****************************************************************************** From: "Little, James L " Date: Fri, 26 Mar 1999 08:11:15 -0500 Subject: Accurate Mass Data Organization: * I wanted to share some accurate mass data with the mass spectrometry community. I wanted to know what types of precision and accuracy other labs obtained. We normally obtain about 5 ppm error for GC/MS (one std. dev.) and about 2 ppm on solid probe (slower scanning, average more scans). This seems reasonable in comparison to several thousand data points I obtained from the expermental section of the Journal of Organic Chemistry which showed an error of about 5 ppm. I assume most of these values are solid probe with peak matching, but the authors gave few details. My data points and the literature data points can be found at: http://users.intermediatn.net/slittle/ James L. Little Bldg 150 Eastman Chemical Company Kingsport, TN 37662-5150 Tel 423-229-8685 FAX 423-229-4558 Sailing Club Web Page: http://www-research.kpt.emn.com/sailingclub/ ****************************************************************************** From: Ken Matuszak Date: Fri, 26 Mar 1999 11:51:39 -0600 Subject: Re: Accurate Mass Data Organization: Abbott Labs Pharmaceutical Products Division See our posters from the last two years at ASMS. 1997 - Rich Burton - on characterizing FTMS (mostly ESI and some MALDI) for exact mass measurements 1998 - Paul West - on characterizing our JEOL (FAB and EI) for Exact mass measurements Typically their results are better than 2 or 3 ppm (1 to 2 ppm in MS/MS), but this has been without chromatography Rich is now working on LC/MS and LC/MS/MS using the FTMS for exact mass measurements. He will have a poster this year at ASMS, but it will be more of an application study on metabolites rather that a more formal study on exact mass measurements. Early results indicate 2 to 3 ppm for the LC/MS and LC/MS/MS. Ken Matuszak Group Leader Analytical Mass Spectrometry Abbott Labs. ****************************************************************************** From: "steve" Date: Sun, 28 Mar 1999 09:34:44 +1000 Subject: Re: Accurate Mass Data Organization: Telstra Big Pond Cable - Sydney Site 1 Why bother with the expense and complications of FTMS when there are commercially available LC-TOF (Micromass LCT) instruments and LC-QUAD-TOF (Micromass QTOF) instruments that will give accurate mass information BUT with much higher throughput and very routinely (including chromatograhy and Automated analysis of Combi Chem libraries). Micromass have now released there GC-TOF product which will also give accurate mass determinations from an instrument not much larger than the GC and certainly a lot less expensive than a sector-MS and less complicated to boot. Steve ****************************************************************************** From: Ken Matuszak Date: Mon, 29 Mar 1999 06:26:02 -0600 Subject: Re: Accurate Mass Data Organization: Abbott Labs Pharmaceutical Products Division While the throughput of the ESI-TOF products (Micromass and Perseptive) is wonderful, they typically give measurements of within 5 ppm, only about 60% of the time. Given this fact, if you are doing combi-chem type batches and are trying to determine whether each sample meets a 5 ppm criteria, you may get different passes and fails if you run the same rack more than once. This will be especially true if your samples aren't very pure. I have runs lots of combi-chem focussed libraries. Few of the samples are very pure. I suppose, if you were running med-chem type samples to get a 5 ppm accurate mass for publication purposes, you could run the sample many times until you convinced yourself you were within the 5 ppm criteria. However, this practice to me would seem a little questionable. We however, often have to run unknowns. Even at 5 ppm, there are often too many possible formulas that fit. Remember, ESI-TOF instruments measure their res at FWHM (as do FTMS instruments). A sector instrument typically measures the res at 10% of the peak height. A good rule of thumb is that you need at least twice the res on a TOF or FTMS to get the same effective res as a sector (for example, we measure exact masses on our sector at 12,000 res. and we measure exact masses on our FTMS at 20,000-30,000 res). Have you looked at the peak shape on a TOF? The peaks often broaden greatly at the bottom. This is not too bad of a problem when all the peaks are of the same height. However, in practice, this is rarely the case (even when you put in the same amount of sample). If your peak of interest is close in mass to another peak which is of much greater intensity, there is a large chance that the measurement of the mass of the target compound will be in error. There was a great discussion of this topic at last year's FACSS meeting (Austin, TX) in the session that Sue Weintraub organized. We had representatives from Micromass and Perseptive, as well as combi-chem and pharm companies. I anxiously await ESI-TOF instruments that have resolutions of 10,000-20,000. Current ESI-TOF instruments measure at about 5000 res. I love automation. I think these instruments have a great future. They just aren't there yet. We have a Micromass QTOF, but DO NOT USE IT for exact mass work. We love the instrument for LC/MS and MS/MS, but confirm all our mass measurements of significance on the FTMS. This has proven to be extremely important. It is easy to fool yourself into thinking something is good science just because it is easier. In our industry we need to be a bit more rigorous. Ken Matuszak Abbott Labs Group Leader - Analytical Mass Spectrometry ****************************************************************************** From: kabwbb@aol.com (KABWBB) Date: 29 Mar 1999 12:55:32 GMT Subject: Re: Accurate Mass Data Organization: AOL http://www.aol.com Ken, Nice Job of educating. I believe until resolution improvements in future generations of ToF instruments, traditional high-resolution instruments will have a place in most labs. Brad Barrett ****************************************************************************** ****************************************************************************** From: clee751159@aol.com (CLee751159) Date: 29 Mar 1999 20:28:11 GMT Subject: Re: Accurate Mass Data Organization: AOL, http://www.aol.fr It's all a matter of price; TOFs give less resolution than magnetic machines from the big 2, and they have other operational difficulties, particularly in tuning for m/z less than about 300 and because of detector dead time, which shifts peaks on the mass axis. However they are half the price and don't have the same (false) aura of complexity. Having said that, the TOF is quite a revolution in this area, and nothing should be taken away from the achievement. Micromass launched a very interesting core-less medium-resolution magnetic machine in the 1980's, but they told me they had to abandon it because of lack of interest. At that time, people were not used to having accurate mass "on tap", whereas self-service nmr was common. I still think there is a market for small-ish magnetic machines. The German firm AMD make them , but (perhaps for the above reasons) havn't yet made it into the big time. I hope this helps to put the subject into perspective. ****************************************************************************** From: "Little, James L " Date: Tue, 30 Mar 1999 07:55:12 -0500 Subject: Accurate Mass Data Organization: * I remember the ASMS statement that talks about the acceptance of accurate mass data. If you really wanted to hit this target legitimately, I believe you would have to keep a control chart of the performance of the instrument. The data should then be reported with the standard deviation and an idea of the number of data points used to arrive at the standard deviation measurement. We don't generally have too many requests for publication, but it would be easy to run it several times or just pick the "magic" scan off the probe analysis that met the requirement for the journal's accuracy. However, I think that would probably be missing the point. If I was reading the journal article, I would be more interested in knowing the standard deviation of the measurement. If I had another proposed structure, I would know if it was reasonable within the error of the measurement. I currently have an ariticle in press in the Journal of Chromatography that uses accurate mass for the ID of unknowns. I list the ppm error in our labs measurement and the number of measurements used to obtain the error. The values were taken over a 10 year period. I am really interested in the error of my measurement because I use it in identifying unknowns where no structure is proposed. You need it to know how far to open the window for the accurate mass window for searching possible formula. I usually go at least 15 ppm with our window. Our data and some journal data is listed at: http://users.intermediatn.net/slittle/ James L. Little Bldg 150 Eastman Chemical Company Kingsport, TN 37662-5150 Tel 423-229-8685 FAX 423-229-4558 "A Little Mass Spectrometry and Sailing": http://users.intermediatn.net/slittle Sailing Club "Intranet" Web Page: http://www-research.kpt.emn.com/sailingclub/ ****************************************************************************** From: Ken Matuszak Date: Tue, 30 Mar 1999 07:40:49 -0600 Subject: Re: Accurate Mass Data Organization: Abbott Labs Pharmaceutical Products Division CLee751159 wrote: } }Ken, } } } }Nice Job of educating. I believe until resolution improvements in future } }generations of ToF instruments, traditional high-resolution instruments will } }have a place in most labs. } } } }Brad Barrett } } } } } } It's all a matter of price; TOFs give less resolution than magnetic } machines from the big 2, and they have other operational difficulties, } particularly in tuning for m/z less than about 300 and because of detector } dead time, which shifts peaks on the mass axis. However they are half the } price and don't have the same (false) aura of complexity. } } Having said that, the TOF is quite a revolution in this area, and nothing } should be taken away from the achievement. } } Micromass launched a very interesting core-less medium-resolution magnetic } machine in the 1980's, but they told me they had to abandon it because of } lack of interest. At that time, people were not used to having accurate } mass "on tap", whereas self-service nmr was common. } } I still think there is a market for small-ish magnetic machines. The } German firm AMD make them , but (perhaps for the above reasons) havn't yet } made it into the big time. } } I hope this helps to put the subject into perspective. No, I disagree, it is not just a matter of price. It is a matter of how accurate and precise your measurements need to be for a given application. The price is merely what you have to pay to achieve that goal. If the errors given to you by a given technology range from 0-20ppm (not uncommon for an ESI-TOF) and you need to have them all measured accurately and precisely to withing 5 ppm, then it doesn't matter even if the instrument is FREE. The instrument can not do the job. One must define the application first. In my previous posts I laid out three applications and indicated the level of acceptablity, at least from my point of view. 1. Combi-chem racks where a 5 ppm cutoff gives you a yes/no confirmation of the product. 2. Med-chem samples for publication (requires 5ppm) 3. Unknowns 5 ppm, 60% of the time, will give you different yes/no answers if you run Combi-chem racks multiple times. You might as well do unit mass resolution measurements on a benchtop quad and save yourself the money and the headaches For med-chem type samples, I said you could run the sample multiple times until you convinced yourself it was the correct mass, although I would find this a bit questionable as a reviewer. For unknowns, this variability would be totally unacceptable. Sorry, even though I have been in mass spec for a number of years now, doing everything from instrument demos and product marketing on the vendor side to sample analysis and now group management in a pharm lab, I still revert back to my training as an Analytical Chemist when it comes to questions like this. Like I said before, I think (hope) that the TOF instruments have a great future. I will be the first to implement them when they get where they need to be for accurate mass work. In the meantime, don't let yourself be fooled into thinking they can do something that they really can't - just because they are relatively inexpensive. Sincerely, Ken Matuszak Abbott Labs Group Leader - Analytical Mass Spectrometry ****************************************************************************** From: cody@jeol.com (Chip Cody) Subject: Re: Accurate Mass Data Date: Tue, 30 Mar 1999 17:14:42 GMT Organization: JEOL USA, Inc. In sci.techniques.mass-spec you write: }I still think there is a market for small-ish magnetic machines. Thank you!!! I am always of the opinion that no mass analyzer is perfect for all applications. There is a place and application for each type of mass spectrometer. That includes magnetic sectors as well as time-of-flight, quads, traps, and FTMS. As previous respondents to this thread have pointed out, magnetic sectors use a different resolution definition than TOF systems. You have to compare "apples to apples". Furthermore, they have excellent reproducibility and dynamic range, making them a good choice for quantitative analysis. The chief disadvantages in the past have been size, price, and ease-of-use. We (JEOL) make benchtop double-focusing magnetic sector mass spectrometers for both GC/MS (the GCmate) and LC/MS (LCmate). We also sell a benchtop ESI/TOF system, the Jaguar. We see these technologies are complementary, not competitive. All of our magnetic sector mass spectrometers are fully automated (setup and data acquisition) and they are many orders of magnitude easier to operate than older-generation magnetic sector mass spectrometers. All have linked-scan MS/MS options. We even have an open access system for both benchtop machines. While not by any means the lowest-cost benchtop quad and trap mass spectrometers, both are much less expensive than the larger sector systems, and apparently less expensive than the (some of the) new TOF systems are being sold for. Our benchtop sector systems are not a panacea. The resolution is still much less than a full-size magnetic sector. The scan speed is fast (70 ms/decade maximum scan speed for GCmate), but it is no comparison to the data acquisition speed of a TOF system. MS/MS is reproducible, rich in information and easy to set up, but one cannot achieve the same precursor selectivity as a hybrid or triple quad. However, one should match the tool to the problem. Like all MS technologies, magnetic sector technology has continued to evolve and improve over the years. Today's sectors should not be judged based on instruments from 10 years ago, any more than one should judge today's TOF spectrometers based on the performance of a 10-year-old system. One should evaluate all reasonable technologies based on their merits and their suitability to the problem to be solved. -- "For purposes of ... New Jersey Right to Know Act. Contents partially unknown." |____________ |_ Robert B. Cody, Ph.D |________________________________ Product Development Manager |__ Mass Spectrometry |________________________ JEOL USA, Inc. |_ http://www.jeol.com |__________ e-mail: cody@nojunkmail.jeol.com |_ (REMOVE 'nojunkmail' TO REPLY) ============== Do not send me spam or advertising via e-mail !! ========= ****************************************************************************** From: clee751159@aol.com (CLee751159) Date: 11 Apr 1999 15:26:07 GMT Subject: Re: Accurate Mass Data Organization: AOL, http://www.aol.fr }No, I disagree, it is not just a matter of price. It is a matter of how }accurate and precise your measurements need to be for a given application. }The price is merely what you have to pay to achieve that goal. We are interested in establishing impurity profiles, and when we find a new impurity, we have to identify it, preferably without going to the trouble of isolating it for nmr. Identification work therefore represents only part of our LC/MS usage, which means that it's quite acceptable to spend some time messing around to make sure we have got the right molecular formula. I'm sure many other users are in a similar situation. In the lab where I work we have several quads and one TOF, which together cost hardly more than a single magnetic instrument from the big 2. I mentioned magnetic instruments in my original posting because from experience I do find them more useful than TOFs; they were ruled out of the last purchasing round only on the grounds of cost effectiveness in a given application. Regards ****************************************************************************** From: clee751159@aol.com (CLee751159) Date: 11 Apr 1999 15:38:07 GMT Subject: Re: Accurate Mass Data Organization: AOL, http://www.aol.fr }I remember the ASMS statement that talks about the acceptance of accurate }mass data. If you really wanted to hit this target legitimately, I believe }you would have to keep a control chart of the performance of the instrument. }The data should then be reported with the standard deviation and an idea of }the number of data points used to arrive at the standard deviation }measurement. } }We don't generally have too many requests for publication, but it would be }easy to run it several times or just pick the "magic" scan off the probe }analysis that met the requirement for the journal's accuracy. However, I }think that would probably be missing the point. If I was reading the }journal article, I would be more interested in knowing the standard }deviation of the measurement. If I had another proposed structure, I would }know if it was reasonable within the error of the measurement. } }I currently have an ariticle in press in the Journal of Chromatography that }uses accurate mass for the ID of unknowns. I list the ppm error in our labs }measurement and the number of measurements used to obtain the error. The }values were taken over a 10 year period. } }I am really interested in the error of my measurement because I use it in }identifying unknowns where no structure is proposed. You need it to know }how far to open the window for the accurate mass window for searching }possible formula. I usually go at least 15 ppm with our window. } }Our data and some journal data is listed at: } }http://users.intermediatn.net/slittle/ } }James L. Little }Bldg 150 }Eastman Chemical Company }Kingsport, TN 37662-5150 }Tel 423-229-8685 }FAX 423-229-4558 } }"A Little Mass Spectrometry and Sailing": }http://users.intermediatn.net/slittle }Sailing Club "Intranet" Web Page: }http://www-research.kpt.emn.com/sailingclub/ } } I suggest it would be very hard to estimate the error in LC/MS applications, because (at least at moderate resolutions) there is know way of knowing if the centroids of the sample and reference mass peaks are being shifted by isobaric interferences. Statistics won't help here. Fortunately, this doesn't matter to many of us, because the structure of the unknown in question (an impurity in a drug substance, for example) is generally established later on by other methods and nobody goes to jail if you get it wrong :-) However, I fully agree with you as far as proof of structure or identity solely on the basis of LC/MS data is concerned. Regards