Then, the best matches are shown in a search report.

I have seen people do a library search, look at the first result in the search report, declare “that is it” and go on their merry way without ever looking at the spectra. This is a recipe for disaster! The library search program will always give you a result, even if it is a bad result. Just because the HQI is a number that comes out of a computer does not sanctify it. Remember, computers are programmed by people, and computer programs make mistakes as easily as people do.

Another pitfall people fall into is over interpreting the HQI. When an HQI of 100 is a perfect match I’ve seen people interpret a 95 as meaning “there is a 95% probably that I identified the sample correctly, or “the samples are 95% the same” or “the spectra are 95% similar”. All of these ideas are wrong. The HQI is not a probability or a percentage, it has no units. The value of the HQI varies with a number of things including the search algorithm and spectral regions used in the search. The HQI simply orders the matches for a given search…that is all!

Another reason to beware the library search is search algorithms. The problem here is that we are trying to automate a visual comparison by substituting a calculation for it. As spectroscopists we know what the peaks mean and what noise and artifacts look like but search algorithms do not. It can happen that two spectra of radically different samples can give a high HQI if by coincidence the noise and artifacts in their spectra are similar. Additionally, the spectra of similar samples can give a low HQI if the noise and artifacts in their spectra are different by chance.

There is one simple solution to the problems of library searching, ALWAYS VISUALLY COMPARE THE SPECTRA! Look at your sample spectrum and the library matches and draw your own conclusions about what is the best match, do not rely on the HQI by itself to make the decision for you. In any competition between your eyeballs and the search algorithm, your eyeballs win. The computer is not smarter than you, it is faster than you. The purpose of a library search is to narrow down the possibilities for you. It is your job as the human being performing the library search to interpret the results to arrive at your own conclusions.

If you follow this advice, you will no longer have to beware the library search.

FTIR for Identity Testing

In many industries there is a need to identify raw materials as they come in the door, and the process is called “performing an identity” or “identity testing”. Because of its molecular fingerprinting abilities, FTIR is well suited for this type of work. Performing an identity normally entails comparing the sample spectrum to the reference spectrum of a known material.

Recently there have been moves afoot, particularly within the pharmaceutical industry, to use near infrared (NIR) spectroscopy or Raman scattering in place of FTIR for identity testing. Any identity testing technique should ideally be fast and easy, specific, and widely accepted in industry. Obtaining NIR spectra can be fast and easy, but this technique fails the specificity test. The features in NIR spectra are broad and diffuse, few in number, and difficult to interpret. Also, in general only functional groups containing O-H, N-H, and C-H bonds are observed in the NIR. These spectra simply lack the specificity we need in an identity testing method.

Recently developed handheld Raman scattering spectrometers (www.ahurascientific.com) make obtaining Raman spectra fast and easy. Raman spectra contain many sharp peaks and give much of the same information as infrared spectra, so they have the specificity needed for identity testing. However, the use of Raman spectra for identity testing is in its infancy, there are few libraries of Raman spectra available, and the technique certainly is not yet widely accepted in industry.

If two molecules have different chemical structures they have different infrared spectra…you can’t get more specific than that. From my own observation FTIRs are already used for identity testing in thousands of companies in dozens of industries, so the technique is widely accepted. In the past the main criticism of FTIR for identity testing was that the sample preparation is not fast and easy. The development of diamond ATRs has solved this problem (see previous blog posts). An FTIR equipped with a diamond ATR accessory can obtain spectra on powders, solids, liquids, and polymers in a matter of seconds. If one desires to take the instrument out of the lab to the loading dock where the sample is, some lab FTIRs such as the Bruker ALPHA that I use (http://www.brukeroptics.com/alpha.html ) are portable enough and rugged enough to be put on a cart and wheeled up to the vessel containing the sample. If one wants to take portability to its extreme, hand held FTIRs do now exist (see previous blog posts and http://www.a2technologies.com/index.html ). FTIR now satisfies all criteria for a perfect identity testing tool. Why would you bother using any other technique?

Debunking FTIR Myths II: “You Can Make a Subtraction Say Anything You Want”

Hi Folks. I’ve been away from blogging while enjoying the holidays but now I am back. I’m going to start off the New Year talking about a pet peeve of mine: FTIR myths. These are totally unfounded pieces of “wisdom” that I hear all too frequently from FTIR users, and worse yet from people who have never touched an FTIR. After looking at the title of this blog post you may be asking yourself, “Where’s Debunking FTIR Myths I?” It’s already been written, but was not given the proper title because at the time I did not realize this was going to become a series. The first entry in this series was my most recent blog post originally entitled “FTIR vs. GC-MS Smackdown.” It is now called “Debunking FTIR Myths I: “FTIR Can’t Identify Things in Mixtures.” Go read it now if you haven’t already.

This post’s myth has to do with spectral subtraction. Spectral subtraction makes use of specialized software to simplify mixture spectra and make mixture analysis easier. The attached figure illustrates the utility of spectral subtraction. The bottom spectrum in red is of the amino acid glutamine dissolved in water, and the middle spectrum in blue is of pure liquid water. The glutamine peaks in the bottom spectrum are very small and masked in part by the strong, broad absorbances of liquid water. It would be hard to tell from the bottom spectrum whether there was anything dissolved in the water at all. Using a spectral subtraction program I subtracted the spectrum of pure liquid water, known as the reference spectrum, from the sample spectrum. The subtraction result, in green, is at the top of the figure. The water peaks have now either been removed or greatly reduced in size, simplifying the mixture spectrum. The glutamine peaks are now apparent for all to see. It would be difficult or impossible to identify the glutamine in this sample without the use of subtraction.

The criticisms I hear of subtraction are that its “arbitrary”, that you are “adding something” to the spectrum when subtracting or that you can “make the result say anything you want.” This is all baloney. Spectral subtraction is like anything else in life, if it is USED PROPERLY it has great utility and legitimacy of purpose. The “arbitrary” critique probably stems from the fact that subtraction involves user interaction. Most of the time when two spectra are being subtracted from each other their absorbances are different because of differences in pathlength or concentration between the two samples. To compensate for this the absorbances of the reference spectrum are multiplied times a number called the “subtraction factor” or “scale factor” which adjusts the reference spectrum absorbances so they match those of the sample. When the subtraction is performed the reference spectrum absorbances subtract out. There is nothing arbitrary about adjusting the subtraction factor; there are rules and procedures to follow that are clearly described in books and training courses on FTIR (including my own). If done properly, there is nothing arbitrary in the setting of the subtraction factor.

The “adding something” myth is the most ridiculous. Subtraction and addition are completely opposite mathematical operations. You can’t “add” something to a spectrum by subtracting something from it; nothing is added during a spectral subtraction. The “you can make a subtraction say anything you want” criticism is the most damning, and the most wrong. If performed properly a spectral subtraction will allow you to more clearly see that which was already present in the sample. In the attached example by following proper procedures I was able to decide upon the best subtraction factor (0.96 in this case), and the result is that the glutamine peaks that were difficult to see before are now obvious. I didn’t make the result say anything I wanted, the result simply shows me more clearly that which was already present in the spectrum.

In the hands of a properly trained user spectral subtraction is a legitimate tool to simplify mixture spectra and make mixture analysis easier. Books and training courses to help you to perform subtractions properly exist (www.spectros1.com), so there is no excuse for not using subtraction in your work, and there is certainly no excuse for badmouthing a useful spectroscopic technique.

Debunking FTIR Myths I: "FTIR Can't Identify Things in Mixtures"

I am not normally one to trash other chemical analysis techniques. But when people mouth untruths about FTIR it raises my dander and I feel compelled to respond. I have talked to several people lately who have said, “FTIR can’t be used to identify things in mixtures. It’s easier to shoot the sample into the GC-MS instead.” I am fully aware, and I always teach in my courses, that FTIR has difficulty with mixtures. However, as I also always teach, there are strategies for getting around the mixture analysis problem, some of which I have talked about in previous blog posts. These techniques include purifying the sample, mixture analysis software, subtraction, library searching, and the “process of elimination”. These techniques make analyzing mixtures doable, and frequently some of the components in a mixture can be identified from the infrared spectrum of a sample as a result. I have identified components in mixtures hundreds of times during my career using these tools. FTIR can be used to identify things in mixtures if the analysis if performed properly.

GC-MS stands for Gas Chromatography-Mass Spectrometry. In this technique the GC purifies mixtures into their components, and then the mass spectrometer identifies each component. Thus, it appears GC-MS does not suffer from a mixture problem like FTIR. Additionally, GC-MS is more sensitive than FTIR and does a better job of handling aqueous solutions. I have in fairness sung the praises of GC-MS.

Here is why the phrase “It’s easier to shoot the sample through the GC-MS” raises my ire. A GC-MS is always more expensive, more time consuming, and more difficult to use than an FTIR. This is why so many more labs have an FTIR than a GC-MS. Also, GC-MS does suffer from a mixture problem. The retention time of a peak in a gas chromatograph is not unique to that molecule. Molecules with different chemical structures can, by coincidence, elute at the same time. Thus, there is always the possibility that what we think is a pure component coming off a GC column may in fact contain two or more components. The other dirty little secret about a mass spectrum is that it does not provide a completely unique fingerprint of a molecule. Mass spectrometry does not distinguish between structural isomers. These are molecules that have the same chemical formula but different chemical structures. For example, a mass spectrometer normally can’t distinguish ortho-, meta-, and para- xylene from each other. These three molecules are easy to distinguish by FTIR. Infrared spectroscopy can distinguish between structural isomers, and provides a unique fingerprint for a given molecule. A mixture of structural isomers would be difficult to distinguish by MS. It could thus be argued that you can’t identify things in mixtures by GC-MS either.

There is really is no victor in the FTIR/GC-MS smackdown except hopefully the truth. Both techniques have their strengths and weaknesses, and you need to have an understanding of both of these to pick the right technique for the right sample. Since FTIR is faster, simpler, and cheaper than GC-MS I believe it should the first choice for analyzing unknown mixtures.

A Certification Program for FTIR Users

FTIRs have been used for decades to perform chemical analyses including in such critical applications as the analysis of drugs, criminal evidence, and in homeland defense. The user of an FTIR must prepare the sample appropriately, choose the correct scanning parameters, and interpret the data correctly to achieve a successful analysis. In short, a properly trained human being is key to achieving accurate FTIR results. However, there are no formal proficiency guidelines or certification programs that insure the competency of FTIR users and hence the quality of FTIR results…until now.

Starting in 2010 my company, Spectros Associates, will be offering a Certified FTIR User(CFU) program. To earn the CFU certificate FTIR users will be required to take a 5 day sequence of FTIR training courses consisting of my Fundamentals of FTIR, Hands-On FTIR Sample Preparation, and Infrared Spectral Interpretation I, II and III courses. Then, you must pass a written exam consisting of multiple choice questions and several unknown spectra that must be interpreted. Only upon completing the coursework and successfully passing the exam will the CFU certificate be awarded.

The advantage of this program to FTIR users is that it will be your proof to the world that you know how to do your job well. The knowledge you obtain in earning the CFU will make your FTIR analyses faster, better, and cheaper. This will make you more valuable to your employer and enhance your career. The certificate can be held up at review time as an accomplishment, perhaps to be rewarded with a higher salary. Your certification can be put on your resume as proof to potential employers that you possess an important skill set, and might make you more employable than those who are not certified. The CFU can also be used towards the continuing education requirement that some professional societies require of their members.

There are also advantages to employers of the CFU program. By certifying their workers they can rest assured they are getting the best FTIR analyses possible. If regulatory bodies ever question the quality of spectra, employers can point out that the data were measured by certified workers following best practices. Finally, employers hiring FTIR workers can rest assured that job applicants who have earned the CFU possess the skills needed to do the job well.

Please let me know what you think of thie Certified FTIR User idea by leaving a post below or e-mailing me at bcsmith@spectros1.com .

Infrared Spectroscopy in Outer Space!

Last month NASA announced the presence of a thin layer of water and hydroxyl ions on the surface of the moon. Not one, not two, but three different satellites recorded data to support this conclusion. And guess what technology the three satellites used to detect the water? Infrared Spectroscopy of course! As many of you know from looking at the spectra of liquid water or water vapor, this molecule absorbs strongly around 3500 cm-1, or since astronomers like to use wavelength, around 3 microns. The first evidence that there was water on the moon came from the analysis of the moon rocks brought back by the Apollo missions back in the 1970s (it was watching these missions as a kid that inspired me to become a scientist). Water was found in the rocks, but it was assumed that they had been contaminated with terrestrial moisture.

In 1999 NASA’s Cassini satellite flew past the moon on its way to Saturn and detected water as well. However, at the time the data were thought to be erroneous and were not published until this year. More recently, the Moon Mineralogy Mapper aboard the Indian Space Agencies’ Chandrayaan-1 spacecraft found evidence for water on the moon. In June of this year NASA’s Epoxi spacecraft, flying past the moon on its way to a comet encounter in 2010, also detected water on the moon. For all three of these spacecraft the spectrum of the moon was measured in infrared reflectance. A dispersive grating spectrometer was used rather than an FTIR, perhaps because FTIRs have moving parts and are hard to fix billions of miles from home. The spectra are measured in single beam mode since there is not an identical twin moon available devoid of water to provide a background spectrum. However, the spectra from equatorial regions, where it is known to be drier, can be subtracted from the spectra of polar regions to hopefully remove any contribution from the moon’s surface other than OH and water. Examples of the types of spectra measured are seen above.

Now, it’s not time to pack your swim trunks and head to the moon for a dip. The water and hydroxyl are estimated to be present in concentrations of about 1000 ppm. You would have to wring dry a ton of lunar soil to obtain a quart of lunar water. The water is thought to form when energetic protons in the solar wind collide with oxygen bearing minerals on the moon’s surface. More details at these NASA websites.
(cut and paste these URLs into your browser...blogger won't let me post links!).

http://science.nasa.gov/headlines/y2009/24sep_moonwater.htm

http://www.nasa.gov/topics/moonmars/features/moonm3-images.html

Abrasive Sampling: Spectra of Intractable Samples

There is a sampling technique used on powders and solids called Diffuse Reflectance Infrared Fourier Transform Spectroscopy, or DRIFTS for short. I think some graduate student spent several months thinking up that acronym. At any rate, the technique involves using a special sampling accessory that fits into your FTIR sample compartment to bounce the light off the sample. The light is then collected and sent to the IR detector (more on DRIFTS in a later blog post). Abrasive sampling is an interesting application of DRIFTS. Some FTIR accessory manufacturers make flat metal posts that fit into their DRIFTS accessories and SiC disks with an adhesive backing. The SiC disk is adhered to the metal post, and then the disk is rubbed against the sample to abrade off sample particles. The SiC disk with particles is then placed at the focal point of a DRIFTS accessory, the light is reflected off of the sample particles, and is collected and sent to the detector to obtain the sample spectrum. The background spectrum is run on a clean SiC disk.

The attached spectrum is of the white paint on a light fixture obtained by abrasive sampling using an Alpha spectrometer from Bruker Optics (http://www.brukeroptics.com/). This spectrum would have been difficult to obtain any other way. The light fixture was firmly attached to the ceiling and so could not be taken down. In theory, one could scrape a lot of the paint off and make a KBr pellet or cast a film of the paint. However, this would damage the light fixture, involve time consuming trial and error, and may still not work. With abrasive sampling, only a small hidden part of the light fixture was scratched, and the entire measurement process took about 2 minutes. Abrasive sampling is useful for spectra of large, intractable objects such as furniture, large pieces of plastic, or anything that is simply too big to be analyzed by normal FTIR sampling techniques. The beauty of abrasive sampling is that it is fast and easy. However, the SiC scatters the IR beam a lot, so abrasive sampling spectra can be noisy. This can sometimes be dealt with by increasing the number of scans, perhaps to as many as 256.

A note on the SiC disks. I have seen labs try to save a little money by going to the hardware store, buying SiC paper, and then using a cork borer to punch out SiC disks of the proper size. As long as these disks fit into the sample cup that came with your DRIFTS accessory this should at least in theory allow you to obtain abrasive sampling spectra.

Use your imagination…what type of applications might abrasive sampling have at your company?