Table 10.
Area source concentrations of diacetyl and 2,3-pentanedione using evacuated canisters (NMAM 3900).
| Diacetyl | 2,3-Pentanedione | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Source | N | GM (ppb) | GSD | P95 (ppb) | %BDL | GM (ppb) | GSD | P95 (ppb) | %BDL |
| NON-FLAVOR | |||||||||
| Café grinder | 7 | 118 | 6.5 | 2,487 | 0 | 122 | 6.4 | 2,501 | 0 |
| Ground coffee | 52 | 488 | 10 | 21,788 | 0 | 251 | 11 | 12,674 | 0 |
| Heat sealing bags | 3 | – | – | 16* | 0 | – | – | 8.8* | 0 |
| Miscellaneous quality control | 11 | 27 | 4.9 | 366 | 0 | 22 | 5.6 | 368 | 0 |
| Miscellaneous café | 7 | 12 | 2.2 | 44 | 0 | 13 | 2.5 | 58 | 0 |
| Packaging roasted coffee | 18 | 28 | 4.2 | 292 | 0 | 14 | 3.9 | 129 | 0 |
| Quality control grinding | 9 | 50 | 6.0 | 928 | 0 | 42 | 5.7 | 720 | 0 |
| Roasted coffee | 54 | 19 | 4.7 | 245 | 0 | 10 | 4.6 | 125 | 1.9 |
| Roasted coffee in bag | 5 | 76 | 27 | 16,456 | 0 | 68 | 19 | 8,491 | 0 |
| Roasted coffee in container | 53 | 225 | 8.2 | 7,168 | 0 | 140 | 7.9 | 4,213 | 0 |
| Roaster cooling drum | 12 | 6.0 | 3.3 | 41 | 0 | 3.2 | 4.0 | 31 | 8.3 |
| Roaster door | 10 | 8.3 | 2.2 | 30 | 0 | 4.9 | 2.7 | 25 | 0 |
| Roasting | 12 | 21 | 6.0 | 411 | 8.3 | 11 | 4.4 | 123 | 8.3 |
| Sample roaster | 5 | 75 | 7.6 | 2,059 | 0 | 38 | 12 | 2,143 | 0 |
| FLAVOR | |||||||||
| Flavored coffee | 8 | 6.6 | 100 | 11,868 | 38 | 6.3 | 71 | 6,190 | 38 |
| Flavoring | 9 | 24 | 381 | 354,158 | 44 | 1,882 | 17 | 185,446 | 0 |
| Ground coffee | 17 | 59 | 36 | 20,945 | 24 | 143 | 9.7 | 6,038 | 5.9 |
| Miscellaneous quality control | 1 | – | – | 37* | 0 | – | – | 20* | 0 |
| Packaging roasted coffee | 16 | 45 | 4.4 | 497 | 13 | 47 | 3.6 | 378 | 6.3 |
| Roasted coffee | 3 | – | – | 7,386* | 33 | – | – | 1,749* | 0 |
N, number of samples; GM, geometric mean; ppb, parts per billion; GSD, geometric standard deviation; P95, 95th percentile; %BDL, percent samples below the limit of detection; –, not enough samples above the detection limit to obtain an estimate; max*, maximum presented when <5 measurements were above the detection limit; “–,” not enough samples above the detection limit to obtain an estimate.