[asorocha]

Anyone knows what can cause drift on the signal of the iron lamp? Sometimes is too low, others is too high... I'm having an usual behavior for this lamp... What can I do to get a constant signal?

Thank you!

[marquis]

The list is long...

First is sample prep.  There can be many issues, but as a start, is the sample in a suitable acidic matrix?

I assume you are using a premix burner.  The first question is, when was the last time it was cleaned out?
If it was over a month, clean it. 

You also need to check the level of the waste liquid.  There needs to be a fair amount of liquid in it.  If not, you can get varying
pressure and all kinds of sample variations.

Next is acetylene pressure.   Acetylene is usually stored in a pressurized tank dissolved in acetone.  If the pressure is
too low, you can get bursts of acetone coming through.  This will give varying flame temp and ionization.

Your air source needs to be checked.  If it not suitable filtered, you can get sodium flecks, which will also mess things up.
You can see if this is an issue by looking at the air acetylene flame while aspirating no liquid.  It should have a rare (or no)
yellow fleck in it.  If it is continuous yellow, you have a sodium problem.

Your aspirator hose needs to be checked.  If it has a small leak, you can have problems.  Also, have you peaked the aspirator setting?

All of this is before checking burner head alignment, air acetylene mix ratios, lamp age and lamp current, and all kinds of sample
prep issues.

You might get started with the initial instrument checkout and then we can proceed to the last issues.

Good luck

[asorocha]

Thank you for your attention on this subject marquis.

Answering:
yes, the sample is in an acidic matrix.

yes, I already tried to clean and repeat analysis. have no effect on my results.

I will check the level of waste liquid!

Acetylene tank is brand new!

I will check for sodium problem too.

yes I peaked the aspirator settings... but I can do it again, just to check.

burner head: aligned. air acetylene mix ratio: as recommended. lamp age: 6 months. lamp current: as suggested by manufacturer (75% max current, to get more sensitivity)

I really appreciate your *delete me* Let me check the issues that needs to be checked and I'll come back to you later.

[asorocha]

Today tests: (after checking all the conditions, everything was OK)

blank - .003717
std 1ppm - .02277
std 3ppm - .05796
std 5ppm - .08685

DB - .02117
FS - .06110
std 1 - .04034
std 3 - .07648
std 5 - .1109

blank - .001775
std 1- .02277
std 3 - .05648
std 5 - .08670

conditions: I=12mA; flow: 65L/h; Abs max (std 1,6ppm): 0.057; burner height: 10mm

then:

I=6mA; burner height: 8mm

analysing standards 3 times in a row (aspirating blank only in the beginning): results higher each time (acceptable I think...)

analysing standards 3 times, but interrupting between each set (aspirating blank each time): similar results (R² between .99 and .995)

tomorrow I'll try with a new sample...

[marquis]

The first set of data (blank, 1ppm, 3ppm, and 5 ppm) and the third set of data look pretty good.

Don't recognize the DB and FS on the second set of data.  Can you fill me in?

Thanks,

[asorocha]

DB - Digestion Blank
FS - Fortified Sample

[asorocha]

The set after sample is strange right? It doesn't make any sense...
Can you understand why I'm having those results? Saturated atmosphere?? I have no idea... Even the digestion blank has a very high result (almost 1ppm); contamination? ok, possible. But the standards I'm using are always the same, it shouldn't, but they are; so I was expecting the same range of results. You agree?

[marquis]

Contamination is the first thing that comes to mind.

Your blank in the second set shouldn't be running that high. 

You may hear people talk about "matrix" differences.  Because the
composition of the sample is different from the standards, you can
get different readings. 

Try this.  After getting the strange readings, try running your regular
blank.  You should get a value of close to 0.000A.  If the value is higher,
running DI water or 1% nitric or whatever your blank is.  Keep aspirating
it until the value goes down to near zero.   You may have to aspirate the
water for five minutes. Then run your next sample and repeat. 

It maybe that your matrix is messing up the readings and you need to
get the burner cleaned out before running your next sample.

Some people try doing an autozero in this case.  Doesn't usually work
too well.  Your zero keeps drifting and you never know where your base
point is.

[marquis]

If it does turn out to be a matrix effect, you may need to
change your test method.  Many times, standard addition
methods are suggested so the standards and the samples
see the same matrix.

Good luck.

[asorocha]

Thank you one more time.

I'll try what you suggested, then I'll report my new results.

[Irlanur]

How long do you wait after you put on the lamp? they have to heat up!

[asorocha]

2h minimum... (it's a new one)

[asorocha]

New results. (analysis: standards; samples (DB and FS); Calibration blank (auto zero); standards; standards)

Conclusion: happening the opposite. I'm still having a drift of the signal but negative this time!

The set of results (2nd standards) in between, are too much lower.

I ran my standards and samples three times and only the first one had R^2>0,995... So, calibration blank between samples and 2nd reading of standards, makes a difference but it doesn't solve my problem, example:

.002356
.01339
.03522
.05848

DB .006782
FS .02722
.004359
.02966
.05035

.000548
.01247
.03181
.05297

look at this: (1st and 3rd reading of standards; 3rd repetition)
0,001576
0,02344
0,03765
0,05460

-0,005270
-0,00231
0,03277
0,04685

It seems that auto zero "cleans too much" both nebulizer and burner, but if I doesn't aspirate any water I "get more iron" than it is supposed to... strange isn't it? Maybe not, I need to find a compromise between this two situations.

This was a great *delete me* Thanks.

[marquis]

Please check your instrument settings.

You will have some degree of baseline drift in any instrument. 
When you adjust your data(by assuming the blank is zero and
adjusting your absorbance up or down, accordingly) your
answers are more consistent.  You might be able to take care
of this drift by zeroing more often.

Please check your integration time.  The longer you integrate,
the better your results are.  The Perkin Elmer instruments
usually default to a 0.5 sec integration time.  3 seconds is
much better.  You can go longer, if you have the sample
volume.

Hope this helps and good luck.

[marquis]

I think my last message was as clear as mud.

Here is the math.  We'll take your values and round to the nearest
0.001.  Then we will calculate the delta by subtracting the blank from
each sample.  Here is the data for your first data set:

sample         rounded      delta (sample -blank)

blank             0.002           n/a
1 ppm           0.013          0.011
3 ppm           0.035          0.033
5 ppm           0.058          0.056

Do this for each data set you have, using the blank
run with that data set.  Then look at your calibration
again.