What does linear velocity (35 cm/sec) mean?

[Peter Apps]

"I have used this feature in B.02.01 Chemstation, it is fun. But in B.03.01, it is not available any more. "

A set and forget linear velocity function was one of Agilent's big selling points - I wonder why they dropped it (and if they really have).

Peter

[krickos]

Terry: Maybe I was unclear again, was always talking about setting the mode to constant flow, have not seen the constant velocity option.

Now if I have not suffered from a complete brain breakdown after partying friday/saturday, the Velocity is determined by column length (cm) divided by t0(tm) (seconds). As the column volume is constant and the column flow (ml/min) is constant in the constant flow mode, how can the mobile phase velocity change for a unretained analyte?

After testing on the software (a.09.03) it looks like it should on the plot when running constant pressure ie a drop of velocity with increased temp in constant pressure mode, but strangely the velocity increases in constant flow mode with temperature? ie the opposite to Terry.
Seems like a software issue or do I need to soak my brain in water?

[JI2002]

I don't think there is such a thing as "constant velocity mode". Actually, the more accurate term here should be "average linear velocity". The linear velocity of the carrier gas is different across the column. It's slower at the inlet of the column and faster at the outlet of the column. In the Agilent GC method translator software, the outlet linear velocity can't be determined for GC/MS because it's very large.

Constant flow doesn't mean same average linear velocity.
Same average linear velocity means same average flow rate(Fc). Fc is different than "flow rate" that chemists usually use.

It's easy to physically determine the average linear velocity: measure the retention time of a unretained compound (to), average linear velocity = L/t0
where L is column length in cms and t0 is in seconds.

It's a little complicated to calculate average linear velocity:

Average linear velocity = (Pi - P0)*j*dc^2/(32*L*n)

Pi is inlet pressure
Po is outlet pressure
j = 3/4 * (p^2-1)*(p+1)/(p^3-1) where p = Pi/Po
dc is column diameter
L is column length
n is carrier gas viscosity, it's parameter dependent on temperature T:

n(T) = n0(T/To)^x

n(T) is the viscosity at temperature T
n0 is the viscosity at zero degree C
x is a constant of carrier gas, for helium x = 0.646.

You don't have to calculate all these manually, just use the GC method translator from Agilent to do the math.

[JI2002]

the Velocity is determined by column length (cm) divided by t0(tm) (seconds). As the column volume is constant and the column flow (ml/min) is constant in the constant flow mode, how can the mobile phase velocity change for a unretained analyte?

Carrier gas average linear velocity increases with the pressure during temperature ramp in constant flow mode. Constant flow is diffwerent than Fc, which is average flow rate.

[Terry]

JI2002, thank you for the explaination. It is very informative, I would refer to some basic GC theory books for details.

Krickos, probably we should consult this difference with Agient service people

Time for bed here in China, bye for now

Terry

[Terry]

I did not see JI2002's second reply when I posted my last message

[Peter Apps]

JI2002

When I worked on a 6890, running off Chemstation there were three ways you could set the controls for the carrier gas; pressure, volume flow, and linear velocity. Depending on which one you chose, you then entered a number in the corresponding screen box (which already had the units) and the software did all the appropriate calculations to give you what you asked for by controlling the pressure at the head of the column. If you selected anything to stay constant, it remained the same during the run, not (necessarily) along the length of the column. So in that sense there is such a thing as constant linear velocity.

I do not have Agilents in my current lab and I do not recall the exact screens and keystrokes.

Peter

[JI2002]

I knew it's doable to run at constant average linear velocity, but I didn't know there was this option. Thanks for pointing that out, Peter.

[krickos]

JI2002

Thanks for the update on velocity, had a nagging feeling all day in my backhead that something was off.

[Terry]

...but strangely the velocity increases in constant flow mode with temperature? ie the opposite to Terry. Seems like a software issue or do I need to soak my brain in water?

Krickos, my poor memory does not serve me correctly. Back to my lab, I have just checked the plot again, it is the same as yours, not opposite. I would flush myself down the toilet.

[Ron]

In most cases when a method specifies a linear velocity, the assumption is made that the velocity is set at the initial temperature. For the USP method this will probably give good results, if not I would try setting the linear velocity at the midpoint temperature of analyte elution.

There are GC systems out there that allow constant linear velocity to be set, when this is done the pressure is ramped to maintain the same average linear velocity throughout the temperature program.

[Terry]

Ron, thanks for the clarification. I got it.

Terry

[mbicking]

I have read the USP method many times now, hoping that it would eventually make some sense. Unfortunately, ...

Anyway, here are a few other comments about the method.

1. Both procedures (A, B) employ a 20 minute isothermal hold after injection. All of the Class 1 analytes elute in this window in both Procedure A and B. Most of the Class 2 analytes elute in 20 minutes for Procedure A, and I recall even more for Procedure B. So, the issue of linear velocity, flow, and pressure changes, etc. are not important for most of the analytes.
2. I doubt most auditors would appreciate the subtleties of this excellent discussion about GC operation. They will be concerned about your results more than your settings. If you can meet the system suitability requirements (resolution and signal-to-noise) and demonstrate some level of accuracy and precision, that's all you should worry about. Obviously you have to document what you do, but don't worry as much about the settings if your system works.
3. The method has no precision/accuracy criteria; odd for a standard method, don't you think? I am told that the general USP limits apply. Based on our early pre-validation experiments, there is a serious problem with standard and sample stability. This would explain why the method, as written, is really only designed to analyze one sample at a time.

Oh, and the reference standards are available from USP at $185/3 ampules. And the method uses up these standards at an alarming rate. Sounds like a nice revenue scheme.

Oh (#2), the Class 1 standards are back ordered for 6 - 8 weeks. But, you can get a similar mixture from Restek. They had 200+ in stock today.

[Consumer Products Guy]

I have read the USP method many times now, hoping that it would eventually make some sense. Unfortunately, ...

Hard to believe, beloved USP !!!

The method has no precision/accuracy criteria; odd for a standard method, don't you think? .

Hard to believe, beloved USP !!!

The reference standards are available from USP at $185/3 ampules. And the method uses up these standards at an alarming rate. Sounds like a nice revenue scheme.

Hard to believe, beloved USP !!!

[krickos]

Oh, and the reference standards are available from USP at $185/3 ampules. And the method uses up these standards at an alarming rate. Sounds like a nice revenue scheme.

See some Q&As

http://www.sepsci.com/chromforum/viewtopic.php?t=9538