Table 1.
Recent studies of HS-GC-IMS with NTS.
| Reference (Main Author/Yeat) | Aim of Reference | Matrice (Number of Samples) |
IMS Type (Ionization Source) | Separation (Length × Inner Diameter (ID), Film Thickness (ft)) | Unsupervised and Supervised Methods | Complementary Analysis | Method and Number of Compounds Identified | Data-Split; (Cross-) Validation |
|---|---|---|---|---|---|---|---|---|
| Arroyo-Manzanares/2017 [89] |
classification between pig’s food sources | dry-cured Iberian ham (24) | FlavourSpec by GAS (tritium, 6.5 keV) | FS-SE-54-CB stationary phase (30 m × 0.32 mm ID, 0.25 µm ft; 94% methyl, 5% phenyl, 1% vinyl silicone) | PCA PCA-LDA (90%), kNN (k = 3, 90%) |
library search (LS) 5 |
80/20 | |
| Arroyo-Manzanares/2019 [84] |
detection of adulteration (sugar cane or corn syrup) | honey (198) | FlavourSpec by GAS (tritium, 6.5 keV) | HP-5MS UI (nonpolar) (30 m × 0.25 mm ID, 0.25 μm ft) | OPLS-DA (97.4% prediction of class, 93.75% prediction of adulteration degree) | 0 | 80/20 | |
| Brendel/2021 [90] |
detection of botanical origin | citrus juice (47) | OEM IMS by GAS (tritium, 300 MBq) | ZB-5 ms column (30 m × 0.25 mm, 0.25 μm ft; 5% phenyl-methylpolysiloxane) | PCA PCA-LDA (91.5%) |
HS-GC-MS/IMS | reference substances (RS) 9 |
leave-one-out (LOO) 4- and 6-fold CV |
| Cavanna/2018 [85] |
detection of food freshness | egg (132) | FlavourSpec | FS-SE-54-CB-1 stationary phase (15 m × 0.53 mm ID, 1µm ft; 94% methyl, 5% phenyl, 1% vinyl silicone) | PCA OPLS-DA (97%) |
RS, SPME-GC-MS 5 |
leave-one-out CV | |
| Chen/2020 [30] |
detection of geographical origin | Chinese yellow wine (122) | FlavourSpec by GAS (tritium, 6.5 keV) | nonpolar column (30 m, 95% methyl, 5% phenyl) | PCA QDA (95.35%) |
LS 12 |
70/30 | |
| del Contreras/2019 [33] |
quality assessment/classification | olive oil (701) | FlavourSpec by GAS (tritium, 6.5 keV) | SE-54-CB (30 m × 0.32 mm, 0.25 µm ft; 94% methyl, 5% phenyl, 1% vinyl silicone) | PCA non-targeted (PCA-LDA, kNN k = 3: 79.4%), targeted OPLS-DA: 74.3%) |
LS, RS 20 |
80/20; 7-fold CV |
|
| Garrido-Delgado/2011 [80] |
quality assessment/classification | olive oil (49) | portable UV–IMS instrument (UV ionization source: 10.6 eV) and FlavourSpec by GAS (tritium, 6.5 keV) | MCC (nonpolar) OV-5 (20 cm) | PCA UV-IMS: kNN (k = 3, 86.1%); GC-IMS: kNN (k = 3, 100%) |
0 | 71/29 and 64/26; leave-one-out CV |
|
| Garrido-Delgado/2012 [78] |
quality assessment/classification | olive oil (98) | FlavourSpec by GAS (tritium, 6.5 keV) | MCC (nonpolar) OV-5 (20 cm, 1000 parallel glass capillaries) | PCA Targeted: kNN (k = 3, 79%), non-targeted: kNN (k = 3, 87%) |
Organoleptic analysis | LS, RS 10 |
bootstrap validation (B = 100) |
| Garrido-Delgado/2015 [32] |
quality assessment/classification | olive oil (55) | FlavourSpec by GAS (tritium, 6.5 keV) | two different types of columns: MCC (20 cm × 3 mm ID, 900 parallel glass capillaries: 40µm ID, 0.2µm ft) and CC (30 m × 0.25 mm, 0.5µm ft) | PCA MCC-IMS: kNN (k = 3, 79%); CC-IMS: kNN (k = 3, 83%) |
LS, RS 26 |
bootstrap validation | |
| Gerhardt/2017 [34] |
detection of geographical origin | olive oil (40) | FlavourSpec by GAS (tritium, 6.5 keV) | NB-225 (25 m × 0.32 mm × 0.25 μm ft; 25% phenyl, 25% cyanopropyl methyl siloxane) | PCA PCA-LDA (98%) and kNN (k = 5, 92%) |
RS 4 |
10-fold CV | |
| Gerhardt/2018 [52] |
detection of botanical origin | honey (74) | advanced IMS by GAS (tritium, 300 MBq) | DB-225 (25 m × 0.32 mm, 0.25 μm ft; 25% phenyl, 25% cyanopropyl methyl siloxane) | PCA PCA−LDA (98.6%), kNN (k = 5, 86.1%), PLS-DA (PLS = 5, 97.0%) |
1H NMR | RS 18 |
87/13; 10-fold CV |
| Gerhardt/2019 [79] |
quality assessment/classification | olive oil (94) | FlavourSpec by GAS (tritium, 6.5 keV) | DB-225 (25 m × 0.32 mm × 0.25 μm ft; 25% phenyl, 25% cyanopropyl methyl siloxane) | PCA, HCA PCA-LDA (10 PCs, 83.3%), kNN (73.8%), SVM (88.1%) |
Organoleptic analysis | RS 25 |
10-fold CV |
| Gu/2020 [88] |
detection of quality changes during storage | wheat kernels infected with mold (90) | FlavourSpec by GAS (tritium, 6.5 keV) | MXT-WAX (nonpolar) (30 m × 0.53 mm ID, 0.1 μm ft) | PCA, HCA GA-SVM, Artificial samples: (Classification: 100%), (Quantification > 93.9%); Natural samples: (Classification: 60–86.7%, Quantification: 72.6–88.9%) |
LS 13 |
67/33 | |
| Gu/2020 [86] |
detection of quality changes during storage | milled rice infected with Aspergillus species (108) | FlavourSpec by GAS (tritium, 6.5 keV) | SE-54-CB | PCA kNN (k = 3, 94.44%), PLSR (90.9%) |
electric nose (E-nose) | LS, partially RS 25 |
67/33; 10-fold CV |
| Gu/2021[87] | detection of quality changes during storage | peanut kernel infected with aflatoxigenic fungi (180) | FlavourSpec | Rtx-WAX (30 m × 0.32 mm ID, 0.25 μm ft, RT-12,424) | PCA OPLS-DA (93.3%), low-level data fusion (90%), mid-level data fusion (96.7%) |
fluorescence analysis | LS 3 |
67/33; 7-fold CV, 200× permutation test |
| Li/2019 [5] |
classification between aroma | green tea (23) | real-time IMS: self-made positive photoionization (PP-IMS) with KR lamp | - | PCA PLS-DA (95.6%) |
sensory evaluation | 0 | 10-fold CV, 1000× permutation test |
| Schwolow/2019 [81] |
detection of botanical and geographical origin | honey and olive oil (64 honey, 54 EVOOs) | advanced IMS by GAS (tritium, 300 MBq) | DB-225 (25 m × 0.32 mm, 0.25 μm ft) | PCA Honey: PCA-LDA (33%); Olive oil: PCA-LDA (78%), PCA-LDA + data fusion (100%) |
FT-MIR | previous studies (18 honey, 7 olive oil) | 87/13; 10-fold CV |
| Shuai/2014 [44] |
detection of adulteration (with vegetable oils) | flaxseed oil (78) | IMS-KS-100 by Wuhan Syscan Technology (pulse glow discharge) | n-hexane extraction | recursive support vector machine (R-SVM) (93.1%) | 0 | 10-fold CV | |
| Vega-Marquez/2019 [91] |
quality assessment/classification | olive oil (701) | FlavourSpec by GAS (tritium, 6.5 keV) | SE-54-CB (30 m × 0.32 mm, 0.25 µm ft; 94% methyl, 5% phenyl, 1% vinyl silicone) | Deep learning (88.8%), SVM (83.1%), kNN (84.5%), Tree (78.3%), Regressor (85.5%), XGBoost (85.7%) | LS, RS 20 |
80/20 | |
| Wang/2019 [82] |
detection of botanical origin | honey (40) | FlavourSpec by GAS (tritium, 6.5 keV) | FS–SE–54–CB-0.5 (15 m × 0.53 mm ID) | PCA OPLS-DA (95%) (VIP > 1.5) |
HS-SPME-GC-MS8 | 200× permutation test | |
| Wang/2019 [83] |
detection of botanical origin | honey (120) | FlavourSpec by GAS (tritium, 6.5 keV) | FS–SE–54–CB-0.5 (15 m × 0.53 mm ID) | PCA PLS-DA (84%) |
LS 25 |
CV-ANOVA | |
| Yuan/2020 [92] |
metabolomic studies | rat feces (30) | FlavourSpec | FS-SE-54-CB-1 (15 m × 0.53 mm ID) | PCA OPLS-DA (86.6) |
LS 11 |
200× permutation test | |
| Zhu/2020 [31] |
quality assessment/classification | wine (143) | FlavourSpec | MXT-WAX (polar) (30 m × 0.53 mm ID, 0.5 μm ft, polyethylene glycol) | PCA PCA-LDA (65.7%), PLS-DA (58.7%), kNN (k = 5, 60.8%), SVM (51.8%), XGBoot (81.8%), ANN (89.5%) |
sensory evaluation | RS >30 |
85/15; 10-fold CV |