MS Tip - Evaluating multipliers to determine when to replace

Mass Spec Tips is a collection of tips relating to the operation and service of mass spectrometers. Some of these Tips have been reported in our newsletter, "The Mass Spec Source." If you have a question relating to the operation or service of mass spectrometers or would like to add your comments to the below tip and would like to have it included in this forum, please contact us. Input to this database of MS Tips is needed from our readers to complete this section.

Answer 1

We posed these questions in our last newsletter because they are often asked by mass spec users. The determination of the gain or proper operation of the electron multiplier may be quite difficult to determine. Factors such as the cleanliness of the mass spec source, filament operation or alignment and vacuum system performance can all erroneously convince the operator that the multiplier is in need of replacement. On many of the early magnetic sector mass spectrometers, a sample was introduced and the ion beam was defocused so that only a few ions per second would strike the multiplier. These single ions were then detected on the oscillographic recorder and the peak area in coulombs was calculated. From the charge ratio, the multiplier gain could be determined. This method was accurate but cumbersome to use. However, today oscillographic recorders are rare and defocusing of the ion beam to obtain single ions may be difficult or impossible to achieve on modern quadrupole mass spectrometers. Most users determine the performance and usefulness of the multiplier based on experience and how old it is. Normally if the mass spectrometer is not performing up to specs, and the multiplier is more than one year old, the operator will replace the multiplier. On most mass spectrometers which use autotuning programs to adjust electron multiplier voltage, the tune voltage will increase with age. When the mass spectrometer begins to tune in excess of 2800 volts, it is time to replace the multiplier. More accurate methodology is needed in order to determine multiplier performance and to eliminate errors in judging multiplier performance due to the performance of other mass spectrometer components. This is the reason that we proposed this question to our readers.

The following responses were received from our readers in the last month. If you have any additional comments or suggestions, we would like to hear from you.

Answer 2

We use a Hewlett Packard GC 5890, with a Hewlett Packard MS-5970. Type of samples run are wastewater, soils, and liquid waste.

Answer 3

In the July 1994 issue of the Mass Spec Source you requested when a multiplier needs replacement. Coincidentally, just yesterday I wrote a requisition to you to order a new multiplier for my INCOS-50B.

Older Finnigan instruments had a cup or shield connection on the back flange, as well as a scope display, and the way to determine multiplier gain was well laid out in the manual. The newer instruments, such as the INCOS series, however, do not have this ability and the multiplier gain must be determined in a round about way. Frankly, there is no way of getting a numerical value for the gain, but one can see when the multiplier is getting sick enough such that it needs to be replaced.

Over time the multiplier will require more and more voltage to perform it's job. That is the first indication of multiplier wear. A new multiplier will work fine at 1000 volts and an old one may need over 2000 to do the same job. Furthermore, it seems, at least in my experience, that one loses high mass sensitivity more than low mass sensitivity. For that reason, in order to keep spectra consistent, I consider 1800-2000 volts my upper limit. If I have to use more than that to get decent sensitivity it is time to replace.

Of course, there are several things that can cause a loss of sensitivity, from a dirty source to something as simple as a leaky septum causing sample to be lost. The way to convict the multiplier of the crime is to determine the noise level. All multipliers have noise spikes, but, like sensitivity, you need to go to higher and higher voltages to see the noise. In the INCOS-50 series of instruments go to MTUN and with the multiplier and filament off set the zero to give a few noise spikes. Turn on the multiplier. Starting at a low voltage incrementally increase the voltage and see at what voltage the spiking gets worse. A new multiplier might do that at as low a voltage as 1400 volts or so. (Use the ES command in MTUN to set the voltage.) You might have to use the U command to reset the multipliers upper voltage limit so you can go to higher levels. In the case of the multiplier in my instrument now, the one in need of replacement, 2600 volts gave very little appreciable increase in noise level.

Answer 4

Before assuming that the multiplier is bad, check the gain by following the procedure described either in the 400 manual or in the Service Short from 1/18/91, which deals with performing a multiplier gain check. If the multiplier appears to be capable of reaching a gain of 105, this may be enough to allow you to continue to run samples. However, excessive noise may cause dropouts to appear in either the total ion chromatogram or the individual mass chromatograms.

Sources other than the multiplier may be causing the noise. An ion volume which needs cleaning can greatly reduce the signal to noise ratio. Just as detrimental, is an air leak of enough magnitude to impede either the sample flow, or the flow of ions from the ionizer to the multiplier. Check to see that the chamber pressure is at a normal level. Typically, for a singularly pumped system, this pressure should be in the low 10-5 torr range or below. For a differentially pumped system, the front, or source chamber, will normally be in the low 10-5 range, while the rear, or analyzer chamber, will be approximately an order of magnitude lower. In general, the lower the pressure, the less obstruction there will be to the ion flow through the quadrupole resulting in better sensitivity.

Another factor which could make the multiplier appear noisy, if you are trying to observe peaks produced from the cal gas, is the amount of cal gas actually being ionized. Check to see that there is a pressure increase of approximately 1 to 2 to the 10-6 torr in the source chamber when cal gas is turned on. If your chamber pressure is normally in the 10-5 range, you may have difficulty determining if this increase occurs. Either turn down the GC column flow or remove the column completely, capping off the glass interface with a blank ferrule (included in the spare parts kit). This should lower the chamber pressure to the 10-6 range at least, making the cal gas flow easier to observe.

Even though the cal gas flow may be correct, it should be confirmed that an air leak does not exist, with either the cal gas on or off. If the pressure changes correctly when the cal gas solenoid is toggled, but there is more air than cal gas being delivered, the signal to noise ratio will be low. If a leak is detected, take a logical approach to resolving it by observing peaks in the 10 to 40 mass range while squirting methanol on various joints. By toggling the divert and cal gas solenoids separately, you shouldn't let the glass interface intimidate you, but be cautious not to over tighten any of the fittings. Always hold the fitting firmly with the appropriate size wrench while tightening the nut against it! if new ferrules are indicated, replace them.

If a GC column is installed, be aware of the flow rate. There is a limit to how much flow the system can handle; thus the reason for a jet separator in cases where the flow exceeds 1 to 2 mls./min. Most GC manufacturers provide a test column with the instrument which may be only a few meters long. This column is designed to be used with a flame detector and not a mass spectrometer. Although the vacuum gauge may indicate a reasonable operating pressure; within the mid 10-5 range, most vacuum gauge controllers are calibrated with nitrogen and therefore may display an inaccurate reading. Excessive carrier gas flow from a short column or one with too large an I.D., may cause so much background noise that the baseline on the oscilloscope will actually rise when the multiplier voltage is increased. Too high of a chamber pressure, which also creates background noise, will cause this baseline rise to occur as well. A multiplier that exhibits this symptom under normal conditions is most likely contaminated and should be replaced.

Answer 5

We have a HP 5890 GC with a J&W DB5-MS column interfaced to a 5971A MS running under HP ChemStation software. We determined that the multiplier (detector) needed replacing from information directly off the Standard Autotune output and general machine performance.

The autotune output yields a large amount of information pertinent to multiplier performance. The values for EMVolts, X-Ray and the shape of the 502 PFTBA peak are what we used as diagnostics. Originally, we noticed that the EMVolts started to rise well above our normal operating range of 1700 to 2000. In fact it reached 3000, and stayed there. At this point, we assumed that our source was dirty, so we cleaned the source. This had no effect. Tunes still produced EMVolts at or near 3000. Our second clue was that the shape of the 502 PFTBA peak was fragmented and the abundance for the 69 peak was very low; sometimes 70% - 80% of normal. Our third clue was the value for the X-Ray skyrocketed up to 130.

In conjunction with the values from the tune, was how the machine was operating. Our quality control abundance's and the signal to noise ratio both dropped by 50%.

After replacing the multiplier, all problems were solved and the machine is back to running as it normally does.

The gain of an electron multiplier can most simply be defined as the ratio of the output electron current to the input ion current at a specified second dynode voltage. The output current is not directly measurable without auxiliary test equipment, but it can be inferred from the height of an ion in SIM mode. The input ion current, on the other hand, cannot be measured on most mass spectrometers short of instrument modification, and the latter is not recommended. As mentioned by Barry Austern in the September, 1994 issue of the Mass Spec Source, older machines such as the Finnigan 3200/3300 could be used to measure the input ion current, as they were equipped with a normally grounded x-ray shield directly on axis with the quadrupole rods. One could simply turn the multiplier voltage off and transfer the preamp lead from the multiplier signal output to the x-ray shield.

Various methods have been proposed over the years to circumvent the problem of not being able to measure the input ion current, and thereby determine the multiplier gain. Even if I can determine the gain of a multiplier, I would not replace a multiplier based solely on its gain. Rather I base the replacement decision upon whether or not routine analyses which could be performed in the past, can still be performed. I stress the word routine since many factors can effect instrument sensitivity.

In a research facility such as ours there are not many routine analyses. Therefore I have devised my own tests which are based on the functional tests performed by the instrument manufacturers technicians at the point of initial installation. In order to run the tests, I use Methyl Stearate in Electron Impact (EI) mode (M+=298) and Positive Chemical Ionization (PCI) mode with Methane ([M+H] + = 299), and Octafluoronaph-thalene in Negative Chemical Ionization (NCI) mode (M- = 272). I make sure that the tests are performed with a clean source on a well pumped down machine using the same column (DB-1, 12 meter, 0.2 mm ID, 0.33 u film), temperatures, flow rates, and acquisition parameters each time. Then, I test to determine two things:

1) What is the minimum amount of material I need to inject to produce a signal to noise ratio of 10:1 when monitoring the molecular ion (or M+H ion in PCI)? (I inject progressively smaller amounts of material while varying the multiplier voltage until I determine this amount/voltage point).

2) By using 1000 times more material and the same multiplier voltage as in 1 above, what is the relative standard deviation of the M+1/M, M+2/M and M+3/M ratios? This latter test is important to a research lab such as ours because much of our work involves ratio measurement using stable isotopes. The first test may actually suffice for most labs.

When a multiplier is new the first test yields low amounts of material at low voltage. Then, as the multiplier ages, both will rise.

These tests will yield numbers which will provide comparative ratios over a period of time, and therefore help determine continued effectiveness of an electron multiplier. However, my ultimate criteria is still whether or not routine analyses can be performed.

All the later VG 70/ZAB's and our MAT CH5 mass spectrometer have Faraday cups at the end of the ion beam path. The signal can be measured there and compared to the output of the multiplier. On older 70/ZAB's and MAT instruments the signal could be measured at either the first dynode of the multiplier or through the resistor string with the voltage on.

Another method is based on measuring the signal strength of single ion events employing an oscilloscope of known display response and gain. The gain can be measured on any mass spectrometer with an analog scope display and adjustable frequency response whose detector pre-amplifier's input impedance is known. It does not require any special connections or modifications to the detector.

Multipliers must be replaced when they need replacing. Generally we replace conventional venetian blind or discrete dynode multipliers when they saturate at 107 gain. Multipliers such as the old Kramer 17-dynode (FK7 or FK17) as well as the EMI 119 can achieve gains of >109 when new. The gain of the photodetector systems on the VG 70-SE4F and 70-VSE in the Lab have not appreciably changed in 7 and 5 years respectively. This is the advantage of a photodetector system, the EMI photomultipliers are sealed systems and are in a high vacuum environment as well.

MS TIP No. 18 - What techniques or methods do you use to determine if the electron multiplier is in need of replacement? How do you determine multiplier gain? What is an acceptable signal to noise level?

Answer 1

Answer 2

Answer 3

Answer 4

Answer 5