Advice for ICP-MS maintenance, adjustment, and repair

Adjusting the detector voltage

1. For a new detector FIRST RESET THE VOLTAGE TO +900 v (Pulse) and -1800 (Analog). If you don't you will kill the detector. When the detector is new you can use the optimization method in the software, which should result in values close to +900 and -1800 V.
2. To optimize the detector after it has been used for awhile. First record the pulse and analog voltages. Then increase the pulse voltage in steps of 50 V. The last increase that gives you a count rate increase of 10% is the increase to keep (let's say it was 100 V). Take 40% of this increase (40 V) and add it add the 40 V to your recorded pulse voltage (e.g., 900 + 40 = 940 v). Take 60% of this increase (60 V). and subtract it from your recorded analog voltage (e.g., -1800 - 60 = -1860 V).
3. I think you can do the same thing using the optimize program, just take its recommended pulse voltage increase and divide it 60/40 as shown above.
4. When the instrument was new the recommendation was to keep the voltage change equal, or for this example add 50 V to the pulse value and subtract 50 V from the analog value.
5. Logic, or at least philosophical symmetry, might suggest the split should be 67/33 (= 1800/900) rather than 60/40 or 50/50, but that's what the pro says.

Cleaning the lens

From: "Jarrett, Jeff" Subject: Re: Auto Lens Cleaning To: PLASMACHEM-L@LISTSERV.SYR.EDU (edited). We clean the ion lens of our ELAN ICPMS instruments in a 5% (v/v) nitric / 5% (v/v) hydrochloric acid solution for 10 minutes in an ultrasonic bath (as was suggested to us by PerkinElmer). This usually discolors the lens, but doesn't seem to affect it's reusability. Still, I wonder how many times we should try to reuse the lenses. So far, we resuse them until we see that the optimized lens voltages don't drop back to the 4 - 7V range when they are first re-installed. Sometimes rust-colored spots start to develop on the inside of the lens. At that point, we discard it. We clean the lens whenever the optimum voltage for In in the autolens optimization gets up around 9V. At that point, we begin to see degradation in the precision of measurements of the lowest mass analytes (i.e. 9Be). The lens in the ELAN 6000 and 6100 instruments are pretty easy to clean, and we keep a clean one in the drawer to swap out quickly. We usually do the cleaning at the end of the day, though, and let the vacuum pump down overnight. If we try to run the instrument without an overnight pump-down time, our QCs tend to not pass. We have had more luck with programming in a 12 hour delay time before the start of the first run, though. In the ELAN 6100 DRC, the lens is more difficult to get out and put back because the dynamic reaction cell partially blocks the way to the screw on the lower side. It takes a little longer due to that, but we clean the lens the same way.

 

From: "David Jones" Subject: Re: Auto Lens Cleaning To: PLASMACHEM-L@LISTSERV.SYR.EDU (edited). I'm doing mostly mid to high range masses and don't take the lens out until it reaches 10.5 or so. I have noticed the corrosion issue on old lenses and wanted to avoid it so I switched to the Microgrit powder. And I will have to find out how long I can polish them before the inner ridges are worn down too much. I was also shown a fine bristle fibre glass electronics cleaning brush which seemed to have a very good effect on the inside surface but haven't used one yet. With the DRC I was shown that removing the interface by the four Allen bolts is the quickest way of getting it out. Just be careful of the O-rings but with somebody to hold the interface for you once it's unscrewed removing the lens is almost like changing a light bulb, and no awkward positioning of the ball driver required. And the great thing with the DRC vacuum system is that you really can be running in 2 hrs with no problems.

Lighting the torch

This section stemmed from a problem where the torch lit but extinguished within a few seconds. Replacement of the load coil and much fiddling with the torch, injector, spark rod, load coil centering, checking for gas line leaks, etc. had no effect. A call to Rich Barrows, our PE service engineer, suggested the following checks:

1. The mass flow controller follows the following program during the torch ingnition sequence: 1 lpm for ~75 seconds, off for 15 seconds, then to the current DAC value after the torch has lit. If the nebulizer flow controller has screwed up, it will give the wrong flow values. In particular, if the gas flow goes up to 2 lpm the torch will be blown out. Our test was fine.
2. For the RF generator, the plate voltage should be >3900 V when the torch lights. During operation the plate voltage should be lower (ours start voltage was ~3980 V and running voltage was ~3175 V), the plate current should be >0.5 mA (ours was ~0.52 mA), filament voltage should be 6 to 7 V (ours was 6.25 V), and the grid current should be >100 mA (ours 118 mA). These tests were therefore fine.
Rich suggested reducing the DAC nebulizer flow value to 0.2 lpm for startup, and this worked; the torch lit as normal and I was then able to reset the nebulizer flow to the usual 0.95 lpm or so. The instrument ran without a hitch for the rest of the day.
3. There are two likely problems according to Rich. Most likely is that the RF tube is nearing the end of its life, which typically shows signs of age after ~1500 hours. Ours was at 1585 hours, so this is a possibility. The other possibility he suggested is no longer completely clear in my memory; it had to do with a problem with the voltage taps on the RF transformer. He suggested moving the leads from the current taps (presumably 208 V) to the next lower voltage (200 V?). I think he said if the system then worked fine, then the taps were incorrectly selected. Since instrument has been working fine for 2.5 years this seems unlikely to me. Another possibility that comes to my mind is that the impedance matching electronics may be screwed up. There are automatic adjustments made during startup especially, and if these are far off the torch will not light or will not stay lit.

And from our own Annals of Disaster: After a major power failure we had damage to the computer, the mass flow controller box, and the plasma RF generator. For the last, the torch was hard to light, and later would not light at all. There was no sound of the high voltage turning on, and the READY light blinked off for a second when the torch was supposed to be lighting. I think the sound of the torch sparker could be heard. The problem was a dead controller board within the RF power supply, which had to be replaced. Our PE technician thought the problem was in two transistors attached to big heat sinks, but there was no outward sign of any damage.

RF power level

Our service engineer says that the plasma power for any but organic matricies is usually 1100 to 1150 W. At 1150 W we usually see ~2% CeO/Ce. At 1300 W we see typically 1% CeO/Ce at good sensitivity, though naturally M2+ ratios are higher. However, our engineer says that the RF tube life expectancy drops with increasing plasma power. RF tubes can usually run for over 1500 hours at 1150 W or so, though our engineer has seen them fail with as little as 1000 hours and has also seen them running fine at 5000 hours.

Poor precision

Poor precision can result from peristaltic pump pulsation. We suggest:

1. Using a 0.25 mm diameter tube to the nebulizer.
2. Not cutting short the pump tubing on the nebulizer side.
3. Adjusting the peristaltic pump roller pressure for the sample tube while looking at the X-Y adjustment method graph. Poor precision will be easily seen in a very spiky sensitivity curve.

Autosampler does not go to the correct position

The AS-91 autosampler starts not returning to the rinse or standby positions correctly. Dirty worm drive. Remove covers and give the worm and sliders all a clean and relube. Other possible problems: The proximity sensors (little black things) for the X, Y, or Z search for the zero position with each full probe movement. If they don't find the zero position, there can be positioning problems, typically causing the motor to jam when the X, Y, or Z positioner bottoms out. The proximity sensors or circuitry may be broken, or the metal plates that the sensors look for may be out of alignment. The plates can be adjusted a little, and so can the sensors. Such adjustment can also help get the sampling probe to be centered on the sample tubes. Bad motors or broken cables are another possibility. Check all cables for brakes by inspection and also by testing wire continuity. Don't forget the ribbon cable to the mother board.