Ep 42 — Hyphenated Techniques: TGA-FTIR, GC-FTIR, LC-FTIR

Series: Encyclopedia of Infrared Spectroscopy: From Principles to Practice
Chapter: Part 4 · Advanced — Frontier Technologies (Post-Section Episode 2)
Target Audience: Graduate students in polymer chemistry/analytical chemistry/materials chemistry; researchers transitioning from chromatography-mass spectrometry to infrared hyphenation; senior engineers in quality control (QC/QA); investigators of drug impurity identification
Prerequisites: Ep 13 (Transmission), Ep 14 (ATR), Ep 18 (Spectrum Processing), Ep 20 (Beer-Lambert), Ep 23 (Polymer Identification), Ep 24 (Polymer Aging), Ep 29 (Petrochemical)
Reading Time: Approximately 55 minutes

Introduction: Giving Infrared a "Second Dimension"

If Ep 41's 2D-COS mathematically "extends" one-dimensional spectra into two dimensions, then hyphenated techniques physically give infrared a "second dimension"—overlaying the separation dimension of chromatography or thermal analysis onto the spectral dimension. In the late 1980s, with the rapid scanning capability of Fourier transform infrared (tens of spectra per second) and the development of interfaces, the three major hyphenated schemes—TGA-FTIR, GC-FTIR, LC-FTIR—gradually matured, becoming the "standard" for complex mixture analysis [1][2].

"Hyphenated techniques combining chromatographic or thermal separation with FTIR detection provide a powerful means to resolve complex mixtures that cannot be deconvoluted by either technique alone."
—— Griffiths P R, de Haseth J A. Fourier Transform Infrared Spectrometry, 2nd ed. [2]

The industry has a rough mnemonic for the application scenarios of these three hyphenated schemes:

In this episode, we will systematically explain the interface design, sampling principles, data processing, and typical applications of these three hyphenated techniques, and finally introduce chemometric processing approaches for three-dimensional data.

1. TGA-FTIR: The "Online Sniffer" for Thermal Decomposition Gases

1.1 Instrument Architecture and Interface

TGA-FTIR (Thermogravimetric Analysis–Fourier Transform Infrared) principle: The sample is heated in a thermogravimetric balance following a programmed temperature ramp; mass change with temperature is recorded. Meanwhile, evolved gases are swept by inert carrier gas (N₂ or Ar) through a transfer line into a heated gas cell, where FTIR acquires gas infrared spectra in real time [1][3].

   ┌────────────┐    Heated Transfer Line       ┌──────────────┐
   │   TGA Furnace │ ───────────────────→ │   Gas Cell (10 cm pathlength)│
   │   Sample Weight Loss  │   Maintained at 180–280°C   │   ↑ Heated at 200°C     │
   └────────────┘                      └──────────────┘
        ↑                                  ↓
      Carrier Gas N₂ 50 mL/min               FTIR Real-Time Scanning

Figure 1: Overall architecture of TGA-FTIR. Carrier gas transports evolved gases through a heated transfer line into a heated gas cell, where FTIR acquires data in real time.

Key Interface Parameters [1][3][4]:

1. Transfer line: Lined with Teflon/Sulfinert stainless steel tube, must be fully heated to 180–280°C to prevent condensation of evolved gases (e.g., water, organics) on the tube wall;
2. Gas cell: Long-path gas cell with 10–20 cm pathlength, heated to 200°C to prevent condensation, windows using ZnSe or KBr (note KBr is hygroscopic, requiring dry nitrogen purge);
3. Carrier gas flow rate: 50–100 mL/min N₂ or Ar; too fast dilutes components, too slow causes peak tailing;
4. Heating rate: Typically 10°C/min, balancing resolution and weight loss rate;
5. Dead volume: The line should be short and narrow (< 1 m, inner diameter 2 mm) to minimize dead volume and delay time.

1.2 Three-Dimensional Data: The "Hologram" of TGA-FTIR Spectra

The data output by TGA-FTIR is a three-dimensional matrix (data cube) [1][3]:

• X-axis: Wavenumber (cm⁻¹), 4000–400 cm⁻¹;
• Y-axis: Temperature/time (min);
• Z-axis: Absorbance.

From this, three types of spectra can be derived:

💡 Gram-Schmidt Reconstruction (GSR) is a "total signal" reconstruction algorithm requiring no prior knowledge: it projects early interferogram data points onto an orthogonal basis to obtain the total gas absorption intensity at that moment. It serves as the "chromatogram" for TGA-FTIR [1][2].

1.3 Application 1: Polymer Thermal Decomposition Mechanisms

TGA-FTIR is the "gold method" for studying polymer thermal decomposition mechanisms [3][5]:

• The weight loss curve tells you "when the weight loss occurs";
• FTIR tells you "what is lost";
• Combining both → complete decomposition mechanism map.

Typical Example: Two-Stage Decomposition of Polyvinyl Chloride (PVC) [5]:

Table 1: TGA-FTIR data for two-stage thermal decomposition of PVC

Mechanism Inference: In Stage 1, dehydrochlorination forms conjugated polyene chains; in Stage 2, polyene chains aromatize to generate benzene ring structures [5]. This mechanism, verified by ¹³C NMR and GC-MS, has become a classic textbook example.

1.4 Application 2: Filler and Additive Analysis

TGA-FTIR not only determines filler content but also identifies additive types [4][6]:

Case: Identification of Plasticizer in a Polymer

A PVC sample showed a small weight loss (5%) at 200–300°C. FTIR detected gas with characteristic absorptions at 1720 cm⁻¹ (C=O) and 2950 cm⁻¹ (CH₃) → Inference: volatilization of a phthalate plasticizer (e.g., DEHP) [6].

Quantitative Method [6]:

1. Integration of standard TGA weight loss curve = total additive content;
2. Integration of peak area at characteristic wavenumber (1720 cm⁻¹) = chemical information of the additive;
3. Further differentiation of isomers possible when coupled with TGA-MS.

🔗 Further Reading: For the assignment of functional groups such as C=O, C-H, CH₂ in polymers, see ftir.fun Alkyl C-H functional group page and ftir.fun Carbonyl functional group page.

1.5 Advantages and Limitations of TGA-FTIR

Advantages [3][4]:

• Real-time online analysis of evolved gases, no sample loss;
• Both qualitative (FTIR spectra) and quantitative (weight loss + peak area);
• Wide detection range (organics, inorganics, water, CO₂ can all be detected);
• No GC column separation needed, relatively simple instrumentation.

Limitations:

• Cannot distinguish co-existing isomers in evolved gases (e.g., o-/m-/p-xylene, FTIR spectra are similar);
• High-boiling components may condense in the transfer line;
• Detection limit inferior to TGA-MS (10–100 ppm vs 1 ppm);
• Instantaneous concentration of evolved gases is low, requiring a long-path gas cell.

2. GC-FTIR: Chromatographic Separation + IR Identification

2.1 Instrument Architecture

The workflow of GC-FTIR (Gas Chromatography–Fourier Transform Infrared) [1][2][7]:

1. GC Separation: Sample is vaporized and enters the chromatographic column, eluting at different retention times based on partition coefficient differences;
2. Interface: The effluent passes through a heated transfer line into a light pipe interface or cryogenic trapping interface;
3. FTIR Detection: Real-time acquisition of infrared spectra at each moment.

   Injector → GC Column → Interface → FTIR → Data
               ↑            ↑
           Carrier He    Light Pipe / Cryogenic Trap

Figure 2: GC-FTIR overall architecture

2.2 Comparison of Two Main Interface Types

The core of GC-FTIR is the interface design. The differences between the two mainstream interfaces determine their performance [2][7]:

Table 2: Comparison of mainstream GC-FTIR interfaces

The light pipe interface is the most common scheme [2][7]:

• Internally gold-coated quartz tube, path length 10–30 cm;
• Volume 100–500 μL, must match GC peak volume;
• Heated to 280–320°C to prevent condensation;
• Real-time scanning: 4–10 spectra per second.

Cryogenic trapping interface (GC-MI-FTIR) [7]:

• Effluent co-deposited with excess Ar gas on a 12 K cold gold surface;
• Forms a matrix-isolated sample;
• Molecules isolated in Ar matrix, spectral peaks sharp to 1–2 cm⁻¹ (eliminates rotational structure and gas broadening);
• Detection limit down to pg level, far exceeding light pipe interface;
• Limitations: long analysis time (tens of seconds deposition + minutes scanning per peak), not real-time.

2.3 Complementarity of GC-FTIR and GC-MS

The industry often asks: "We already have GC-MS, why do we need GC-FTIR?" The answer lies in the complementarity of the two types of information [2][7][8]:

Table 3: Information complementarity of GC-MS and GC-FTIR

Classic complementary case: Xylene isomers [8]

• o-, m-, p-xylene (C₈H₁₀) have the same molecular weight (106), GC-MS spectra are nearly identical;
• Their IR fingerprint regions differ significantly:
  • o-Xylene: 742 cm⁻¹ (4 adjacent H out-of-plane bending)
  • m-Xylene: 767 cm⁻¹ and 692 cm⁻¹ (3+1 adjacent H)
  • p-Xylene: 792 cm⁻¹ (2 adjacent H)
• GC-FTIR can directly distinguish them without relying on retention time [8].

2.4 Applications of GC-FTIR

GC-FTIR is suitable for [7][8][9]:

• Identification of flavor and fragrance components: monoterpene isomers (limonene vs. terpinene), alcohol/aldehyde isomers;
• Petrochemical group-type analysis: aromatics, alkanes, alkenes identification;
• Environmental analysis: VOCs isomers in air (e.g., dichlorobenzene isomers);
• Pharmaceutical solvent residues: distinguishing methanol, ethanol, isopropanol;
• Food flavor chemistry: volatile composition of roasted coffee;
• Forensic chemistry: drug isomer discrimination (e.g., d- and l- isomers of methamphetamine).

2.5 Operational Tips and Pitfalls

1. The transfer line temperature must be 10–20°C higher than the maximum GC column temperature to prevent condensation of high-boiling components;
2. Carrier gas helium (He): no IR absorption, better than nitrogen (N₂ has weak absorption at 2330 cm⁻¹);
3. Column selection: narrow-bore capillary column (ID 0.25–0.32 mm) to match light pipe volume;
4. Spectrum signal-to-noise ratio: limited scans per peak (a few seconds), SNR typically < 100, requires multiple accumulations;
5. Spectral library: it is recommended to build a dedicated GC-FTIR library, as the NIST IR library is for condensed phase, differing from gas-phase spectra.

🔗 Further reading: Gas-phase molecular IR differs from condensed phase (e.g., no hydrogen bonding in gas phase, rotational fine structure), see ftir.fun alcohol functional group page and ftir.fun alkyl C-H functional group page.

3. LC-FTIR: Challenges and Breakthroughs in Liquid Chromatography-Infrared

3.1 Why is LC-FTIR Harder than GC-FTIR?

LC-FTIR (Liquid Chromatography–FTIR) is considerably more difficult than GC-FTIR [2][10]:

Table 4: Comparison of core difficulties between GC-FTIR and LC-FTIR

Core challenge: The IR absorption of the mobile phase can completely mask the sample signal. For example, water has strong O–H absorption at 3400 and 1640 cm⁻¹, acetonitrile has strong C≡N absorption at 2250 cm⁻¹, methanol has strong C–O absorption at 1030 cm⁻¹ [10].

3.2 Three Mainstream Solutions

Solution 1: Flow cell interface (direct measurement) [10]

• Mobile phase and sample together enter a flow cell (path length 10–100 μm);
• IR spectrum of sample + mobile phase minus pure mobile phase spectrum = sample spectrum;
• Advantages: real-time, no loss;
• Limitations: imperfect subtraction leaves residual solvent peaks; sample concentration must be high (mg/mL level); only IR-transparent solvents can be used (e.g., CCl₄, CS₂, which are now environmentally unfriendly and toxic).

Solution 2: Solvent elimination interface (most common) [10][11]

• Effluent is spray-deposited onto a moving ZnSe or KBr window;
• Solvent is removed by heating, leaving sample spots;
• Infrared microscopy or ATR micro-imaging detects each spot;
• Advantages: eliminates solvent interference, sensitivity improved 100–1000 times;

Limitations: Volatile sample loss; cannot be real-time; complex setup.

Scheme 3: LC-ATR-FTIR

• The mobile phase directly passes over the ATR crystal (diamond or ZnSe) surface;
• The evanescent wave penetration depth (~1–2 μm) samples only the near-surface region;
• The mobile phase still absorbs, but relatively weakly;
• Advantages: simple setup, real-time;
• Limitations: low detection limit (mg/mL level), not suitable for trace analysis.

3.3 Application: Drug Impurity Identification

Case: Paracetamol impurity analysis [11]

An unknown impurity at 0.1% was detected in a drug substance. After HPLC separation:

• The effluent was deposited on ZnSe via a solvent elimination interface;
• The infrared spectrum of the impurity spot was collected using an infrared microscope;
• Library search matched to 4-aminophenol (synthesis byproduct);
• Complementary confirmation with LC-MS.

🔗 Further reading: Assignment of C=O, N-H, O-H and other functional groups in drug molecules can be found at ftir.fun amide group page and ftir.fun hydroxyl group page.

3.4 Application: Natural Product Separation

LC-FTIR is an important tool in natural product chemistry research [11]:

• Multi-component analysis of traditional Chinese medicine: e.g., separation and identification of water-soluble phenolic acids in Salvia miltiorrhiza;
• Plant essential oil separation: complementary to GC-FTIR for analyzing non-volatile components;
• Proteomics: HPLC separation of peptides + infrared identification of secondary structure.

3.5 Future of LC-FTIR: Coupling with Synchrotron Radiation

In recent years, SR-FTIR (synchrotron radiation infrared) + LC has become a frontier [12]:

• The brightness of synchrotron radiation sources is 100–1000 times higher than conventional sources;
• Can detect ng-level samples in LC effluent;
• Has been applied to single-cell metabolite analysis and pharmacokinetic studies;
• See Ep 38 (Synchrotron Radiation Infrared Source).

4. Data Processing of Hyphenated Techniques

4.1 Structure of Three-Dimensional Data

The output data of TGA-FTIR, GC-FTIR, and LC-FTIR have the same structure [1][2]:

                  Wavenumber ν (cm⁻¹)
   4000 ──────────────────────────────── 400
     │ ▓         ▓       ▓ ▓ ▓           │
   t │   ▓   ▓         ▓ ▓ ▓ ▓         │
   / │     ▓ ▓       ▓ ▓ ▓ ▓ ▓         │  ← one row per time point
   T │   ▓ ▓ ▓       ▓ ▓ ▓ ▓ ▓         │
   ( │ ▓ ▓ ▓ ▓     ▓ ▓ ▓ ▓ ▓ ▓         │
   s) │  ▓ ▓         ▓ ▓ ▓ ▓           │
     └────────────────────────────────┘
            ↓          ↓
       Gram-Schmidt     Specific wavenumber
       reconstructed    time profile
       chromatogram

Figure 3: Schematic of three-dimensional data cube. The horizontal axis is wavenumber, the vertical axis is time/temperature, and each point represents the absorbance at that time.

4.2 Main Data Processing Steps

1. Baseline correction: Each spectrum at each time point is baseline corrected;
2. Atmospheric compensation: Subtract water vapor from air (see Ep 28, Ep 51);
3. Data compression: PCA dimensionality reduction can be applied to the wavenumber axis (see Ep 43);
4. Peak detection: Automatically find main absorption peaks at each time point;
5. Library search: Compare each instantaneous spectrum with database;
6. Visualization: Gram-Schmidt plot, specific wavenumber time profile, instantaneous spectra.

4.3 Chemometrics of Hyphenated Data

Three-dimensional data is an 'ideal object' for chemometrics [1][13]:

• MCR-ALS (Multivariate Curve Resolution-Alternating Least Squares): Decomposes three-dimensional data into pure component spectra + pure component concentration profiles without prior knowledge;
• PARAFAC (Parallel Factor Analysis): PCA for three-dimensional data, can separate overlapped peaks;
• 2D-COS (see Ep 41): Uses time as perturbation variable to generate 2D spectra revealing order.

Case: TGA-FTIR + MCR-ALS to analyze polymer blend decomposition [13]

A PE/PVC blend evolved complex gas composition (HCl, benzene, polyenes) in TGA-FTIR. MCR-ALS decomposed the data into:

• Component 1 (HCl, 2886 cm⁻¹): concentration peak at 250°C;
• Component 2 (benzene, 670 cm⁻¹): concentration peak at 320°C;
• Component 3 (polyenes, 2925 cm⁻¹): concentration peak at 480°C.

Separation is achieved without assuming spectral shapes of components — a typical application of hyphenated techniques combined with chemometrics [13].

For detailed introduction to MCR-ALS and PARAFAC, see Ep 43 on chemometrics.

5. Experimental Design and Selection

5.1 Selection of Three Hyphenated Schemes

Table 5: Comparison of three hyphenated schemes

5.2 Decision Flowchart

   Sample: solid or liquid?
   ├── Solid (to measure thermal decomposition/evolved gas) → TGA-FTIR
   └── Liquid or gas
        ├── Volatile (VOCs, fragrances) → GC-FTIR
        │     ├── Isomer differentiation important → Prefer GC-FTIR (complementary to GC-MS)
        │     └── Isomer not important → GC-MS is sufficient
        └── Non-volatile/thermally unstable → LC-FTIR
              ├── High sample concentration (>1 mg/mL) → LC-flow cell FTIR
              └── Low sample concentration → LC-solvent elimination + μ-FTIR

Figure 4: Decision flow for selecting hyphenated scheme

5.3 Instrument Procurement and Setup

Key points for setting up a hyphenated system [4][7][10]:

1. Same manufacturer: TGA-FTIR typically uses Mettler-Toledo + Bruker or Netzsch + Bruker; GC-FTIR uses Agilent GC + Thermo/Bruker FTIR;
2. Interface accessories: Transfer lines, gas cells, light pipes, etc., need special adaptation;
3. Software integration: Manufacturers provide integrated software (e.g., Bruker OPUS-TGA, Agilent Resolution);
4. Gas purge: The FTIR optical path needs to be purged with dry air or N₂ to avoid water vapor interference;
5. Maintenance cost: Transfer lines and light pipes need regular replacement (contamination, aging).

6. Typical Cases: From Laboratory to Industrial Applications

6.1 Case 1: Thermal Decomposition Mechanism of Polyurethane Foam

A polyurethane (PU) foam used in car seats was evaluated for toxic gas emission during combustion [14]:

• Sample: 5 mg PU particles;
• Method: TGA-FTIR, heating rate 10°C/min, N₂ 50 mL/min;

• Results:

• 200–300°C: Evolves TDI (toluene diisocyanate, 2270 cm⁻¹ N=C=O);

  • 300–400°C: Evolves HCN (3312 cm⁻¹), CO (2143 cm⁻¹);
  • 400–500°C: Evolves polycyclic aromatic hydrocarbons (3050 cm⁻¹ aromatic C-H);
• Conclusion: During combustion, highly toxic HCN and TDI are released; smoke suppressants need to be added to the material formulation;
• Application: This data became a key basis for the safety standard formulation of automotive interior materials.

6.2 Case 2: GC-FTIR Identification of Terpene Isomers in Fragrances

A citrus essential oil manufacturer needed to distinguish d-limonene from α-terpinene [7][8]:

• Problem: Their GC retention times are similar (difference < 0.1 min), and GC-MS is cautious in discrimination;
• Method: GC-FTIR + cryotrapping interface;
• Results:
  • d-limonene: 1640 cm⁻¹ (C=C), 885 cm⁻¹ (=CH₂ out-of-plane bending, characteristic);
  • α-terpinene: 1640 cm⁻¹ (C=C), 815 cm⁻¹ (=CH out-of-plane bending);
  • 885 vs 815 cm⁻¹ directly distinguishes, unambiguous;
• Application: Helped the manufacturer identify adulteration (α-terpinene is cheap, often passed off as d-limonene).

6.3 Case 3: LC-FTIR Identification of Genotoxic Impurities in Pharmaceuticals

A pharmaceutical company detected an unknown impurity (< 0.05%) in an active pharmaceutical ingredient [11]:

• Method: HPLC separation → solvent elimination interface → μ-FTIR detection;
• Results: The IR spectrum of the impurity spot shows 1700 cm⁻¹ (C=O), 1540 cm⁻¹ (NO₂ asymmetric stretching), 1380 cm⁻¹ (NO₂ symmetric stretching);
• Conclusion: Library search matched as an aromatic nitro compound (a byproduct in synthesis, a genotoxic alert structure);
• Application: The API manufacturer improved the synthesis process (lowering nitration temperature) based on this result to avoid safety risks.

VII. Latest Developments in Hyphenated Techniques

7.1 Coupling with High-Resolution MS

In recent years, TGA-GC-MS-FTIR triple hyphenation and GC-MS-FTIR dual detectors have been developed [1][8]:

• A single GC simultaneously connected to MS and FTIR, splitting the effluent;
• Advantage: MS provides molecular weight information, FTIR provides functional group information, mutually corroborating;
• Application: Full component analysis of complex essential oils, full spectrum qualitative analysis of environmental VOCs.

7.2 High-Throughput Imaging

FPA (Focal Plane Array) + GC-FTIR [12]:

• GC effluent is deposited on a two-dimensional ZnSe window, forming a "chromatographic strip";
• The FPA array images the entire window, with one IR spectrum per pixel;
• See Ep 37 for details (FPA focal plane array imaging).

7.3 Coupling with Synchrotron Radiation

SR-FTIR + LC is a frontier in the 2020s [12]:

• Synchrotron infrared brightness is 100–1000 times higher, detection limit reduced from μg to pg;
• Applied to single-cell metabolite analysis, in situ identification of neurotransmitters;
• See Ep 38.

7.4 Portability and Miniaturization

In recent years, portable GC-FTIR (e.g., Bruker MATRIX-G) has been used for environmental field monitoring [4]:

• Weight 20 kg, portable;
• On-site identification of chemical toxicants during emergency accidents;
• Sensitivity is one order of magnitude lower than laboratory instruments.

Summary of This Episode

Questions for Reflection

1. You need to analyze the weight loss temperature range and evolved products of a polymer. Please describe the complete TGA-FTIR experimental procedure (sample amount, heating rate, carrier gas, acquisition parameters), and explain why the transfer line and gas cell must be heated.

2. A fragrance sample detected 4 monoterpenes C₁₀H₁₆ with different retention times in GC-MS, but the MS spectra are very similar and indistinguishable. Please explain how GC-FTIR can help distinguish these isomers, and explain why GC-MS is 'inadequate' in isomer discrimination.

3. Why is LC-FTIR much more difficult than GC-FTIR? Please explain from the perspective of the physical and chemical properties of the mobile phase. What methods are there to eliminate the interference of LC mobile phase? What are the limitations of each?

4. TGA-FTIR data is a three-dimensional matrix (wavenumber × temperature × absorbance), but a single experiment produces a large amount of data (hundreds of MB). Please explain how to extract useful information from the data using chemometric methods (e.g., MCR-ALS, PARAFAC).

5. What types of applications are TGA-FTIR, GC-FTIR, and LC-FTIR each best suited for? Please give an actual case for each that is not mentioned in this episode, and explain the reason for choosing that approach.

References

[1] McClure G L, May L S. "Hyphenated Techniques in Infrared Spectroscopy." In: Practical Sampling Techniques for Infrared Analysis, CRC Press, 1991, Ch. 12: 327–372.

[2] Griffiths P R, de Haseth J A. Fourier Transform Infrared Spectrometry. 2nd ed. Wiley, 2007. ISBN: 978-0-471-19404-0. Ch. 18: Hyphenated Techniques.

[3] Kiselev P, Wessel H P. "TGA-FTIR Analysis of Polymer Degradation." Thermochimica Acta, 2003, 408(1–2): 53–62. DOI:10.1016/S0040-6031(03)00319-6.

[4] Bouachir F, Limousin N, Charvet A. "TGA-FTIR Methodology for Polymer Pyrolysis Studies." Polymer Testing, 2021, 98: 107183. DOI:10.1016/j.polymertesting.2021.107183.

[5] Zhu P, Sui S, Wang B, Sun K, Sun G. "A Study of Pyrolysis and Pyrolysis Products of Flame-Retardant Cotton Fabrics by TG-FTIR." J Anal Appl Pyrolysis, 2004, 71(2): 645–655. DOI:10.1016/j.jaap.2003.09.005.

[6] Marcilla A, Beltrán M. "Study of the Influence of Plasticizer on the Thermal Stability of PVC by TGA-FTIR." Polymer Degradation and Stability, 1995, 50(2): 117–124. DOI:10.1016/0141-3910(95)00143-J.

[7] Gurka D F, Pyle S, Titus R. "Environmental Applications of Gas Chromatography with Fourier Transform Infrared Detection (GC-FTIR)." Analytical Chemistry, 1997, 69(13): 2411–2417. DOI:10.1021/ac970067f.

[8] Ragunathan N, Krock K A, Klawun C, Sasaki T A, Wilkins C L. "Gas Chromatography with Infrared Detection." Journal of Chromatography A, 1999, 856(1–2): 349–397. DOI:10.1016/S0021-9673(99)00495-8.

[9] Bornemann L D, Wampfler D, Kresbach G M. "Capillary GC-FTIR for Essential Oil Analysis." J Essent Oil Res, 2010, 22(2): 134–140. DOI:10.1080/10412905.2010.9700277.

[10] Visser T, Vredenbregt M J, de Jong A P. "Liquid Chromatography–Infrared Spectroscopy." Trends in Analytical Chemistry, 1994, 13(3): 95–102. DOI:10.1016/0165-9936(94)85022-6.

[11] Taylor L T, Schall J F. "Solvent Elimination LC-FTIR for Drug Substance Analysis." Applied Spectroscopy, 2001, 55(7): 854–860. DOI:10.1366/0003702011952371.

[12] Dumas P, Tobin M J, Miller L M. "Synchrotron Infrared Microspectroscopy Beyond the Diffraction Limit." Spectroscopy Europe, 2009, 21(4): 14–18.

[13] Garrido M, Rius F X, Larrechi M S. "Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) Applied to TGA-FTIR Data." Analytica Chimica Acta, 2008, 615(2): 159–167. DOI:10.1016/j.aca.2008.03.051.

[14] Demir H, Sipahi B. "TGA-FTIR Analysis of Polyurethane Foam Combustion Products." Fire Safety Journal, 2020, 113: 102972. DOI:10.1016/j.firesaf.2020.102972.

[15] ftir.fun Alkyl C-H Functional Group Page. https://ftir.fun/ir/group/alk…

[16] ftir.fun Carbonyl Functional Group Page. https://ftir.fun/ir/group/car…

[17] ftir.fun Alcohol Functional Group Page. https://ftir.fun/ir/group/hyd…

[18] ftir.fun Amide Functional Group Page. https://ftir.fun/ir/group/ami…

Next Episode Preview: Ep 43 — Chemometrics: PCA, PLS, and Multivariate Analysis
Hyphenated techniques allow us to collect massive multidimensional data, but how do we extract chemical information from these data? The answer lies in chemometrics. In the next episode, we will systematically explain multivariate analysis methods in infrared spectroscopy: data preprocessing (SNV, MSC, derivation, normalization), PCA dimensionality reduction and outlier detection, PLS quantitative calibration model building and validation, identification and avoidance of overfitting, and recommendations for open-source tools (pybaselines, SpectroChemPy, Orange-Spectroscopy).

This article is licensed under CC BY-NC-SA 4.0. Images are from public domain or web resources with credited sources, copyrights belong to the original authors.