I attempted to use my Agilent 5973 with a high temperature, aluminum clad, capillary column. I got very, very poor sensivity. However, the instrument tuned perfectly. I checked for leaks inside and out and could not detect any. I thought that perhaps the column ferrule on the detector end was constricting the flow. I loosened the fitting until the source pressure went up then slowly tightened the fitting until vacuum was restored. Still no change in sensitivity. The source pressure goes up when an injection is made, so I know that the solvent is migrating through the column. When I swapped the column with a trad. polyamide coated column, the system worked fine again. Is it possible that the metallic column coating is causing a problem?

What's your calculated flow rate? Sensitivity goes through the basement at flowrates significantly higher for which the system is rated - higher source pressure means less electrons available for analyte ionization.

Which pump do you have on the 5973, the diffusion (1.5 mL/min rating), standard turbo (2 mL/min rating, or performance turbo (4 mL/min rated)? With a megabore column your pressure drop will be very low and you'll need to watch what you set your head pressure.

For a 5 m x 0.53 mm column: 5 psi = ~74 mL/min at 100C (outlet to vacuum).

For a 30m x 0.53 column: 5 psi = ~12 mL/min at 100C (outlet to vacuum).

Both conditions would exceed all the MSDs flow handling capabilities that I know of and would most likely result in sensitivity losses.

At 1 psi x 30m x 0.53mm, you're still putting about 8 mL/min into the source, but if you had the performance turbo you could probably get away with it, especially at the higher temperatures.

I don't know about the metallic coating affecting anything. My guess is no, because the transfer line jacket is metal and also protrudes into the source...

Spencer

I assumed that you were using a megabore column because I'm not aware that metal-clad columns commonly come in smaller dimensions. They very well could, but I wouldn't know who supplies them. My answer still applies, but if you are using a smaller i.d., please post!

Spencer

One way to tell if the flow is significantly higher is to compare the source pressure using the Al clad column with the polyimide clad column. If the source pressure is significantly different, too high a flow is likely the problem. Even though the performance turbo option on the 5973 will pump out the 4 mL/min, the source pressure will be higher and the sensitivity lower than at 1 mL/min.

I don't know what compounds you are analyzing, but the aluminum clad columns usually have a very thin phase coating, as the rationale was to operate at elevated temperatures. If the compound is adsorbed or degraded due to contact with uncoated column sections (not that uncommon with a 0.1 um phase) that could be a reason for loss of sensitivity.

A metal column with the same dimensions as a polyimide-clad fused silica will produce a negligible pressure difference in the source. I say negligible because He can diffuse out of the fused silica column, albeit at very slow rates. Knowing your column dimensions and head pressure renders the experiment above unnecessary.

Spencer

I would be extremely careful trying to use an aluminium clad capillary with a mass spectrometer (I presume a 5973 is a mass spectrometer?). I do not know the exact operational mode of the 5973 or the layout in the ion source, but mass spectrometers use high potentials and aluminium is a pretty good conductor of electricity. I think you have the possibility of damaging your instrument or even creating a safety hazard. I would not be surprised if the lack of performance is connected with this. I would consult the manufacturer or the instrument manual before going any further. In fact I thought this was one of the main reasons why aluminium clad columns had fallen out of fashion. The other, I thought, was that the cladding can become disconnected from the fused silica with repeated temperature programme cycles leading to breakage.

Our Agilent engineer was asked this a few months ago and thought metal or metal clad columns would be okay, but we haven't tried them yet. Ask Agilent. Yes, they do come in smaller bores than 0.53mm.

Anonymous,

Please post your source for <0.53 mm metal columns. Thanks.

Spencer

QA source for 0.25mm i.d. metal columns is Quadrex 1-800-275-7033/203-393-3112 www.quadrexcorp.com. We have used their aluminum clad high temp columns in the past (available in three phases , three lengths each) but the columns became brittle and broke after about 4 years so they were very useful to us (400-1HT, 400-5HT, 400-65 HT). We then decided to try their 0.25mm i.d. stainless steel capillary column instead and have been very pleased with it; it was UAC-1HT-10-0.15F (methyl silicone type, 10 meters, 0.15u film). It looks like this type is available in 8 phases.

Thanks...

Sorry, I haven't gotten back sooner. I am indeed surprised how much feedback this has generated. The column is an Quadrex 5% phenyl, 15m x 0.25mm; 0.1um. The 5973 is a turbo model, I believe it is 2ml/min. turbo. the I hadn’t considered the loss of sensitivity due to higher source pressure. We usually run around 6x10-6 with a 30m x 0.25mm. I don’t recollect the pressure being high, but I can’t honestly remember what it was running at. I ran it with a head pressure of 1.7psi (I’m not sure if the vacuum would let me go any lower) which the software reports as 50cm/sec, 0.9mL/min (I didn’t confirm this). The flow rate appears to be well within the systems pumping capacity, but perhaps there is still a loss of sensitivity due to higher source pressure. I was injecting an amount that gave easily 100:1 S/N ratio on the polyamide, zilch on the aluminum clad. I would prefer to use a 30m column, but that’s a rather expensive experiment if I still don’t get the sensitivity. I have asked my Agilent service rep. for advice, I’ll see if he comes back with ideas. Thanks for the input all, I’ll add more info as I get it!

The 6890 EPC does not function well at less than 5 psi pressure. There is a fixed restrictor for the 6890 for the septum purge. With the column you are using for the analysis, the septum purge is the least restrictive path out of the inlet. If you are injecting is splitless mode, when the pressure spikes up during the injection the flow going into the injection port is shut off completely until the pressure drops to below the setpoint, the flow is turned on again, but may not be turned on until the splitless sampling time is over. Meanwhile, the sample is exiting the septum purge instead of going out the column, so the sensitivity is extemely low. I have seen this on other GCMS systems from other manufacturers.

As a rough guideline, columns less than 25 m length should be less than 0.25 mm id, preferable .20 or less.

The flow restrictor theory is an interesting one. I also have a PTV on this system (Gerstel). The PTV does not have a septum, so I can put that theory to the test.

My Agilent service rep. said that “the factory” has seen problems with polyamide columns that are inserted too far into the source. The presence of the column can effect the electronics on the lenses. His suggestion is to ramp all the lenses with a regular cap. Column then put the Al clad column in and do it again. I doubt this is the problem since the instrument tuned fine with the Al clad column installed, but I’ll give it a try. I probably wont re-post for another week though. Thanks for the input!

If the column is inserted too far, it can actually have the column end in the electron beam, so that the analytes do not enter the ionization zone. This is not an electronic problem, it a physical problem.

I doubt that you will have success with the Gerstel inlet with a 15 m 0.25 column, as the different inlet will not solve the problem of not enough back pressure for the EPC to function correctly.

I was at a site at one point where they were using a 15 m 0.25 column on a GC, then tried to transfer the method over to a mass spec using the same column. The sensitivity was very poor until I increased the split ratio from 10:1 to 50:1, when there was a dramatic increase in response since a small portion of the analyte was now reaching the column. The next morning they had a 30 m column and the analysis worked much better. This was a non-Agilent system, but the same principles apply. Dimensions of the column for MS are much more critical than for GC.

I did not intend to say that the column was in so far as to block the beam, but in far enough so that arcing between the column and the entrance nut might occur.

What do you think of this experiment. 1-Set up the column as before and verify poor response. 2-Connect a 15m x .25 piece of 5%phenyl, polyamide column on the front end. 3-Adjust the headpressure to get 32cm/sec (7.1psi at 100C). Obviously, I won’t be able to do the high temperature analysis, but this should determine if the problem lies in the column dimensions.

Seems like a very good way to nail down the problem. If you haven't talked to any of the former J&W guys now working at Agilent I would recommend contacting them also. They have a tremendous amount of information on issue like this.

What is the application that you are trying to do (what analytes are you trying to run)? Perhaps you can adjust column dimensions without dramatically affecting your analytes of interest? Running 15 meter x 0.25 mm ID columns on an MSD is very difficult to do for reasons explained above.

Your experiment sounds reasonable. That should help to determine if the presence of the aluminum-clad column in the source is causing a problem (I doubt that this is happening if the column is installed properly). You may also be able to make the system work by installing a restrictor at the front of your column -- something like 5 meters of deactivated fused silica tubing with an ID of 0.18 mm. This should help get your headpressure up to a reasonable amount to keep the inlet happy, however calculating actual column flow will now be very difficult due to the differences in pressure drop across the two pieces of tubing.

Best regards,
Jason Ellis
Technical Support
Agilent Technologies (J&W Site)

I would need to use bare fused silica in order to maintain high temp. capability. I'll try my first idea and take it from there. Thanks for your response.

More information about the application would be helpful, but here is some additional information to consider.

The temperatures that you may need to elute the analytes may be too high for the interface temperature limits of the 5973. I don't remember the limits for this instrument, but I know that continued operation above the low 300C range was not recommended. Having the interface 100C cooler than the column is not the ideal situation.

I would consider a metal column, with one of the no vent type interfaces. With this type of setup, the metal column would terminate outside the mass spec transfer line, with fused silica going into the source. I know that Restek makes metal columns that have temperature limits above 400C, and I am sure that the other column companies make similar columns.

Finally, is there significant value in running your samples on mass spec? There are some samples that can be run easily on GC-FID which simply don't work well on the mass spec due to the columns/conditions required. High temperature columns were primarily developed for applications such as waxes where pattern was important and the major peaks can be identified solely by retention time.

Good luck.