GC-MS Screen for the Presence of Melamine, Ammeline, Ammelide and Cyanuric Acid
(Version 2.1)
This method for Melamine, Ammeline, Ammelide and Cyanuric Acid should be regarded as interim. Because of the need to rapidly provide this information, the method has not undergone the rigorous internal and external validation required for an official method. The performance of the method may change when different equipment and supplies are used or when different sample matrices are encountered. The user should validate the performance of the method in their laboratory and pay particular attention to the recommended quality control elements. Users should document with their results the version or date of the method used.
PURPOSE:
This procedure was developed by PRLNW and FCC to screen various matrices for the presence of melamine and some related compounds at the established minimum reporting level (MRL) of 10 µg/g and above using gas chromatography/mass spectrometry. Samples are extracted using a mixture of acetonitrile/water/diethylamine and the analytes are subsequently converted to trimethylsilyl derivatives for analysis.
SCOPE:
This procedure has been evaluated using dry protein materials (wheat gluten, rice protein, corn gluten, and soy protein), wet and dry pet foods, and dry animal feeds. It is anticipated that the method will also be applicable to a variety of other matrices.
RESPONSIBILITY:
It is the responsibility of the analyst to note any modifications to or deviations from this procedure in the documentation associated with the analysis.
DEFINITIONS AND ACRONYMS:
TMS – trimethylsilyl
DEA – diethylamine
SAFETY CONSIDERATIONS:
Accepted safety measures should be employed when working with chemicals and pressurized gases. It is advisable to work in a fume hood whenever the procedure calls for the use of diethylamine and pyridine. Use caution with diethylamine as it is a volatile strong base and can cause chemical burns.
EQUIPMENT AND SUPPLIES:
REAGENTS AND STANDARDS:
Prepare a solution which consists of 10 parts (by volume) diethylamine, 40 parts water and 50 parts acetonitrile. Store in the dark. The solution turns yellow with time.
For use as an internal standard. Prepare a stock solution at 1.0 mg/mL in pyridine.
For adding to sample extracts, dilute the stock solution in pyridine to a concentration of 5.0 µg/mL.
Use of 2,6-Diamino-4-chloropyrimidine (DACP) as an internal standard is still undergoing evaluation. It is up to each laboratory to verify that their internal standard does not contain any of the analytes as impurities. The Aldrich standard has been found to be free of any analyte impurities.
Prepare a stock solution of 1.0 mg/mL in a mixture of 20 / 80 : DEA / H2O (v/v)
Prepare a stock solution of 1.0 mg/mL in a mixture of 20 / 80 : DEA / H2O (v/v)
Brief sonication may be required to solubilize this standard.
Prepare a stock solution of 1.0 mg/mL in a mixture of 20 / 80 : DEA / H2O (v/v)
A modest amount of sonication may be required to solubilize this standard.
Prepare a stock solution of 1.0 mg/mL in a mixture of 20 / 80 : DEA / H2O (v/v)
Store stock solutions of standards in the refrigerator to retard hydrolysis. It has not been established how rapidly the solutions degrade but the potential does exist. Stock standards have sat on the benchtop at room temperature for three weeks and shown no evidence of degradation.
QC ELEMENTS:
A 1:1 (v/v) mixture of Sylon BFT and pyridine should be run at the onset of each analysis and then occasionally throughout the analysis to show that there is no carryover.
A method blank which consists of 20 mL of extraction solvent taken through the entire procedure including the addition of the internal standard should be evaluated to make sure that there is no contamination from the reagents, the nylon filter or the containers.
Condition the system at the beginning of each sample set by making two injections of the high standard (see below). These injections will not be used for semi-quantitation.
A low standard, which consists of each analyte at 0.10 µg/mL, should be analyzed at the beginning of the sample set to show that the necessary sensitivity is being attained by the instrument.
A high standard at 1.0 µg/mL of each analyte is analyzed at the beginning of the sequence and after all samples in the batch have been injected to provide a basis for semi-quantitative evaluation and to demonstrate whether the amount of drift during the analysis of the set of samples is tolerable.
A control sample which is representative of the type of samples which are being analyzed is fortified with each analyte (viz. melamine, ammelide, ammeline and cyanuric acid) at a level of 10 µg/g. Analysis of this spiked control sample must indicate that the compounds are present which serves to demonstrate effective system performance at the desired reporting level. See Sample Fortification, below.
PROCEDURE:
This procedure should be used with the GC operating in the splitless mode for greater sensitivity.
A. Extraction Procedure
Weigh out approximately 0.5 g of a representative portion of the sample into a 50mL polypropylene centrifuge tube.
Add 20 mL of extraction solvent (10:40:50 DEA : H2O : Acetonitrile).
Mix well to thoroughly wet the entire sample.
Sonicate for 30 minutes.
Centrifuge for 10 minutes at 5000 g (or better).
Filter a portion of the supernatant using 0.45µm nylon filter discs (a two-stage or molecular weight cutoff filter may be used for difficult extracts). For example, a 2 mL filtered portion will allow 200 µL for the derivatization step and additional filtrate in the event that the derivatization needs to be repeated or if further work is necessary.
B. TMS-Derivatives
Sample Extracts:
Transfer 200 µL of filtrate from Step A to a GC vial.
Note: A smaller aliquot may be used provided that the necessary sensitivity level (10 µg/g of sample) is achieved. Reducing the amount of matrix present improves the general performance of the evaporation/derivatization step and saves wear and tear on the instrument.
Evaporate to dryness at 70°C (a low flow stream of dry air or nitrogen may be used).
Note: Taking the filtrate completely to dryness is a critical step in the derivatization process. The presence of water prevents formation of TMS derivatives of the analytes. If the internal standard response is much lower than usual (less than 30%), there may have been problems associated with the derivatization step. In addition, if the vial warms significantly to the touch after addition of the derivatization reagents, residual water was present and a new aliquot of filtrate must be prepared.
Add 200 µL pyridine.
Add 200 µL Sylon BFT.
Add 100 µL of the internal standard solution at 5.0 µg/mL in pyridine. This produces a concentration in the extract of 1.0 µg/mL. If you are not using the internal standard, then add 100 µL of pyridine.
Shake well or vortex to mix.
Incubate at 70oC for 45 minutes.
Ready for Injection.
Note: If insoluble material is observed at the bottom of the vial after 45 minute incubation, transfer liquid portion to another GC vial or filter before analysis.
C. Instrument Parameters
GC Conditions:
Column 30m DB-5MS 5% phenyl 95% dimethyl-polysiloxaneInlet Temperature 280oC
Detector Temperature 290oC
Injection Mode Splitless
Injection Volume 1 µL
Carrier Gas Flow He at 35 cm/sec (constant flow)
Oven Program 75oC (hold 1 minute) to 320oC at 15oC/minute (hold 2.67 min) for a total run time of 20 min.
Note: Alternate GC conditions may be used provided that adequate resolution is obtained between the target analytes. Any such deviations from the method must be noted in corresponding documentation. With small volume liners, some peak splitting has been observed under the above conditions. Using a higher starting temperature (100oC) alleviated the problem. To help overcome interferences, significant additional resolution may be obtained by decreasing the ramp to 4oC/min over the interval 150oC to 200oC, which will shift the retention times.
MS Conditions (Full Scan Mode):
Tune Autotune (to maximize sensitivity across mass range)
A +306V multiplier bump may be added after Autotuning
If the sensitivity which is required to detect analyte spikes at 10 µg/g in the matrix cannot be achieved in full scan mode, use selected-ion monitoring (SIM) parameters below.
MS Conditions (SIM Mode): Select three or four ions to track: M, M+2, M-15 and another
Group |
Start Time a |
Ions b | |||||
M c |
M + 1 |
M + 2 |
M - 15 |
Other |
Other | ||
Urea / Biuret from |
6 |
276 |
|
|
189 |
171
|
|
Cyanuric Acid |
9 |
345 |
346 |
347 |
330 |
188 |
|
Ammelide |
9.7 |
344 |
345 |
346 |
329 |
286 |
198 |
DACP (internal std.) |
10.18 |
288 |
289 |
290 |
273 |
275 |
237 |
Ammeline |
10.5 |
343 |
344 |
345 |
328 |
285 |
198 |
Melamine |
11 |
342 |
343 |
344 |
327 |
285 |
197 |
a Start Times may need to be adjusted based upon the retention time of standards on your system. |
D. Peak Identification and Results
The approximate retention times of the tri-TMS derivatives are as follows (minutes):
Cyanuric Acid 9.2
Ammelide 10.0
DACP 10.3 (di-TMS derivative)
Ammeline 10.7
Melamine 11.2
These need to be confirmed on your system and the parameters above adjusted to fit. Figure 1 is a standard chromatogram showing the cyanuric acid tri-TMS derivative (9.2 minutes), ammelide tri-TMS derivative (10.0 minutes), ammeline tri-TMS derivative (10.7 minutes) and melamine tri-TMS derivative (11.2 minutes). Figures 2 through 5 represent the mass spectra of each of these compounds. Figure 6 represents the mass spectrum of 2,6-diamino-4-chloropyrimidine di-TMS derivative.
When full scan mode is used and the matrix is even somewhat complicated, it is usually necessary to examine tracings of individual ions (extracted ion chromatograms) to observe the analytes. The extracted ion chromatograms may be generated for each analyte (see SIM table above for appropriate ions) which will help simplify the process of confirming positive or negative results (i.e. pulling things out of the grass). Ideally, each of these analytes and the internal standard can be entered into the data analysis portion of the Chemstation software as a target compound. By entering the appropriate retention time, ion ratios and peak area for each compound, the software can calculate sample results automatically. Results can be reviewed using the “QEdit” view which also provides an excellent visual display for documentation of sample results (see Figure 7 for a sample QEdit report for melamine).
In full scan mode, the criteria for identification of target analytes include the agreement of the retention times with those of the standards to within 0.05 minute. Also, the mass spectra need to correspond to those of the standards with no significant peaks absent. There may be additional peaks present due to overlap with other components but this should be examined carefully.
In SIM mode, the criteria for identification of target analytes include the agreement of the retention times with those of standards to within 0.05 minute. Mass spectral confirmation of an analyte is based upon the ratios of the integrated areas for selected ions to the integrated area of the most abundant of the selected ions being tracked. The criterion is that each ratio (as a percentage) corresponds to that observed for a standard to within 10 units. Using melamine as an example, track the ions at 342 (M), 327 (M-15) and 343 (M+1), use the data in the table above as representative of a standard. For a peak nominally identified as melamine based upon retention time, the ratio of the area of peak (at 11.2 min) from m/z = 342 (M) to that of the area of the peak at m/z = 327 (M-15) needs to be between 44% and 64% while that of the ratio of the peak at m/z = 343 (M+1) to that of the area of the peak at m/z = 327 (M-15) needs to be between 6% and 26%.
E. Sample Fortification and Mixed Standard Preparation
A Mixed Standard Spiking solution is prepared by mixing portions of the stock standard solutions with 10/40/50 DEA/Water/Acetonitrile to create a solution which is 100 µg/mL of each analyte.
Spike Preparation
Weigh a 0.5 g portion of the matrix of interest into a 50 mL centrifuge tube. A previously analyzed, representative “blank” matrix may be used. If a representative matrix is not available, select one of the samples to be analyzed to use for spiking.
Add 50 µL (Low Spike) of Mixed Standard Spiking solution directly to the representative control sample and proceed with the method. The Low Spike must be observed to declare negatives as less than 10 µg/g. If it is not then negative samples will need to be re-analyzed.
A High Spike may also be prepared by adding 250 µL of Mixed Standard Spiking to another 0.5 g portion of the control. The High Spike provides an additional check in the event the Low Spike is not observed. The absence of the High Spike would serve as an indication of a major problem either with the sample preparation or the instrument.
High Standard
Dilute the Mixed Standard Spiking Solution to 10 µg/mL. Place 50 µL of this solution in a vial and take it through the derivatization reaction including the addition of the internal standard. This produces a High Standard at a concentration of 1.0 µg/mL of each analyte.
Low Standard
Dilute the Mixed Standard Spiking Solution to 1 µg/mL. Place 50 µL of this solution in a vial and take it through the derivatization reaction including the addition of the internal standard. This produces a Low Standard at a concentration of 0.10 µg/mL of each analyte.
F. Reporting
In the event that the analytes were observed in the representative control sample that was fortified at 10 µg/g, and no analyte signals were observed in the samples at levels which approach those in the fortified control, then the samples are not contaminated with melamine, ammeline, ammelide or cyanuric acid at levels in excess of 10 µg/g.
When it is close (within 50%), consider preparing additional portions from those samples along with a couple of additional fortified controls.
If it is clear that one or more of the analytes are present in the samples (based on the identification criteria above) and at levels in excess of 10 µg/g, then a semi-quantitative estimate may be obtained by comparing analyte responses to those from standards as described below.
G. Obtaining a Semi-quantitative Estimate of Target Analytes
Choose an ion to use for estimating the amount of the target analyte (usually the molecular ion). Using the High Standard (1.0 µg/mL) which was run at the beginning of the set of samples and the High Standard that was run at the end of the set of samples, calculate the average peak area for the selected ion. Apply the following formula:
Concentration in Derivatized Test Sample Extract (µg/mL) = Area (Test Sample) / Average Area (Standard) X 1.0 µg/mL.
Then, Concentration in Sample (µg/g) = Concentration in Derivatized Test Sample Extract (µg/mL) X 0.50 mL X 20.0 mL / 0.20 mL X 1 / Sample Wt. (g)
If the internal standard areas are very different (> 20%) between the test sample and the standard then adjust the area of each analyte response by dividing by the integrated area of the internal standard (from its molecular ion). Use these ratios in place of the integrated areas above. Consider re-analysis if the difference is extremely large since that may indicate a partially clogged syringe or other instrumental problem.
Finally, if the signal is more than 25 X larger than the signal from the standard, then prepare a new extract using 0.25 g of sample and 40 mL of the extraction solvent since the solubility limit for ammeline and ammelide in 10/40/50 DEA/water/acetonitrile is being approached.
As a general rule-of-thumb, the closer the analyte signal from the unknown is to the analyte signal from the standard, the better the estimate of the unknown concentration. So if more accurate numbers are required, prepare additional standards and re-analyze the samples in question.
H. Contacts for Further Technical Information
Jonathan Litzau, Forensic
Email: jonathan.litzau@fda.hhs.gov
Greg Mercer, Pacific Regional Laboratory – Northwest (PRL-NW),
Email: greg.mercer@fda.hhs.gov (Primary contact for “QEdit” reporting option)
Kevin Mulligan, Forensic
Email: kevin.mulligan@fda.hhs.gov
Table 1: Feasibility Study with Analytes Spiked At and Near the Action Level
Average Percent Recovery Matrix Spiking Level( µg/g) Cyanuric Acid Ammelide Ammeline Melamine Wheat Gluten (a) 10 84 90 86 79 50 87 88 86 74 Rice Protein 10 98 101 96 96 50 107 108 108 101 Wet Pet Food 10 106 103 114 105 50 107 107 111 106 Corn Gluten 10 139 166 172 152 50 103 103 98 90 Soybean Meal 10 120 118 164 152 50 97 90 90 88 Dry Cat Food 10 158 142 134 126 50 81 82 74 77 Notes: (a) 0.5 g extracted into 20 mL of 10/40/50 DEA/H2O/CH3CN (v/v/v). 0.20 mL extract to 0.45 mL silylated preparation. Analysis using scan mode on a 5975i MSD. Extracted ion chromatograms were used to determine area response for molecular ions.
(N=2, Range in Parentheses)
(16)
(20)
(23)
(10)
(13)
(16)
(13)
(11)
(a)
(2)
(4)
(3)
(2)
(1)
(3)
(1)
(0)
(a)
(1)
(5)
(13)
(6)
(3)
(0)
(7)
(5)
(b)
(86)
(96)
(105)
(93)
(4)
(3)
(5)
(4)
(b,c)
(44)
(8)
(150)
(76)
(7)
(12)
(10)
(11)
Salmon Flavor
(b)
(41)
(33)
(44)
(26)
(3)
(4)
(5)
(5)
(b) 0.5 g extracted into 10 mL of 10/40/50 DEA/H2O/CH3CN (v/v/v). 0.20 mL extract to 0.45 mL silylated preparation. Analysis by SIM on an Agilent 5973 MSD.
(c) Rubbery precipitate in some final extracts that can clog the syringe
Figure 1: Standard chromatogram showing the TMS derivatives of target compounds
Figure 2: Cyanuric Acid tri-TMS derivative mass spectrum
Figure 3: Ammelide tri-TMS derivative mass spectrum
Figure 4: Ammeline tri-TMS derivative mass spectrum
Figure 5: Melamine tri-TMS derivative mass spectrum
Figure 6: 2,6-Diamino-4-chloropyrimidine di-TMS derivative mass spectrum
Figure 7: Example of a QEdit report for melamine