Dear Joe,

If your GC-MS analysis was ran with a temperature program (a ramp, for example a 5 degrees per min rate) then you should calculate Van den Dool and Kratz indices. If the program was isothermal, your choice are the Kovats indices. However, if you ran a complex program with several ramps and isothermal periods it would be very difficult to calculate any sort of RI.

Let us assume you had a single ramp (or a ramp with an isothermal period in the end and your compound eluted in the ramp part). First, find two consecutive alkanes that are closest (in retention time) to the peak of the compound you want to calculate RI. For example, these n-alkanes have n and n+1 carbon atoms. The difference between the retention times of the two alkanes in RI units amounts to 100, by definition. Next, you calculate the difference of retention times between your peak of interest and the alkane with n carbons (you could use the other one as well but should be careful with the next step). This value (I will designate with x) should be divided with the retention time difference for the two alkanes and multiplied by 100. The retention index for your compound is then RI = 100*n + x.

This is really simple, and perhaps it would be best if you made an Excel spreadsheet with such a function so that you only supply the retention times of the nearest alkanes and you analyte of interest.

Please, be free to ask if you need any further help

All the best,

Niko

Dear Niko,

Thank you for your clear and comprehensive response, it is much appreciated. As for the settings I used to run my samples, I started at 60C and ramped to 300C at a rate of of 10C/min so I believe I will be looking to calculate Van den Dool and Kratz indices. 

The issue I have is that I actually had to send my samples away to another lab for analysis as I didn't have the facilities. As such, they provided me with an n-alkane run under the same conditions as my sample. Now you detail that I should use the two closest n-alkanes to my peak of interest, but the lab running the samples used 9, 10, 11, 12, 14, 16, 18, 20, 22, 24, 26,28, 30, 32 and 34 carbons. So I am missing the retention times of the odd numbers of carbons. How would I proceed to calculate the retention indices using this data? I'm guessing the difference between the n-alkanes I have is 200 in RI units?

Best wishes,

Joe

Do a regression with your data points RI units vs RT and use it to calculate the RI of your compound of interest. 

 Check the linearity of your regression. It would be preferable to have an r2>0.99.

Chapter 3 (Identification based on retention) of this book deals with the subject very thoroughly: Identification techniques in gas chromatography. NY: Wiley Interscience. DA Leathard, BC Shurlock - 1970 - Wiley and Sons Ltd

Some stationary phases for packed columns were reproducible between columns; I don't know if the bonding and cross-linking of capillary phases have a significant influence in that respect.

The book is still advertised on the web and I have a copy. Does someone know the rules about sending a scan of a single chapter to a RG member?

Hello Joseph!

I recommend you to use the Van den Dool and Kratz indices, you should have a look at the IUPAC goldbook regarding this topic (http://goldbook.iupac.org/R05360.html). As you only have some n-alkanes (it would be better to have them all), I suggest that you use what you have, use the last formula of the link I've shown here. It should work with a bit of error, but it should work. Good luck!

If you got all the raw data from your external lab, you might save a lot of time if you look into a free software called AMDIS. It should be part of your NIST library and might already be installed on your PC. This software can combine identification by comparing spectra with applying penalties to the matching score for not suitable retention indices.

Retention Index based identification of a compound is more authentic than Retention Time based identification. This is because, retention time is dependent on analysis conditions, whereas retention index is independent of the conditions and will remain same on a given stationary phase. For the determination of Retention Index, you require a mixture of n-alkanes ( saturated n-paraffins) from C7 to C37. These alkanes can be separated by a suitable capillary column. By convention, these alkanes willhave well defined RIs which is nothing but the carbon number multiplied by 100. Depending on the elution of your compound  ( between which alkanes) the RI will be automatically calculated by the s/w. For example, if your compound elutes between C21 and C22, then the compound's RI will be between 2100 and 2200. Shimadzu offers a facility for it's users to do a library search based on Retention Index in their GCS system which enhances the probability of identification considerably.

For an example of how to do this Retention Index/Kovats Index and make it understandable to readers, this publication has been popular for a long time.  

I analyzed hydrocarbons extracted from tsetse fiies that had been pinned in cabinets at the London Museum of Natural History for 100+ years. There were no computer programs at the time, so it was done by comparison with n-alkanes of both even and odd carbons that eluted before and after the sample peaks. Measurements were made of the many peaks eluting between n-alkane standards. The Carlson et al. paper shows how methyl-branching position changes retention times of compounds with the same molecular weight, and how to interpret them.

The neat trick with natural products is that tsetse flies, for example, have "families" of compounds with the same methyl branch positions, as the chain length increases from 25 to 39 carbons.

Many mass spec id programs will give total rubbish for structures of long-chain hydrocarbons because the programs ignore KI values.  

From ResearchGate below- also another Katritsky et al. in my RG resume.

Article: Elution Patterns from Capillary GC for Methyl-Branched Alkanes

David A. Carlson, Ulrich R. Bernier, Bruce D. Sutton

[Show abstract]

Journal of Chemical Ecology 10/1998; 24(11):1845-1865. DOI:10.1023/A:1022311701355 · 2.24 Impact Factor

Hi David,

Thanks for your informative response - it has cleared a few points up for me. However, I feel that one of the most confusing things I am finding whilst trying to interpret my data is the difference between Kovats RIs and Linear RIs. I understand that Kovats are calculated using isothermal GC settings, but it is quite difficult to find Linear RI values in the literature. How do you tackle this problem?

Best wishes,

Joe

Also from ResearchGate in D.A Carlson publications.

Article: QSPR correlation and predictions of GC retention indexes for methyl-branched hydrocarbons produced by insects

Alan R. Katritzky, Ke Chen, Uko Maran, David A. Carlson

[Show abstract]

Analytical Chemistry 02/2000; 72(1):101-9. DOI:10.1021/ac990800w · 5.83 Impact Factor

Thanks for the articles David, I will have to study them this weekend. How do you know when to use Kovats RIs or other RI types?

When you have even and odd chain compounds, its dead easy, because they are all the same. Otherwise you have to exterpolate-  figure out where a missing peak should be located, as your GC's linear temp program continues.  Plot them on some system with excel. or by hand on graph paper (if you can find any--- to be sure).  NIST gives wrong answers very easily.  

Menon's answers were good by the way.

What are your samples from?   

Hi David,

Sorry for the delayed response. I have created a RI calculator in excel in order to aid with identification (attached). However, in the litertaure both Kovats and Linear retention indices seem to be utilised interchangeably. How do I know whether to use Kovats or another RI type?

The samples are from two-spotte spider mite (Tetranychus urticae) hexane washes of a sample size of 100.

Thanks for your help.

KOvats are dead easy- just add 100 to the carbon number.

Yeah, I understand that. But I thought Kovats and linear RIs are calculated in different ways as Kovats are used for isothermal analysis, while linear are for temperature ramped analysis. Or is that wrong?

I never did any isothermal plots. all were deltaT.  There are new definitions then. There are so many peaks in most insects, if you try to plot only isothermal, you will never get done.  I'll bet the differences are lost in noise/ measurement differences, or maybe too large peaks sometimes when you guess where the top is.

Here is the point- When you have "families" of compounds: the Retention Indeces will come out very similar to each other in each group.

3-methyl alkanes from C27- C37,                        KI xx65

internal monobranched-- 13-methyl C27-C37,  KI xx25

internal dimethyl-- 13,17-dimethyl C27-C37.      KI xx55

internal trimethyl-- 13,17,21-trimethyl C27-C37.  KI xx75

As the methyl branches approach the center of the chain, the molecule gets "rounder", and the di- and tri-Me branched compounds remain within the same carbon number, despite having  2 CH3-s or 3 CH3-s.  The computer program cannot figure this out yet.  

I don't know if this is clear enough yet.  I comes out pretty well in my tsetse fly papers, or the fly papers.. 

So what do your spider mites have?

Hello Joe

I use the Kovats index. Attached is a small presentation about the metod.

Hi Eduardo,

No, you did not use the Kovats Retention Index in the attached pdf, you used the Lineair Retention Index (LRI) or Van den Dool and Kratz Retention Index.

Unfortunately it is a common mistake to call Lineair Retention Indices "Kovats Retention Indices". They are different RI's calculated with completely different formula's.

The Kovats Retention Index is calculated with the formula with "log" in it and should only be used when running the GC-oven isothernal.

Reagrds, Alex

Hello Joe

Hi Joe, you can calculate the retention index by doing an injection of the linear alkanes. But there is another way, so if you want more information, you can gladly contact me by email: [email protected]

Several corrections/explanations appear in this thread by Joe, Nico, Tom, Alex and Henry. Kovats Indeces are very old, from the time of packed columns and isothermal ovens and no computers. Me too. Use of "KI" without explanation is misleading, and obsolete. It was used to get entomologists to understand how useful and universally present hydrocarbons are in closely-related insects. Linear Retention Indeces (LRI) established for specific natural insect methyl-branched alkanes don't change, using temperature-programmed non-polar capillary columns. Do modern MS search engines provide for use of this definitive tool?

Besides from the practical advice, this thread is a useful reminder of the importance of retention indices, which provide a direct link between chromatographic theory and practice. At the risk of going a bit off-topic, it may be worth adding some comments to my reply of 3 years ago by generalising the discussion to to the unsatisfactory situation regarding liquid-phase chromatography with bonded stationary phases.

I'm interested in biochemical separation methods where you want to be able to define an as-yet unidentified metabolite unambiguously in terms of its chromatographic properties. At the early stages of an isolation, it is not possible to associate a spectral fingerprint with the biological activity of a chromatographic peak. However, with modern HPLC, we no can longer define retention indices that would be useful when profiling and identifying metabolites (the current term for that is, unfortunately, "metabolomics").

The different GC retention indices are in fact accurate surrogates for the thermodynamic gas-liquid distribution constants, whenever the stationary phase can be treated as a homogeneous true liquid of defined composition on an inert support. The stationary phase can be of any polarity. With capillary columns, the stationary phase is usually cross-linked and bonded to the column wall, though I understand that manufacturers maintain a range of columns that are equivalent to the old packed ones in terms of the retention indices of appropriate test analytes. An important point is that care is taken to ensure equivalence between manufacturers. This might be considered simply in terms of compliance with official analytical methods, but it also means that retention indices can still be used as an identifier - for all time - of compounds within a given class and, with greater certainty, to establish that a given compound is not present in a specimen.

This link between retention and physical constants was, more or less, a feature of nearly all the liquid-phase separation techniques that preceded HPLC. The obvious example is liquid-liquid counter-current chromatography, which was used for isolating many of the "classical" natural products, metabolites and hormones. Polymers including polystyrene and paper could in practice be made in a highly reproducible manner. By contrast, I have read numerous accounts of attempts since the 1980s to separate hormones and other metabolites by HPLC on bonded silica (a proprietary composite material), where biological activity was the only means of detection. In general, it has not been possible to compare one report with another, especially when manufacture of the columns in question ceased a quarter of a century ago. Attempts have been made to classify columns of a general type such as C-18, but they fall well short of defining a column in terms of the retention indices of a wide range of likely analytes.

GC is still widely used for establishing profiles of metabolites that are volatile or can reliably be converted to volatile derivatives. Attempts are being made to obtain comprehensive profiles of all that metabolites of in a given class that are detectable in a given biological material. We can be sure that such methods will stand the test of time, provided suitable columns are chosen.

However, leaving aside the problem of mass spectral artefacts, attempts to establish comprehensive and universally applicable metabolic profiles or maps using HPLC seem doomed. You can't properly define the stationary phases. There seems to be a strong case for reverting to liquid-liquid counter-current chromatography, for which instrumentation continues to make progress.

Hello Christopher R. Lee Could you tell me where I find more information about this? "liquid-liquid counter-current chromatography".

Emmanuel Anosike

I have a recent review of the literature: DOI: 10.1039/c5np00061k

The literature may not reflect usage because many applications are industrial. I haven't seen anything on metabolic profiling, though a multidimensional version of the technique might be one way of realising the dream of mapping the metabolome for all time (recent article: 10.1021/acs.analchem.7b00925 ; see also LCGC Europe, April 2018, p222). Some work may be required to adapt the approach to the wider range of solvents: for example having to purify them oneself, interfacing with MS.

The equipment manufacturers I know of are:

Chromaton

http://www.kromaton.com/en/

and Gilson

https://www.gilson.com/system-verity-compact-cpc-system.html

Alex Van Renesse sir pls send me any pdf regarding KI calculation