 |
 |
Matt, Your Cl theory sounds reasonable to me. The Cl may be finding H liberated by big signals hitting the multiplier. With our new machine I thought we had a terrible HC problem on our extraction line. The machine would be nice and clean in the morning, then I would run an air shot and it would suddenly be full of hydrocarbons. We think that the large air shots were heating up the multiplier and driving off HC. Once we cleaned the multiplier up (see below) the problem went away. Multipliers seem to take a long time to hit their stride following venting, but you can hurry up the process. The thought is that as you hit a multiplier with a large signal, the electron cascade heats the dynodes up substantially and liberates loads of garbage. We have found that the best burn-in procedure is to raster a moderate beam over the multiplier overnight. On our Johnston we use a 1e-13 mol air shot, but you have to refresh it as it gets ionized away. In addition to cleaning it up, we have seen a 10-15% increase in sensitivity doing this. This technique isn't really ideal because you are liberating crud without pumping it away. I don't think we can get to H2 with our MAP 216. On the ETP multipliers on our Nu we rastered 500,000 cps (~5e-15 moles) of H2 overnight and it really helped. We just partially close the getter until you get the right signal size. Be careful because if it is dirty, lots of H2 is liberated and you can get a bad feedback going. I've even been warming up the multipliers by rastering 500,000 cps of H2 for 5 minutes before running every morning. They become more sensitive and more stable. I'm sure it shortens their lives, but they do seem happier. Andy ============ Andy Calvert ([log in to unmask]) Volcano Hazards Team U.S. Geological Survey 345 Middlefield Road, MS-937 Menlo Park, CA 94025 (650) 329-5276, fax -4664 On Feb 7, 2006, at 2:36 PM, matt heizler wrote: > Hi All, > > We have recently changed our multiplier from a Johnston to a > Blazers and due to this have higher than normal backgrounds as the > machine recovers from being vented. Despite fairly high > backgrounds, we proceeded to run some Fish Canyon sanidine to see > where we were at with respect to reproducibility. We consistently > obtained slightly less radiogenic yields, observed a slight age > variation that was correlated to signal size and MSWD values for 6 > analyses of about 2 to 3. > > We determined that we could subtract an additional absolute value > of about 4e-18 moles from mass 36 from our measured value and thus > increase our radiogenic yield values and obtain populations with > MSWD's of 1 or less. We refer to this as a phantom 36 correction. > > In an attempt to understand why our apparent 36 background was less > during a blank than during an analysis, be began looking at mass 37 > and mass 35 with a little more rigor. We found that for an air > analysis or for an unirradiated sanidine we consistently measured > negative 37 and 35 following blank correction. That is, we had less > 35 and 37 during analysis than during a blank run. In fact, running > any sample of sanidine, air, biotite or amphibole always returns at > negative 35 signal of about 4e-18 moles. > > We believe that it is no coincidence that the negative 35 signal > matches in magnitude the correction that we need to make to mass 36 > in order to get quality Fish Canyon sanidine populations. We > hypothesize that introduction of hydrogen from a sample or air > split is reacting with our chlorine backgrounds in the mass spec > such that 35 and 37 are reduced during analysis (compared to a > blank) and in turn we create HCl (thus the mass 36 interference) as > well as minor mass 38 (37Cl+H). > > Monitoring mass 2 during gas introduction shows that our getters > come to a consistent steady-state value that is non zero (3e-15 > moles) no matter what we put in the instrument. Much of the H is > pumped out by the ion pump after analysis. Also, closing the mass > spec getter with the machine static shows a pretty dramatic > hydrogen degassing of the mass spec that is readily pumped upon > opening the getter. Thus, our getters seem to be working normally. > > Our present Mass spec backgrounds are (moles): > > 40 - 2e-18 > 39 - 5e-19 > 38 - 5e-19 > 37 - 8e-18 > 36 - 2.5e-18 > 35 - 2.5e-17 > > As you can see, 35, 36, and 37 are higher than we'd like, but not > outrageously bad. We believe we have not observed this prior to > venting the machine because our backgrounds were in the E-19 range > and thus any introduction of hydrogen had nothing to react with to > cause our HCl interference. > > We would be curious to know if others have seen any similar > behavior to what is described above. If any one has at present the > unfortunate circumstance of relatively high backgrounds, we would > appreciate information on analysis of unirradiated samples (i.e. > negative 37) or any other runs to test for negative 35. Also, > anyone willing to take a look at hydrogen could help us evaluate > the quality of our getters. We think we have exhausted any > electronic issue that could cause this apparent behavior (i.e. > decay of signal following measurement of a large beam, signal non- > linearity, etc.). The chemistry of our problem seems to make sense. > Any feedback would be welcomed. > > Thanks, > > Matt > > > > > -- > Matt Heizler > NM Bureau of Geology and Mineral Resources > NM Tech > 801 Leroy Place > Socorro, NM 87801 > > Office 505-835-5244 > Argon Lab 505-835-5271 > Main Office 505-835-5420 > FAX 505-835-6333 > http://geoinfo.nmt.edu/staff/mheizler/home.html