Confusion Regarding "Surrogate Standard"

[mattmullaney]

Hello All:

I think that I have a handle on External and Internal Standardization (at this point, maybe I'm not even sure about Internal Standardization).

Please, can someone explain how "Surrogate Standardization" works? Please let me know where I am wrong...a surrogate standard compound is introduced into each sample to be analyzed at the same concentration level. Linearity of the surrogate standard compound, should it not be determined within the sample as a portion of method development/validation?

My thanks to you all, in advance...I'm really feeling lost and fairly embarrassed.

[mattmullaney]

Hello Again,

It occurs to me that I'm not giving enough information, maybe, or maybe in a proper way.

In this case, I've a separation of multiple analytes samples; I do not have standard material(s) to perform external standard calibrations for these analytes. I'm asked to use a "surrogate standard" material which does not appear in the samples to estimate the concentration(s) of the analytes of interest. The "surrogate standard" material is chemically and structurally similar to the analytes of interest.

It seems to me that I can prepare a standard curve of the "surrogate standard" in the fashion of external calibration while including an internal standard material along with the calibration set at a fixed concentration while that of the "surrogate standard" material is varied. This way, I can obtain relative response factors for the "surrogate" vs. the internal standard. Then, to the samples I can add the internal standard material at the same concentration level as in the calibration, and I'd be able to determine the relative response of each analyte vs. that of the internal standard material. I'd assume that the response factors of each analyte is the same as the "surrogate standard" material to estimate the concentration of each analyte.

I understand that this is not the typical use of surrogate that I've read about elsewhere, and I'll have to ensure that this is what is being "looked for" where I'm working.

Does this approach seem reasonable...given also that a reasonable choice for the internal standard is made?

My thanks again for your consideration, in advance.

[tom jupille]

The term "surrogate" is apparently used more in environmental analysis than anywhere else. I've never really made the distinction, but Mark Tracy's definition in this thread (http://www.chromforum.org/viewtopic.php?f=1&t=3461) clarifies it:

In the terms used by environmental analysts, "internal standard" is added to the sample immediately prior to the determination (e.g. HPLC) to correct for unstable instrument calibration. A more general definition is that it is added prior the the stage of the method where the uncontrolled variability is introduced. The term "surrogate" refers to a standard added at the beginning of the sample preparation to check for adequate recovery.

[mattmullaney]

Hi Tom,

Great to hear from you--I think that, where I'm working at present, "Surrogate Standard" is taking on a different meaning...as a Substitute for an Explicit Standard, in my way of thinking.

In this case, I've a number of analytes for which there literally are no standards that I can purchase to use as external calibrants. It's likely that they could be made...for an exorbitant cost perhaps, whom knows? Anyhow, there aren't any plans to go that route. We know what these analytes are using MS, and we can purchase a primary standard grade material to use that is similar to these analytes, instead of using a standard substance per each analyte.

My thanks for the thread link, I understand the definition Mr. Tracy kindly provided--this definition does jibe well with my own environmental analysis experience. I stand here now less confused.

Hopefully the process I suggest below will satisfy "the management" where I'm working! My thanks for your response, kind Sir.

Probably doesn't much matter...for this particular analysis...I don't think I could even properly introduce a surrogate standard in the fashion as defined by Mr. Tracy, as crazy as the whole thing is. I'm fairly certain that other readers will think I'm bonkers, which may be the case anyway...

[tangaloomaflyer]

You might be overthinking this.

Tom's definition is the best way to think about it. Internal standard is added to the extract prior to running on the instrument. When you use internal quantitation, any variation in instrument sensitivity is normalised by the internal standard response (sensitivity drops, analyte and internal standard responses drop, but the ratio of I.S. to analyte stays the same).

Surrogates are added at the start of preparing the sample for analysis. They are chemically similar to the analyte of interest, but should be rare in nature so they are unlikely to be found in the sample and they should be distinguishable from your analytes. That is why thy are called surrogates. They act like the analyte, but you know how much was spiked so you can calculate the recovery when the sample is analysed.

Two ways you can use surrogates:

Quantitate them as you would an analyte. Compare the result against the amount that was spiked to get a recovery. You can normalise your results according to the surrogate recovery if your method permits.

or, use the surrogate the same way you would an internal standard. By doing this you normalise any loss from prep. you can only do this is the surrogate closely matches the analyte chemically.

Most environmental labs would quantitate with internal standard and report the surrogate recoveries on a QC report.

[mattmullaney]

My thanks, tangaloomaflyer (and Tom Jupille, again). I agree completely with what you both say...I am less sure of how to exactly implement it in practice.

The samples are these polymer "plaques" prepared by mixing two liquid silicones together followed by two heating steps that serve to cure the mixture into a "solid" form. The plaques are extracted and the content of cyclic polydimethylcyclosiloxanes are separated...these range from D4 through roughly D25, similar to work Dennis Jenke has published.

The suggestion made to me was to use D4 or D5 as a surrogate to estimate the concentrations of all of the observed cyclic polydimethylcyclosiloxanes found, more or less. Seems to me, if either D4 or D5 can be in the sample to begin with, it's not such a great surrogate to begin with (if there is something I'm missing here, please correct me). Also...due to what goes on in the process of making these plaques...I really don't know if that if I spike D4 or D5 into the pre-cured mix to what degree it would survive...the plaques I've worked with have all been prepared by a third party.

I certainly could spike D4 or D5 into the extract...but then it's acting as an IS...not a surrogate. Very much agreed on this point.

I'm not sanguine about the accuracy of "semi-quantitation" of using D4 or D5 for compounds such as D15 through D25, either...my detection is to be FID, GC separation.

Please, see what you think, and thank you?

[Peter Apps]

Hi Matt

The preparation of the plaques is not part of "sample prep" as an analyst would understand it. The sample prep begins when you have the plaque and you dump it into the extraction solvent.

I agree that D$ and D5 cannot be used as surrogates or internal standards, because they are in the sample to begin with.

How many samples do you have to do, and how variable are they ? It might be best to do an external calibration with a select few of the cyclics (if you can get them as pure compounds) and then to estimate recovery by extracting some samples twice and comparing the peak area from the first and second extracts. Then apply the recovery to correct the results. Or use an exhaustive extraction.

Peter

[mattmullaney]

Hi Peter,

Quite agreed on the prep of the plaque...agreed also on the extraction of the plaque as the beginning of the analysis.

Agreed also on the suitability of D4 and D5 as either a surrogate (out of the question) or an IS.

Idea is to create a QC method (unfortunately?) so many samples going forward. Variability of these cyclics is high even within a single plaque...can be on the order of 20-30% absolute per the worst-case cyclic.

Undoubtedly...you're idea of external calibration is spot on. I'm getting some resistance on this idea. And (as usual, it seems) agreed on the approach you suggest for estimating recovery (extracting twice). The procedure of extraction seems to me to be exhaustive (work done before my time, and with great care, it seems to me)--reviewed the data for this...exhaustive, and yet does not deformulate the plaque, which is good news.

[Peter Apps]

Hi Matt

If your extraction is exhaustive then your main source of variability will be sample mass, and the various volumes along the extraction and concentration (if you do that) pathway. You can eliminate the effects of sample mass by always calculating back to actual mass, and a lot of the volume errors either by adding a standard into the extraction solvent (this could even be an alkane for example, all it is doing is tracking volume) or by check weighing at each stage. I don't see much advantage to using one standard to correct for sample prep and another one for instrument variability - one standard can do both jobs.

You could do this with an internal standard calibration - use e.g. alkanes, all you need is a constant ratio of response per unit mass of standard and analyte, your cyclics are chemically uniform so no problem there. The most likely source of problems at the intrument stage is inlet discrimination so you need ISs with a range of MWs similar to your cyclics'. I don't see any particular advantgaes to this approach compared to external calibration.

Peter

[mattmullaney]

Good Evening, Everyone,

Peter, my thanks for being a sensible fellow and a good teacher. Yes, my plan is to calculate these cyclics back to their mass in the original sample. Very much agreed with the idea that there is little advantage in using one standard to correct for prep and one for instrument variability...this is an argument I've also tried to fight, and lost (hence my weird ideas above).

Tangaloomaflyer's idea (and thank you again for that) appearing earlier in this thread is pretty much the same as the one Peter suggested immediately above...I can use the "surrogate" in this case as an internal standard...and it would be far more appropriate to have mulitple ISs to cover the range of analytes that will elute.

Thank you, and Tom Jupille, for your responses!

[James_Ball]

I am a little late seeing this post but it reminds me of some of the early work I did looking for Glucoraphanin in broccoli extracts. The reference works all used Sinigrin as a "surrogate standard" where they calibrated using Sinigrin and used that calibration curve to quantify Glucoraphanin. This was done because Sinigrin was much less expensive than Glucoraphanin(at that time we paid $800 for 10mg). We always used the actual Glucoraphanin to calibrate with, and we had different results than the people using Sinigrin. The response factor is close between the two compounds, but not exactly the same.

If you want a general idea of how much of a compound you have you can use a similar compound or "surrogate" standard to calculate the compound of interest, but if you are looking for a more accurate quantification you will need the actual compound to be sure of your values.

In this work, the surrogate was very different from how it is defined in Environmental work.

[mattmullaney]

Hi James_Ball,

Your description of the situation I'm in is apt...in my case, there aren't any readily available standards that I can purchase or "easily" obtain for the work I'm asked to do. The word "surrogate" seems to be used in the sense as "a substitute for" the standard(s) I don't have rather than as how an environmental chemist would understand a "surrogate standard"...as an internal standard added to the sample as early as possible in a sample preparation routine, as after Tom (noting Mark Tracy's definition) tangaloomaflyer and Peter all note above.

My thanks for your input as well. Interesting how word usage can cause such horrible confusion...at least to me, I have to say!

[Peter Apps]

Matt, you are not the only one confused by the "surrogate" as the environmental analysts use it. I could understand what the "surrogate" standard was doing in the analysis, but I could never understand why it was called a surrogate. From James' and your posts it seems that the word has been carried over from the substitution of one compound for another as a calibration standard - where the use of "surrogate" makes perfect sense.

Peter

[mattmullaney]

Peter,

Once again, my thanks. If nothing else, perhaps I'm not so crazy as I think I am...then again, perhaps not. My confusion this time, at least, is remedied!

[James_Ball]

Matt, you are not the only one confused by the "surrogate" as the environmental analysts use it. I could understand what the "surrogate" standard was doing in the analysis, but I could never understand why it was called a surrogate. From James' and your posts it seems that the word has been carried over from the substitution of one compound for another as a calibration standard - where the use of "surrogate" makes perfect sense.

Peter

Twenty years ago when we were doing EPA CLP contract work, they actually named the "surrogate" compounds correctly when they called them "system monitoring compounds (SMCs)". That always made more sense to me than the word Surrogate.

If you really want to be confused, start a discussion of what to call the lowest amount of analyte you can see when analyzing a sample