Hi everyone,
I am applying the EPA 552.2 method. Two of the peaks are overlaping one each other. What can I do?
I was thinking to reduce the injection volume (1μl instead of 2μl) or inject with split mode or put an extra ramp in the oven about 10oC before the elution temperature of the compounds.
Any thoughts?
Another way to solve the overlap is optimize the column temperature and split ratio.
I would say that optimising GC oven temperature ramp, using a lower temperature rate before the time the overlapping happens, or even going for isothermal for a while, might help.
Try a longer GC column or, if you can't change it, slow down the ramp (even 2 °C/min)
yup, reduction in ramping time will be helpful to you. you can also try isothermal between those peak retention time..
Important way to solve your problem is optimize the carrier gas velocity (according to Van Deemter method). Probably it is necessary to change the linear velocity of carrier gas, try to reduce the velocity.
Thank you really all for your advice!! I will try to apply these changes and i will let you know the results from here.
Hi Alexandros, indeed, optimizing the column temperature, carrier gas velocity it's of a great importance. I know you may know what I'm about to say, but you need to know the physicochemical profile of your compounds in order to achieve a better optimization, because this information are going to tell you whether or not you can work with high, low or isometric temperature ramps. As soon as you know that the next stage is to learn if the column you are using is the proper one for the compounds you are using.
Now, if you are still having the same problem you could think about changing the temperature ramp, lowering that so maybe the compounds could separate, that usually solve the problem.
Just saw you have the same answer before. Well, hope it was of some help.
Normally, EPA std programs are optimized for the compounds they are developed to analyze. If you have the same column (phase, length and film thickness), and these two peaks are overlapped, you can try to separate them not only with the oven temp rate (slight changes) and the split ratio or the injection volume, but also trying to diminish the column flow in order to produce a better separation.
This is copy of temperature programming for column DB5.625 and backup column DB1701 table 2/ caslab for EPA 552.2.
Yes, as you say, reducing injection volume and reducing sample size can help.
I think they have already experimented and adjusted the temperature programs quite well for maximum separation of these brominated and chlorinated short-chain acids. You didn't say which compounds are overlapping on your column, but some of these standards elute quite close to each other. If both your columns fail to separate analytes when they should, your column may be damaged or have interfering residues. (Cut off a few cm?).
Column A: DB-5.625, 30 m x 0.25 mm i.d., 0.25 µm film thickness, Injector Temp. =
200°C, Detector Temp. = 260°C, Helium Linear Velocity = 24 cm/sec. at
35°C, Splitless injection with 30 second delay.
Program: Hold at 35°C for 10 minutes, ramp to 75°C at 5°C/min. and hold
15 minutes, ramp to 100°C at 5°C/min. and hold five minutes, ramp to
135°C at 5°C/min. and hold two minutes.
Column B: DB-1701, 30 m x 0.25 mm i.d., 0.25 µm film thickness, Injector Temp. =
200°C, Detector Temp. = 260°C, Linear Helium Velocity = 25 cm/sec. at
35°C, splitless injection with 30 second delay.
Program: Hold at 35°C for 10 minutes, ramp to 75°C at 5°C/min. and hold
15 minutes, ramp to 100°C at 5°C/min. and hold five minutes, ramp to
135°C at 5°C/min. and hold 0 minutes.
@David Arthur Carlson
I'm a bit surprised that an official method still refers to GC stationary phases by their trade names, without any further qualification. Even the pharmacepeias are more up to date in that respect.
An important difference with respect to LC is that the retention properties of the usual phases are fairly well defined by their chemical composition, subject to the relatively small effects of proprietary cross-linking and bonding. I mention exceptions below. I'm writing this reply because it seems worthwhile to maintain the ability to define GC (and electrophoretic) methods in a way that make them valid for all time, without this being eroded by commercial interests.
DB5 is an old trade name for 5% diphenyl, 95% dimethyl polysiloxane; one manufacturer used the '5' for 100% dimethyl polysiloxane, reflecting its polarity with respect to the reference hydrocarbon phase. This shows how important nomenclature is in this field.
The '1701' phases are 14% cyanopropyl phenyl, 86% dimethyl polysiloxane and should be named as such.
I don't have shares in Restek, but for many years they have provided a convenient reference table: http://www.restek.com/Chromatography-Columns/GC-Columns/GC-Column-Cross-Reference-Columns-by-Phase
'DB' belongs to Agilent, but their column selection guide didn't have DB5-625 (http://www.chem.agilent.com/cag/cabu/propphase1.htm). Presumably this is one of those proprietary phases, containing several substituents, that are designed to produce a particular separation. They may be named after an official method (like the '624' columns) but still, in general discussions like this one, it's important to indicate this because not everyone is up to date with the particular practices of the EPA and other official bodies.
If you could have provided the specific sample and probable impurities I could have suggested u the most apropriate method for separation.However,there are certain thumb rules for peak separation in GC which may be helpful: (i) Reduce the injection volume,(ii)dilute the sample in an inert solvent,(iii)Change the stationary phase or column as per your sample and impurities(iv)apropriate temp. program,(v)suitable GC detector,(vi) use of software facilities in GC equipment,(vii)Use of 2D GC.
Hope this helps you.Best of luck!
I reduced the injection volume and the He flow rate but i did not see any significant improvement. But when I reduced the temperature rate increase before the co-elution of the compounds of interest, they seemed to be separated quite well.
I run also unknown samples and i noticed that the co-elution was so intense only for the standard samples, where the peaks were much larger (both in area and height) due to their higher concentration in comparison with the unknown samples.
For those who asked, the co-eluted peaks are Trichloroacetic acid (TCAA), Bromochloroacetic acid (BCAA) and the Internal Standard (1,2,3-trichloropropane).
The column in the oven is a Restek's Rtx-5MS (30m x 0.25 i.d. x 0.25 μm).
The derivatives are methyl esters, which are quite polar and could be difficult to chromatograph. The method prescribes a specific commercial column DB5-625, which may have been designed for this kind of separation; it can't be assumed that a general-purpose 5% pheyl methylsilicone will work. If Agilent were to take it off the market the official methods that depend on it would effectively cease to exist. The internal standard, a relatively much less polar and reactive chlorocarbon, seems to be totally unsuitable, but who am I to question the wisdom of the EPA?
Reducing the programming rate is often useful. Note that exchange dynamics are not necessarily reproducible between columns (film thickness, cross-linking, polymer lenghth) and official methods should allow adjustments.
Finally, note that an electron capture detector has a non linear response that may vary between days. Any linearisation algorithm in hardware or software (not specified here) will have to assume a response function that may not in fact be correct. The response curve is "saturating", meaning that as you increase the amount injected, the peaks will become relatively broad at the base and may appear to be less well resolved than is the true physical case. Quantitation is a nightmare in such situations, and it is hard to understand why the ECD was not replaced by a small MS decades ago - if only on account of the reduced expense of troubleshooting a detector that would probably never have been commercialised if MS had been readily and economically available at the time.
First off I would recommend using the updated EPA method 552.3 July 2003
The DB-5.625 column is still available from Agilent (part# 122-5631) - it is made specifically for this type of analysis (more inert).
EPA method 552.2 was published 20 years ago 1995, if you recall this time period in the GC column world there were not as many choices of column manufacturers, especially with the reproducibility that made J&W (The DB brand) famous. That is why they specified a certain brand of column, and the fact that this is a drinking water method that requires strict adherence to all steps in the procedure.
EPA 552.3 still specifies the same columns, however they now add the "or equivalent" language to allow flexibility. They have also switched the primary column to a DB-1701. Section 6.14 also allows more flexibility in regards to GC.
If you are not getting the required separation, I would pull up the specific instrument parameters of the method (including the columns) and follow them. Believe me it will work. One of the requirements of the method is confirmation on a second column. Make sure you are doing this, it is required to confirm the presence of your compounds.
Some questions - Are you using a 2mm quartz liner in the GC? Is it clean? Change it frequently, before each time you calibrate. Also the method recommends a calibration range of 1.0 ug/L to 20 ug/L. Follow this and dilute into this range if necessary.
These are reactive compounds, pay attention to your inlet and GC parameters.
What is unspoken in this series are the rules of thumb, based on the principles of gas chromatographic separation:
The best possible gc separations for peaks are isothermal, with the highest diffusivity gas,using a column with great selectivity for the compounds of interest, a well configured instrument, and the sample concentration well below the region of column overloading.
Once the column, instrument and general method are defined,, the easiest ways to increase resolution are usually to lower the slope of the temperature program, and reduce sample size.
Further to the discussion on changing the temperature program, I forgot to mention (for completeness) that there's a bit more to it than separation dynamics. Sometimes (in both GC and LC) the selectivity is slightly temperature-dependent and a change in column temperature may be enough to tweak up an isothermal separation that isn't quite good enough; of course the effect may be one of the effects of changing a gradient.
@Peter Philbrook:
You recall the need for a confirmatory column. This is a digression on a technical point, but it might be interesting to know if the EPA considers it legitimate for labs that dedicate instruments to particular groups of methods to install both columns permanently in a single oven having one injector and two detectors. That way you save a lot of time, money and some injection vials, and both columns are always ready.
A long while ago, when I was setting up dedicated headspace instruments for volatile impurities (residual solvents) in pharmaceuticals, colleagues needed persuading that this is compatible with the wording of the official method.
I much prefer(red) inserting the columns directly into the injector using a two-hole ferrule, rather than the common and trouble-prone practice of placing a y-coupler downstream. There can be no objection here, because it doesn't matter if different fractions of the analytes get into the two columns (Books have been written on fractionation in split injectors...).
With syringe-type autosamplers it's also possible to use two split injectors, but that places constraints on timing and the temperature program. Also, you may lose what turned out to be a considerable advantage of the data system referencing the two chromatograms as one run.
As for installing the columns, beginners may regret the absence of a second pair of (subminiature) hands, but it isn't too difficult after a little practice, and anyone who has played with GC column switching finds it easy.
@Christopher Lee
Yes we have been doing that here for decades. Single injection, 2 hole ferrule, 2 columns,dual ECDs.
We have gone to building our own gc's so that we can get the best results possible. Attention to a lot of hardware and software details make a world of difference, and the 'big boys' supplying chromatographs have forgotten about that. Some instrument vendors massage the data to make it look better than it really is, and that can compromise the separation and the quantitation.
I have been involved designing, building and using chromatographs for about 45 years (my first gc publication in 1968! ) and am disenheartened with the rather modest performing but shiny and expensive hardware on the market these days.
@Peter Philbrook
Thanks for your confirmation that the EPA accepts that qualified analysts can be trusted to understand what they're doing. Perhaps I was too influenced by attitudes that prevailed in the pharmaceutical sector, where an excess of "quality assurance" sometimes lead to inefficiency and, in the end, bad science.
In order to solve your problem You can use longer GC capillary column or use another capillary column filled in with more polar phase.
@Euzebiusz Jan Dziwiński
Part of the difficulty here is that this sort of analysis may have some regulatory purpose, meaning that the decisions based on the results have to stand up to official review, for example by a pharmaceutical regulator or even (e,g, in case of pollution) a law court. In such cases the analyst must use an official ("compendial") method if one is available. If a new method is required, validation is a long and difficult procedure.
Compendial methods are sometimes out of date, difficult to implement, or plain unsatisfactory, for various reasons, despite the best efforts of those who developed them. However, until recently you were expected to follow them "to the letter". You would certainly not be allowed to use a different stationary phase, and it took the authorities a long time to come to terms with proprietary phases with no proven alternatives, which are practically the norm in LC. There is still debate on how much you can change parameters such as flow rate, column length and temperature program, though optimum values will obviously change from one column to the next.
Generally split mode is always best and gives greater options for reducing overloading, certainly reducing the split ratio may help. Slowing the ramp will also assist and keep the He flow at about 1 ml/min. Decreasing injection volume or buying longer columns are often unnecessary unless this is going to be routine and then you need to think about validation. As for phase just consider the chemistry if the analytes does what you are doing make sense?
Separation processes in chromatograph depends many instrumentation parameters such as temperature program as well as matrix constituents. In some cases, chemical treatments steps are used in analytical procedures. Modification is necessary for standard methods.
I see that EPA 552.2 and .3 are similar: PRIMARY GC COLUMN – DB-1701, 30-meter length, 0.25-mm i.d., 0.25-µm film (14% cyanopropylphenyl-methylpolysiloxane), or equivalent bonded, fused silica column.
552.3-9 mentions CONFIRMATION GC COLUMN – DB-5.625, 30-meter:
Does the Confirmation GC column also not separate your 2 compounds?
Is your problem with dimethyl sulfide coeluting? Can you oxidize this DiMeSulfide in your sample?
Does smaller sample, or cold on-column injection method help?
If the separation is at least partial, consider either of 2 approaches for quantitation: (1) Use Snyder's corrections for overlapped peaks (2) reduce the temperature gradient.
Lloyd Snyder did a great study on the quantitative effect of one peak partially overlapping another. He provided examples for both leading and trailing peaks, peaks of different Rs's and differing ratios of peak sizes. One needs to see those results and be surprised by performance of peak height vs peak area in quantitation. In many cases, peak height works remarkably well !
If you're into programming with Excel, it is easy to model a 2 peak separation and compare the results of integration and peak heights with various degrees of separation and ratios of peak sizes.
One must remember that, in general, temperature gradients ("temperature programming") are used for reducing elution time at the expense of separation. Lowering the time a substance remains in the stationary phase reduces its ability to be retained and equilibrated with the stationary phase. The closer to true equilibration, the closer to the distribution coefficient. When 2 materials of similar distribution coefficients are involved, the separation lies on allowing the materials to approach their the true equilibrium distribution coefficients (which must be different for a separation to take place), the practical result of which are their (hopefully different) retention times.