Overview of Biomass Compositional Analysis at NREL

Christopher ScarlataASTM E48 MeetingPhoenix, AZDecember 2, 2007

Overview

1. Motivation – Why are we interested in biomass analysis?

2. Role – Who we are and what we do.

3. History – Building on prior methods.

4. Application – Not all biomass is equal.

5. The Method – Summative analysis of biomass.

6. New Tools – Increasing throughput, accuracy, and reducing cost.

7. Rapid Analysis – Near-infrared spectroscopy.

8. Questions

Motivation

Irick, T. J., K. West, et al. (1988). "Comparison of Colorimetric and HPLC Techniques for Quantitating the Carbohydrate Components of Steam-Treated Wood." Applied Biochemistry and Biotechnology 17: 137-149.

Motivation

Irick, T. J., K. West, et al. (1988). "Comparison of Colorimetric and HPLC Techniques for Quantitating the Carbohydrate Components of Steam-Treated Wood." Applied Biochemistry and Biotechnology 17: 137-149.

Apologies to the lignin chemists in the Apologies to the lignin chemists in the audience!audience!

Biomass Analysis Section

Biomass Analysis Expertise

• Analytical chemistry• Biochemistry• Chemometrics

Develop and apply standard methods for analysis of biomass feedstocks and process intermediates.

History

• Summative analysis

• Micro method– Furda, I; Simultaneous analysis of soluble and

insoluble dietary fiber, The Analysis of Dietary Fiber in Food, W.P. T. James and O. Theander(Eds), Marcel Dekker, New York, 1981, pp.163 –172.

• Used for Validation of NIST biomass Standard Reference Materials in International Energy Agency sponsored Round Robin– Milne, T. A.; Chum, H. L.; Agblevor, F. A.; Johnson, D.

K.; Standardized Analytical Methods, Proceedings of International Energy Agency Bioenergy Agreement Seminar, Vol 2(1-6), April, 1992, (341-366).

NREL LAPs are based on the Uppsala Method for the analysis of dietary fiber.

NIST Standards

Some NREL LAP’s cite other methods

Laboratory Analytical Procedure ReferenceSummative Mass Closure for Biomass Samples -Determination of Acid Soluble Lignin Concentration Curve by UV-Vis Spectroscopy -Determination of Structural Carbohydrates and Lignin in Biomass ASTM E1758-01Determination of Extractives in Biomass ASTM E1690Preparation of Samples for Compositional Analysis ASTM E 1757 – 01; TAPPI T264; NFTA A 1.1Determination of Total Solids in Biomass ASTM E1756-01; T412 om-02Determination of Ash in Biomass ASTM E1755-01Determination of Sugars, Byproducts, and Degradation Products in Liquid Fraction Process Samples -Determination of Starch in Biomass Samples by HPLC AOAC 996.11; AACC 76.13; ICC 168.Determination of Protein Content in Biomass Mossé (1990)Rounding and Significant Figures -Determination of Insoluble Solids of Pretreated Biomass Material -Measurement of Cellulase Activities IUPACEnzymatic Saccharification of Lignocellulosic Biomass -SSF Experimental Protocols: Lignocellulosic Biomass Hydrolysis and Fermentation -Standard Test Method for Moisture, Total Solids, and Total Dissolved Solids in Biomass Slurry and Liquid Process Samples TAPPI OS-63Determination of Ethanol Concentration in Biomass to Ethanol Fermentation Supernatants by Gas Chromatography -

Laboratory Analytical Procedures (LAPs)

Available at http://nreldev.nrel.gov/biomass/analytical_procedures.html

Biomass Analysis LAPs

• Sample preparation• Extraction• Total solids• Carbohydrates and lignin• Inorganics (ash)• Protein• Starch• Hydrolysate liquor analysis• SSF (solids) and fermentation samples

Calculation Spreadsheets

• Excel format– Automatically performs all necessary

calculations (hundreds)– Provides user-friendly guidelines– Duplicate and averaged values– Error flags page

• Applications–Wood–Corn stover–Hydrolysate liquor–Acid pretreated solids–One step extraction corn fiber–Two step extraction corn fiber

Flow Chart of Biomass Stover Feedstock Analysis

Representative, Prepared Whole Corn Stover (R-P Stover)

sucrose cellulosewax hemicellulosechlorophyll lignin

proteininorganics

H2O solubles Structural Stover

sucrose waxprotein chlorophyll EtOH Extractionnitrate/nitrites cellulosesoluble inorganics hemicellulose

ligninproteinstructural inorganics

EtOH solubles Extractives-Free Stover (E-F Stover)

wax cellulosechlorophyll hemicellulose

lignin 1) 72% H2SO4 Hydrolysis protein 2) 4% H2SO4 Hydrolysis structural inorganics

Hydrolysis liquor #1 Acid Insoluble Residue

glucose galactose acetic acid acid insoluble ligninxylose mannose uronic acids proteinarabinose acid soluble lignin structural inorganics

H2O Extraction

Chemical Analysis Gantt Chart

• Planning is key for successful analysis

• Labor intensive – Are there opportunities for increased throughput?

batch size Operation Fraction of day day 1 day 2 day 3 day 4 day 5 day 6 day 7 day 8 day 9 day 10 day 11 day 12Washing 0.5 0.5Drying 0.25 0.25Milling /Sieving 0.5 0.5Homogenizing 0.1 0.1

Prep work 0.3 0.3ASE Extraction 0.1 0.1Sucrose measruement 0.5 0.5Solvent Removal 0.1 0.1Solids Drying 0

Prep 0.25 0.25Hydrolisis 0.65 0.65UV lignin determination 0.35 0.35Ashing, whole and insols 0.35 0.25 0.1Neutralization 0.25 0.25LC prep 0.15 0.15LC Run time 0LC workup 0.25 0.25Data Entry / Review 0.5 0.5

0.75 0 0 0 1 0.6 0.5 1 0.5 0 0 0.75total: 5.1

Assumes 3 samples w/ size of ~1-2 cft

Assumes a max of 5 samples run in duplicate with standards

Assumes a max of 3 samples run in duplicate

Method Improvement

New Methods

Data Management and QA/QC

Automation

- Accelerated Solvent Extractor (ASE); Dionex Corp.

• Tested and implemented in FY04/05

• Old method (Soxhlet) = 24+ hours

• ASE method = 1 hour

Washing, Drying, Chopping, Milling, Homogenizing

Whole Corn Stover

(as received)

H2O Extraction andEtOH Extraction

HPLC analysis

Drying, ashing,

weighing

Weigh Solids

2-Stage Acid Hydrolysis

Package liquid

samples in barcode

labeled vials

Acid insoluble solids transfer

to filtering crucible

Crucibles with solids go on for further

analysis

Wash solids

Filter liquids

Neutralize? Y/N

Automation

Automated Extraction

Composition of Stover Water ExtractsM

ass

Perc

ent E

xtra

ctiv

es

0%

20%

40%

60%

80%

100%

Kramer 2302-115 2870-061 2302-079 2893-026

unknowninorganic anions

total sugar alcoholsaliphatic acidsinorganic cationsred-brown fractionoligomeric sugarmonomeric sugar

Chen, S. F., R. A. Mowery, et al. (2007). "Compositional analysis of water-soluble materials in corn stover." Journal of Agricultural and Food Chemistry 55(15): 5912-5918.

14% 20% 16% 27% 14% extractives in feedstock

(dry

)

Composition of Stover Water ExtractsM

ass

Perc

ent E

xtra

ctiv

es

0%

20%

40%

60%

80%

100%

Kramer 2302-115 2870-061 2302-079 2893-026

unknowninorganic anions

total sugar alcoholsaliphatic acidsinorganic cationsred-brown fractionoligomeric sugarmonomeric sugar

Should this become a standard method?

14% 20% 16% 27% 14% extractives in feedstock

(dry

)

Washing, Drying, Chopping, Milling, Homogenizing

Whole Corn Stover

(as received)

H2O Extraction andEtOH Extraction

HPLC analysis

Drying, ashing,

weighing

Weigh Solids

2-Stage Acid Hydrolysis

Package liquid

samples in barcode

labeled vials

Acid insoluble solids transfer

to filtering crucible

Crucibles with solids go on for further

analysis

Wash solids

Filter liquids

Neutralize? Y/N

Automated Biomass Compositional Analysis

Automation

Already automated by Dionex

ASE

Justification/Purpose:

• Build rapid NIR/PLS methods – Target ~1 new NIR model/2 months

• Will reduce sample analysis cost.

Need: An automated R&D workstation customized for use in higher throughput biomass chemical analysis.

NREL RFP:

• On the street as of 11/21/07

• Sept. ’08 Milestone

FY07: Developed Faster Analytical Method for Measuring Ethanol and Other ProductsFast HPLC analysis of fermentation products and furans

• Analysis of multiple compounds in under 10 min versus 55 min by old method• E Milestone: Validate and Publish Fast Acid Method (Feb. ’08)

Old Method

New ‘FAST ACID’ MethodNew Method Old Method

Acetic acid 3 15Ethanol 5 22HMF 6 32Furfural 8 50

Analysis time (minutes)

Ethanol

HPLC for Carbs.

Column Stationary phase

Mobile phase

Particle sizes

available

pH tolerance

Max temp. (oC)

Max flow rate

(mL/min)

Max press.

Shodex SP0810

Pb++ type resin Water 7 µm 5 – 9 95 1.0

(at 8 mm id) 30 bar

(430 psi)

TSK-GEL Amide-80

carbamoyl groups

bound to silica

ACN:H20 Plus

modifiers 3 or 5 µm 2 – 7.5 80

1.2 (at 4.6 mm id and 5 µm particles )

150 bar (2200 psi)

Asahipak NH2P-50

amino groups in polyvinyl

alcohol

ACN:H20 Plus

modifiers 5 µm 2 – 13 50

1.5 (at 4.6 mm

id)

90 bar (1300 psi)

ZirChrom WAX

zirconia coated w/

cross-linked polyethylene-

imine

ACN:H2O:MeOH 3 or 5 µm 3 – 9 50

2.0 (at 4.6 mm id and 5 µm

particles)

Not available

Luna NH2 amino groups

bound to silica

ACN:H20 Plus

modifiers 3 or 5 µm 1.5 – 11 n/a

3.0 (at 4.6 mm id and 5 µm

particles)

Not available

HPLC for Carbs.

Column Stationary phase

Mobile phase

Particle sizes

available

pH tolerance

Max temp. (oC)

Max flow rate

(mL/min)

Max press.

Shodex SP0810

Pb++ type resin Water 7 µm 5 – 9 95 1.0

(at 8 mm id) 30 bar

(430 psi)

TSK-GEL Amide-80

carbamoyl groups

bound to silica

ACN:H20 Plus

modifiers 3 or 5 µm 2 – 7.5 80

1.2 (at 4.6 mm id and 5 µm particles )

150 bar (2200 psi)

Asahipak NH2P-50

amino groups in polyvinyl

alcohol

ACN:H20 Plus

modifiers 5 µm 2 – 13 50

1.5 (at 4.6 mm

id)

90 bar (1300 psi)

ZirChrom WAX

zirconia coated w/

cross-linked polyethylene-

imine

ACN:H2O:MeOH 3 or 5 µm 3 – 9 50

2.0 (at 4.6 mm id and 5 µm

particles)

Not available

Luna NH2 amino groups

bound to silica

ACN:H20 Plus

modifiers 3 or 5 µm 1.5 – 11 n/a

3.0 (at 4.6 mm id and 5 µm

particles)

Not available

Peak resolution (R) is proportional to (1/dp)^½

If 7? 3µm, Then R? 50%

Snyder, L. R. A., J. J. A. Kirkland, et al. (1997). Practical HPLC Method Development. Boulder, NetLibrary.

Resolution using the current HPLC column

(Shodex SP0810)

Calculation of resolution of a Carb_36 CVS sample

Analyte Peak Retention Time Width @ half height Resolution Res. Calc.Cellobiose 1 13.3 0.54 1 and 2 2.5Glucose 2 15.6 0.53 2 and 3 1.3Xylose 3 16.7 0.49 3 and 4 1.4Galactose 4 17.9 0.47 4 and 5 1.9Arabinose 5 19.4 0.48 5 and 6 1.7Fructose 6 21.0 0.56

Representative, Prepared Whole Corn Stover (R-P Stover)

sucrose cellulosewax hemicellulosechlorophyll lignin

proteininorganics

H2O solubles Structural Stover

sucrose waxprotein chlorophyll EtOH Extractionnitrate/nitrites cellulosesoluble inorganics hemicellulose

ligninproteinstructural inorganics

EtOH solubles Extractives-Free Stover (E-F Stover)

wax cellulosechlorophyll hemicellulose

lignin 1) 72% H2SO4 Hydrolysis protein 2) 4% H2SO4 Hydrolysis structural inorganics

Hydrolysis liquor #1 Acid Insoluble Residue

glucose galactose acetic acid acid insoluble ligninxylose mannose uronic acids proteinarabinose acid soluble lignin structural inorganics

H2O Extraction

Other areas need improvement

Uronic Acids –•Based on a assumption UA = xylan/6 Polymers in Nature, E.A. MacGregor and C.T. Greenwood, 1990, New York, New York: John Wiley & Sons, Ltd. p 296-297•Need a source of standards

ASL –•Extraction of ASL has low yield and variable results

Rapid Biomass Analysis – NIR Spectroscopy

Predicted vs Measured Constituent Values for calibration samples- stover9.eqa- major constituents

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25 30 35 40 45

Wet chemistry values (%)

NIR

Pre

dic

ted

val

ues

(%)

GlucanXylanLignin_p_corr

Protein

•Current Methods•Corn Stover Feedstock•Hardwood•Softwood•Pretreated Corn Stover solid fraction (washed, dried, and milled)

•Wet chemical methods are used to calibrate rapid analysis methods

•Retain precision and accuracy of calibration methods

•Fast/Less labor intensive•Inexpensive for routine samples

Method BuildingCalibration Samples– Robust model requires ~100 well-characterized

samples– Calibration samples should reflect the composition and

variance expected in test samplesChemical Characterization– Determines precision and accuracy of new method– Requires appropriate analytical methods– Should be robust, reproducible, and sensitive to

compositional differencesMultivariate Analysis Tools– Translates spectroscopic data into compositional dataQA/QC– Calibration checks (well characterized “blind” samples

or standard reference materials)– Outlier flags

Intellectual Property

• At the end of FY08, we will have several (4?, 6?, 8?) NIR calibration models; many researchers want these!

• The unique combination of wet chemistry data, spectral data, and calibration equation(s) appears to be NREL/MRI/Battelle IP

• Any such IP generated using DOE funding (in whole or in part) can and should be protected by a copyright rather than a patent

• How do we distribute these equations without creating an unfunded mandate (“Thanks for calling the NREL NIR Hotline…”)

Questions?Biomass Analysis @ NREL

•David Crocker•Steve Decker•Erik Fisk•Deb Hyman•David Johnson•Bill Michener•Courtney Payne•Darren Peterson•Ray Ruiz•Chris Scarlata•Mike Selig•Amie Sluiter•Justin Sluiter•David Templeton•Jeff Wolfe•Ed Wolfrum•Millie Zuccarello

Prairie grass to biofuels; featured in Science.