Clarus 500 Perkin Elmer

[sadsal123]

Can anyone help with this please. We have recently got this instrument of another company and I am trying to develop methods on it. I have finally managed to get decent separations but I am now having

1) Poor Reproducibility: I have changed Injection Septum and the Wool in the injection port liner which has improved the reproducibility, but its still not good enough. I managed to get the RSD down from ~8% to ~1% on repeated injections but when I run a 20% sample, I am getting results anywhere from 18%-19.5% which is not good enough for us. I used an external calibration method with average calibration factor.

2) When the report is produced, its not giving me the correct amount. Even the 20% sample is coming out at about 98.5% but the raw amount is fine.

If anyone is familiar with the total chrom workstation, can you please help?

Thanks

[Peter Apps]

In order to suggest improvements we have to know:

what are your analytes (identity and concentration), sample solvent, and operating conditions for the inlet, column and detector. This includes the injection volume, speed of injection, wash cycles, gas flow rates, temperature programmes etc etc etc. In detail.

Peter

[sadsal123]

Sorry Peter,

We Process a wide range of waste Solvents. These include Toluene, Acetone, Methanol, Ethanol, IPA, Propan-2-ol, Butan-1-ol, MEK, Ethyl Acetate, Propyl Acetate, DCM, NMP, Xylenes, Cyclohexanone, Butoxy Ethanol, etc, etc.

When we get a load in, we determine the water content and the purity or the contents of the load and then decide if we want to recover the solvents.

I am trying to create various methods for the chlorinated and non-chlorinated solvents. The longer methods seperate about 21 different components and I have managed to get the separation but am struggling quantifying.

For example, I have created a simple method for toluene:

I am using a temperature programme: 120oC for 1min then ramp to 180oC at 20oC/min and hold it for another minute.

Injection Method: Auto Injection.
Syringe Capacity: 0.5ul
Injection Volume: 0.1ul
Injection Speed: FAST

Sampling Rate: 25pts/s
Inlet A: CAP
Detector: FID

Pre-injection Solvent Washes= 1 (WATER)
Pre-injection Sample Washes = 4
Post-injection Solvent Washes = 2

Sample Pumps: 6
Viscosity Delay: 0

Injection Temp: 200oC
Carrier Gas: Nitrogen at 1.5ml/min
Split Ration: 50:1
Column = Elite 624 (6% Cyanopropylphenyl -94% Dimethyl Polysiloxane), 30m, diameter 250um

Detector Temp: 300oC
Attenuation: 0

I think I have put everything here. I am sorry its got really long but I would really appreciate your help Peter coz I have been trying to figure this out for the last few days and not getting anywhere.

Thanks agains,

Salma

[Peter Apps]

Hi Salma

Thanks for all the details. Some suggestions:

0.1 microliters is a very small volume which it is difficult to inject repeatably. I presume that you have a plunger in needle syringe. Can you change to an ordinary plunger in barrel syringe and inject 1 microliter ?

Your temperature programme is too fast for the length of column. I presume that you are ramping fast in order to get the analysis done quicker. It would be better to have a shorter column with a lower start temperature and a slower programme.

The water wash might be a problem (but with the wide range of solvents you analyse it will be difficult to find a substitute) - if traces of water are injected with the sample you could be getting sharp pressure pulses in the inlet that upset the split ratio. Try just "washing" the syringe with multiple pumps in air to evaporate left-over solvent.

Nitrogen is a terrible carrier gas because its optimum flow velocity is so slow. If you switch to helium the analysis time is almost halved, with hydrogen it is reduced to a quarter. Your gas flow rate is way too high so you are not getting the separation that your column could provide under optimum conditions.

Manufacturer's claims notwithstanding, to get injection repeatability of better than 1% needs everything to be nicely tuned and optimised.

Good luck Peter

[sadsal123]

Thanks a million Peter. I will try these....

Salma Patel

[sadsal123]

Peter, Any clues about the second problem? Are you familiar with totalchrom workstation? I get the impression that it has something to do with the setting in the software but I cant figure out what.

[Peter Apps]

Sorry, I've never used Totalchrom.

Peter

[sadsal123]

Hi Peter,

I tried some of your suggestions and have managed to get the RSD down to 0.71 so thanks for those. It was really helpful. The other thing I wanted to know was what should I do to get very sharp peaks. They are sharp but I would prefer them to be sharper and slightly narrower. What do you think?

Thanks.

Salma

[Peter Apps]

Hi Salma

Beware of wasting time making a thing better than good enough.

Remember that sharper peaks are not necessarily better resolved.

How wide are your peaks now (in terms of seconds at the base, or at half height - you can probably make the software print this out) ?.

You will not get sharp well resolved peaks while you are using nitrogen as carrier gas. Use helium at optimum flow rate and their widths will shrink to less than half what they are with nitrogen at 1.5 ml/min.

Peter

[sadsal123]

Thanks Peter, you have been a great help. Currently, I have a very big back log of samples to put through and therefore I would prefer to set up a method with whatever we have but I will definitely optimize and get helium once I am upto date with my back log. Peak width is not that bad so I will stick with it. Thanks once again.

[avrc]

Hi Salma,

I am somewhat familiar with TotalChrom software.

Regarding your second problem, you are saying your raw amount is fine.
Then I assume your corrected amount and adjusted amounts are not o.k?
What about your calibration table? Is your method includes a correct calibration table?Are you using a calib. curve or a calib factor?Are you using 'normalised amount' in your report?Please give datails.

Regarding your first problem, you can change the inj. speed to slow and give some viscosity delay time(3 or 4). This might improve your repeatability.
Or else, you can try with a 5uL syringe instead of 0.5uL and inject 0.5uL and increase the split ratio.

Good luck.

av ravi.

[sadsal123]

Hi Mr. Ravi, Thanks for responding, i carried out some of the suggestions given by Peter and managed to improve the RSD significantly - from ~8% to <1%. I would still like to improve it further and will probably try some of your suggestions but will probably have to wait because the syringe I am using on my Instrument is very new so I will have to wait a while before I order another one (unless I can find someone that sells them cheaper than what they are!!)

By the way, what does viscosity delay do? I am not very keen on using a slow injection speed because the solvents we are working with are very volatile so I dont want to lose any sample before injection. What are your thoughts?

Also, is there a limit to split ratio? For my Acetone method, I am using 0.2ul with a split ratio of 150 and a carrier flow of 0.4. I am currently checking the reproducibility of this and having injected about 5 samples, I have got an RSD of 0.54%. Again ideas and thougths are welcome.

As for my first problem, I am not entirely sure as to what to select in the report body. I am using an average calibration factor on repeated injections of a 100% solvent. Again this is because the strengths of solvents we get in can be anything from 0 to 100% but because of the nature of our work and through put, I dont want to have complicated or time consuming calibration methods.

If ur interested, I can email you the method file, sequence file, etc for you to have a look at and see if you can spot the mistake(s).

Thanks for your help.

Salma....

[sadsal123]

Hi Peter,

Another question for you coz I am sure you'll have some ideas: I have managed to sort out the reproducibility problem by injecting a bigger volume and using a higher split ratio, so for MeOH, I got it down to about 0.3%. I have another problem though, My 22 component standard also contains THF and MEK which are eluting very close to each other, but still seperate. However, when I am trying to calibrate, its not calibrating the MEK because either its not identifying the peak which is strange because the elution time is as I have set it or secondly, its failing it on user criteria, which is something that I have set (this usually means the rejecting outliers which I have set to 3%).

I am trying to overcome this by reducing the flow slightly more to see if these two peaks will seperate further. Waiting for the result but have you any more suggestions?

Also, what is the likely cause for a peak to have 2 apexes?

Thanks.

Salma

[Peter Apps]

Hi Salma

The failure of the peak to integrate is probably a software settings problem - if the signal does not reurn to baseline between the peaks the integrator cannot recognise that a new peak has started.

I do not know the software that you are using, but you will very probably find that it has an autointegrate option which allows the software itself to optimise the settings for peak detection, noise, peak width etc. This is often a useful start.

If you are still using nitrogen the average linear flow rate through the column (inject some methane or lighter gas and divide the column length by the retention time) needs to be 15 cm/s. If you go slower than this the resolution will get worse, not better. For helium the optimum flow rate is 35 cm/s.

Two apices on a peak can be due to two components co-eluting, too much sample on the column, too high a signal, or a setting on the integrator that lets it interpret the brief pause and short term noise at the top of a peak as the end and beginning of a peak. This problem might be fixed by the autointegrate option.

Peter

[sadsal123]

Thanks Peter, you really have been a great help with my method development. Thanks once more.

Salma