Rapid Analysis Methods for Oils and Biofuels Using IR/NIR Spectroscopy

Transcription

1 Rapid Analysis Methods for Oils and Biofuels Using IR/NIR Spectroscopy Ching-Hui Tseng, Nan Wang Eurofins QTA, Inc. July

2 Biofuels Fuel produced from renewable resources, especially plant biomass, vegetable oils, and treated municipal and industrial wastes. Biodiesel is made by processing vegetable oils and other fats. It can be used either in pure form or as an additive to petroleum-based diesel fuel. Bioethanol is produced by fermenting the sugars in biomass materials such as corn and agricultural residues etc. It is used in internal-combustion engines either in pure form or more often as a gasoline additive.

3 Biodiesel Reaction (Trans-esterification) 3

4 Biodiesel Production & Testing Segregation Consistency Price Test Oil Grease ROH Catalyst Optimization Efficiency Consistency Quality Test Test Glycerine Biodiesel Test

5 Tests for Biodiesel Production To produce quality biodiesel: The in-coming fat/oil needs to be tested to see if pretreatment is needed and the pretreatment outcome is satisfactory. The trans-esterification reaction can be monitored by testing the inprocess samples. The finished biodiesel need to meet ASTM D6751 (US) or EN14214 (Europe) standards to be used as biodiesel or blended with diesel fuel.

6 Biodiesel Traditional Testing Methods -- Specification for Biodiesel (B100) ASTM D a Property ASTM Method Limits Units Methanol EN max %mass Water by KF D %mass Kinematic Viscosity, 40 C D mm 2 /sec Sulfur, S15 Grade D max ppm Cloud Point D2500 Report C Acid Number D max Mg KOH/g Free Glycerin D max %mass Total Glycerin D max %mass Oxidation Stability EN min hours Cetane D min Cold Soak Filtration D max seconds

7 Biodiesel Traditional Testing Methods- Specification for Biodiesel (B100) EN14214 Property EN Method Limits Units Methanol, EN max %mass Water by KF EN ISO max mg/kg Viscosity, 40 C EN ISO 3104, ISO mm 2 /sec Sulfur EN ISO 20846, EN ISO max mg/kg Pour Point ISO 3016 Report C Acid Value EN max mg KOH/g Free Glycerin EN14105, EN max %mass Total Glycerin EN max %mass Oxidation Stability, 110 C EN15751/EN min hours Cetane EN ISO min Ester Content, EN min %mass Cold Filter Plugging Point EN 116 C Iodine Value EN max g iodine/100g Density, 15 C EN ISO 3675, EN ISO kg/m 3

8 Traditional Analysis Methods of Biodiesel Production In-coming oil: Free fatty acids: gas chromatography (GC) Moisture: Karl Fischer In-Process: Mono-, di- & tri- glycerides, total & free glycerin: GC method Finished B100 biodiesel: Total & free glycerin, methanol: GC method Acid number, cloud point, moisture, oxidation stability, CFPP, density, ester content, iodine value etc.: different equipment for each test By-product: Glycerin Ash: oven method Glycerin, methanol: GC Moisture: Karl Fischer

9 Traditional Analysis Method Summary In summary, the traditional methods Need a laboratory with all the required instruments Require chemists with different technical skills Expensive and resource intensive Time consuming

10 A Desired Technology One instrument can analyze all the materials, including feedstock, in-process samples, finished biodiesel and by-product Easy to operate so no specialized personnel is required Quick and green (no chemical reagents or waste disposal) Analyzed results are reliable Instrument and methods are easy to maintain 10

11 Infrared (IR) Spectroscopy It has been a very powerful analytical tool for sample identification and component analysis since the 1960s.

12 Principle of Infrared (IR) Spectroscopy hν Low Energy High Energy

13 Infrared Vibrational Energy States Harmonic Oscillation Inharmonic Oscillation Vibrational energy levels 3rd Overtone 2nd Overtone 1 st Overtone Fundamental The mid-infrared (IR) covers mostly fundamental vibrations while the near infrared (NIR) covers the overtone and combination bands

14 What is IR and NIR? Infrared (IR) spectroscopy : fundamental absorption of molecular functional groups, µm ( cm -1 ) Near infrared (NIR) spectroscopy : overtone or combination absorption of molecular functional groups, µm ( ,000 cm -1 ) Oleic Acid MIR Absorbance NIR NIR IR Wavenumber cm -1

15 Electromagnetic Spectrum: Infrard is between visible and microwave regions

16 MIR/NIR comparison MIR Absorption coefficient large (higher sensitive) NIR small (lower sensitive) Pathlength micrometers mm to tens of cm Sample container Salt, ATR quartz, glass Spectrum Sample complicated but unique (can be interpreted) Gas, liquid, paste, solid (small amount, homogeneous, e.g. biodiesel production samples ) simple but overlapped (hard to interpret) Liquid, paste, solid (large amount, inhomogeneous, e.g. bioethanol production samples) Analyzed material Good for organic Possible for inorganic Good for organic Poor for inorganic Available fiber optics 3 m (max.) 100 m (max.)

17 How does IR/NIR work for the qualitative analysis? Qualitative analysis - confirmation of incoming raw materials, identification of unknown samples Method - spectral matching, PCA, ANN etc. 17 8/ 19

18 How does IR/NIR work for the quantitative analysis? Quantitative analysis - determination of contents of chemical composition or chemical properties. Chemometric modeling is required: MLR, PLS, ANN Creating Calibrations Component A B C Units % % % spectrum spectrum spectrum spectrum spectrum spectrum Samples and data 2. Collect Spectrum 3. Build, Optimize Analyzing Samples & Test Model Report 1. Measure Unknown 2. Access Model Sample # Component A 81.55% Component B 5.38% Component C 13.06% 3. Predict Concentrations 8/ 19

19 Challenges of using IR/NIR Need to find optimal instrument and sampling device for the application Need Chemometrics and spectroscopic experts to build and maintain methods Need lots of representative samples and good quality primary data to build the models Need to have the ability to determine and resolve any issues while using IR/NIR methods. Needs to be user-friendly; ideally can be operated by the plant operator

20 Example Solution of Overcoming These Challenges Internet-Enabled IR/NIR 20

21 Why Internet? Monitor instrument performance remotely Solve problems remotely Store and distribute data automatically Use central models for the prediction Develop or update models remotely Expand applications without limit

22 Why Central Models? All the application models are located at the central server Simple analyzer with a smart brain No instrument-specific models- one model for the same application in different instrument No model transfer or adjustment Consistent prediction instrument variance compensated Plug-and-play No technology limit PCR, PLS, ANN etc. Model update simultaneous for all users

23 Demo of the Internet-Enabled Biodiesel Analyzer ( 23

24 Works behind the Biodiesel Analyzer 24/7 helpdesk service with experts monitoring and maintaining performance Thousands spectra of real world samples with reliable primary data included in each calibration Wide range of feedstocks, including soy (virgin, crude and degummed), canola, rapeseed, poultry, tallow, choice white grease, yellow grease, waste vegetable oil, sunflower oil, castor oil, corn, palm, jatropha and blends thereof Participation in ASTM PTP (Proficiency Testing Programs) Methods underwent full round robin utilizing AOAC and ASTM methodologies

25 QTA Participation in ASTM PTP Total Glycerine % Aug-07 Nov-07 Apr-08 Nov-08 Apr-09 Aug-09 = Robust Mean of reference lab X = Mean of QTA results LABS 48 LABS 51 LABS 65 LABS 63 LABS Robust Mean QTA (%) Blue bar = range of accepted reference lab results for labs 63 LABS

26 Example Calibration Curves on Server Monoglycerides of in-process samples Total glycerin of B100 samples QTA Model Statistics Range, % R 2, % QTA Std. Error, % QTA Model Statistics Range, % R 2, % QTA Std. Error, %

27 EQTA IR Models of Biodiesel Process Incoming Oil Reference Method Range Standard Error FFA, % AOCS Ca 5a Moisture, % KF Moisture In-Process Reference Method Range Standard Error Monoglyceride, % ASTM D Diglyceride, % ASTM D Tri-glyceride, % ASTM D Recovered Methanol Reference Method Range Standard Error Moisture, % KF Moisture Crude Glycerin Reference Method Range Standard Error MeOH, % GC/FID Glycerin, % AOCS Ea Ash, % AOCS Ea Moisture, % AOCS Ea

28 EQTA IR Models of B100 Biodiesel Product for ASTM D6751 and EN14214 Specifications QTA algorithms Reference Method Range Standard Error Free Glycerin # *, % ASTM D Total Glycerin # *, % ASTM D Total Acid Number #, mg KOH/g ASTM D Cloud Point # *, C ASTM D Moisture #, % ASTM D Methanol # *, % EN Oxidative Stability #, Hrs EN Sulfur #, ppm ASTM D Ester, % EN Iodine Value AOCS Cd 1b Density, kg/m 3 ASTM D4052/D Kinematic Viscosity, mm 2 /s ASTM D CFPP, C ASTM D Monoglyceride # *, % ASTM D Diglyceride, % ASTM D Tri-glyceride, % ASTM D # AOCS Ck 2-09 standard procedure * Approved alternative methods included in the ASTM D6751

29 Ethanol Production (Example: Corn Ethanol) Corn receiving and storage Fermentation Cleaning milling/grinding more enzyme -- simpler sugar Mixing: H 2 O/enzyme --corn mash Cooking to reduce bacterial Centrifuge Distillation column Molecular Sieve Ethanol Solids Thin stillage Rotary dryer Evaporator DDGS Syrup

30 Traditional Analysis for Bioethanol Production To ensure the quality of the bioethanol produced, quality tests are conducted for incoming corn, processed corn in the slurry tank, fermentation tank, DDGS (Dried Distillers Grains with Solubles) produced. Time consuming analytical methods have to be used to analyze moisture, oil, protein and starch of the incoming corn. HPLC method is commonly used to analyze carbohydrates (glucose, maltose, maltotrios, dextrin etc.), acids (potential inhibitors), ethanol, and other alcohols for samples from the slurry tank and the fermentation tank.

31 Using NIR for the analysis on in-coming corn Protein, moisture, oil, starch Rejects loads of corn on the basis of moisture Tracks corn suppliers more closely More information about the crop for process optimization

32 Using NIR for the Analysis on Corn Mash Measure ethanol, carbohydrates, acids, and glycerol at any time period during fermentation Quality control Trouble shooting Enzyme evaluations Supplement evaluations

33 Using NIR for fermenter monitoring Typical fermenter profile Increase in ethanol Decrease in dextrins, maltose, dextrose Increase in lactic acid Increase in glycerol

34 A Profile of Ethanol Production during Fermentation % Ethanol Hours of Fermentation

35 A Profile of Carbohydrate Reduction during Fermentation 4.0 % C a rbohy dra te Dextrose Maltose Dextrins Hours of Fermentation

36 A Profile of Lactic Acid Production During Fermentation Lactic Acid is produced by the bacteria - Lactobacillus sp. Competes with yeast for dextrose - reduce yield % La c tic A c id Lactobacillus comes from improper cleaning of fermenters undercooked corn contaminated yeast Hours of Fermentation

37 A Profile of Glycerol Production During Fermentation % G ly c e ro l Produced from yeast when under stress Increased ethanol Increased lactic acid High sugar concentration Low ph High temperature Lack of nutrients (nitrogen) Hours of Fermentation

38 Corn Models QTA Starch Model Statistics Range, %DB R 2, % QTA Std. Error, %DB QTA Moisture Model Statistics Range, % R 2, % QTA Std. Error, % QTA Oil Model Statistics Range, %DB R 2, % QTA Std. Error, %DB QTA Protein Model Statistics Range, %DB R 2, % QTA Std. Error, %DB

39 EQTA NIR Models of Corn and DDGS Corn Range R 2, % Standard Error Oil, %DB Protein, %DB Moisture, % Starch, % DB Corn DDGS Range R 2, % Standard Error Oil, %DB Protein, %DB Moisture, %

40 EQTA NIR Models for Biethanol Slurry Tank Corn Mash form slurry tank Range R 2, % Standard Error DP2 (maltose), % DP3 (maltotrios), % DP4+ (dextrin), % Acetic Acid, % Glucose, % Glycerol, % Lactic acid, % Dextrose, % Total Solids, %

41 EQTA NIR Models for Bioethanol Fermentation Tank Corn Mash form fermentation tank Range R 2, % Standard Error Brix DP2, % DP3, % DP4+, % Ethanol, % Glucose, % Glycerol, % Lactic acid, % ph Total Sugar, %

42 Demo of the Internet-Enabled NIR Analyzer ( 42

43 Summary Productions of biofuels, including biodiesel and bioethanol, need analytical tests at different stages to ensure the quality and yield of the products. Traditional analytical methods are expensive, timeconsuming and need many analytical devices and skills. Spectroscopic methods, IR/NIR, can be used to perform rapid analysis for multiple traits but there are some challenges when using them. An Internet-enabled IR/NIR system has been developed and successfully used in production plants worldwide for the rapid analysis of biodiesel and bioethanol.

44 Thank You Question? 44