INDUSTRIAL & MANUFACTURING CHEMISTRY
ISO / EIC 17025 and CDFA Certifications
determination of product & reactant mass balances at each manufacturing stage
A. Refrigerant Gas Analysis from Cargo Containers
From late 2011 through 2012, several deaths have occurred world-wide as a result of the explosion of refrigerant gas in temperature controlled cargo containers. The general consensus is that a counterfeit refrigerant gas was used in place of the genuine R134a (Norflurane or tetrafluoroethane). MAI was the first lab worldwide to have analyzed this dangerous counterfeit gas and remains today as one of only a very few worldwide to have done so. We developed safe extraction methods for sampling in an oxygen free atmosphere that are now being used throughout the industry. Since then, a further consensus has developed that more than one type of counterfeit gas exists, athough the other types found to date are far less dangerous or even harmless in their use as compressor gases.
a) MAI has the following expertise:
b) Contamination Types
Up to present, R134a refrigerant gas contents fall into the following categories.
1) Substantially uncontaminated R134a
2) R134a that is contaminated with historical refrigerant gases, R12, R142b, R133a, and may be thought as contaminated with waste refrigerant gases.
3) R134a that is contaminated with non-refrigerant chlorinated compounds, such as Dichlormethane (DCM or Methylene chloride), Chloromethane (CM) or Dichloropropane, which may be thought of as being contaminated with waste chlorinated solvents/gases.
4) R134a that is contaminated with reactive gases such as
Tetramethylsilane (TMS). Trimethylaluminum
has been hypothesized but its presence has yet to be confirmed in published
data. Reaction products Neopentane & methylbutane are present with TMS and it can be inferred
from reaction stoichiometry that H2,g and CH4,g are
also present. This is the only type of contamination that is known with
certainty to be dangerous, but suspicions prevail about type 3 as well.
Actual real world sample reports follow:
Type 1: R40 Detected
Type 2: R40 NOT Detected
c) Note that these reports have a special quantitation style, as discussed next.
Conventional calibration curves are avoided because:
· Absolute gas phase amount of R134a is not useful because it is dependent upon sample P-T
· Target list changes as new contaminants are discovered
· Custom made gas phase standards are expensive & have long manufacture times.
· Some target compounds are likely unstable and thus have poor utility as analytical standards
· High (±0.5%) accuarcy for R134a is very difficult to achieve using a calibration curve
MAI’s chosen quantitation basis is % Area, derived from target ion area times multiplier to yield TIC area, merits discussion:
· % Area will always sum to 100% of user selected peaks, so that peaks unwanted for summation such as air (N2 & O2 resulting from valve rotation) or lab artifact peaks can be excluded. It provides needed accuracy & target list flexibility.
· Target ion times multiplier reduces to a minimum errors arising from manual peak integration of poorly resolved peaks, and of very small peaks that sit on the tail of a larger peak. The following table demonstrates this point.
Comparison of Methods for determining Compound Areas from GC-MS Data: TIC Direct Integration vs. Target Ion * Fixed Multiplier |
||||||||||||
Sample ID |
R134a |
R12 |
R142b |
|||||||||
WO 1201123 |
Area from target* multiplier |
TIC Area, RTEINT.P |
% calculated / TIC Integration |
Reported % R134a of Total Area |
Area from target* multiplier |
TIC Area, RTEINT.P |
Reported % R12 of Total Area |
% calculated / TIC Integration |
Area from target* multiplier |
TIC Area, RTEINT.P |
Reported % R142b of Total Area |
% calculated / TIC Integration |
-001A |
1.48E+07 |
1.33E+07 |
112 |
99.18 |
1.23E+05 |
1.52E+05 |
0.82 |
81 |
|
|||
-002A |
1.93E+07 |
1.67E+07 |
115 |
99.76 |
4.61E+04 |
2.60E+05 |
0.24 |
18 |
|
|||
-003A |
1.92E+07 |
1.67E+07 |
115 |
99.98 |
4.16E+03 |
1.28E+04 |
0.02 |
32 |
|
|||
-004A |
1.49E+07 |
1.32E+07 |
113 |
99.20 |
1.21E+05 |
1.44E+05 |
0.80 |
84 |
|
|||
-005A |
2.28E+07 |
1.95E+07 |
117 |
97.87 |
4.69E+05 |
5.20E+05 |
2.01 |
90 |
2.63E+04 |
4.28E+04 |
0.11 |
61 |
-006A |
1.89E+07 |
1.66E+07 |
114 |
99.39 |
1.05E+05 |
1.41E+05 |
0.55 |
74 |
1.20E+04 |
2.27E+04 |
0.06 |
53 |
-007A |
1.70E+07 |
1.51E+07 |
113 |
99.90 |
1.44E+04 |
4.60E+04 |
0.08 |
31 |
2.55E+03 |
1.01E+04 |
0.01 |
25 |
-008A |
1.67E+07 |
1.48E+07 |
113 |
99.99 |
9.86E+02 |
0.00E+00 |
<0.01 |
TIC not found |
|
|||
-009A |
1.28E+07 |
1.16E+07 |
111 |
100.00 |
|
|||||||
-010A |
1.32E+07 |
1.19E+07 |
111 |
99.95 |
7.00E+03 |
2.34E+04 |
0.05 |
30 |
|
|||
-011A |
9.28E+06 |
8.43E+06 |
110 |
99.95 |
4.22E+03 |
2.18E+04 |
0.05 |
19 |
|
|||
-012A |
1.61E+07 |
1.42E+07 |
113 |
99.99 |
8.38E+02 |
0.00E+00 |
<0.01 |
TIC not found |
|
|||
-013A |
1.71E+07 |
1.50E+07 |
114 |
100.00 |
|
|
||||||
-014A |
1.78E+07 |
1.55E+07 |
115 |
100.00 |
|
|
||||||
-015A |
1.63E+07 |
1.43E+07 |
114 |
100.00 |
|
|
||||||
-016A |
1.81E+07 |
1.58E+07 |
115 |
100.00 |
|
|||||||
-017A |
1.53E+07 |
1.35E+07 |
113 |
100.00 |
|
|||||||
Average= |
113 |
|
· The above table demonstrates that target ion area times fixed multiplier quite accurately reflects the R134a TIC area, averaging 113%. Also demonstrated is that this calculation is much more accurate than TIC areas for minor peaks (~<0.8%) are highly inaccurate relative to target ion based quantitation.
· BFB target tuning (for example Agilent ChemStation) is recommended, as is using the same NIST 2008 “M reference” fixed multiplier that is used in the above example reports.
· The largest peak on the chromatogram must be within the linear working range of the MS.
d) Reactive contaminant gases are very easy to distinguish from R134a & easy to identify using GC-MS
Reactive gas sample peaks are identified in the following chromatogram.