JOURNAL OF WOOD CHEMISTRY AND TECHNOLOGY, 18(2), 235-252 (1998)
A RAPID MODIFIED METHOD FOR COMPOSITIONAL CARBOHYDRATE ANALYSIS
OF LIGNOCELLULOSICS BY HIGH PH ANION-EXCHANGE CHROMATOGRAPHY
WlTH PULSED AMPEROMETRIC DETECTION (HPAEC/PAD)
Mark W. Davis
USDA Forest Service
Forest Products Laboratory
One Gifford Pinchot Drive
Madison, Wisconsin 53705
ABSTRACT
During the last decade, high pH anion exchange chromatography with pulsed
amperometric detection (HPAED/PAC) has gained increasing acceptance as the method
of choice for analysis of neutral sugars commonly occurring in woods, pulps, and other
lignocellulosics. This paper describes modified chromatographic conditions and discusses
other critical factors that improve the precision and efficiency of this application. The
method involves a controlled loading of acetate onto the column prior to equilibration with
water and injection of sample. In-line solid-phase extraction is used to remove
hydrophobic substances that have the potential to foul the analytical column. Critical
operational parameters for the successful application of the method include a metal-free
flowpath and a consistent application of anions with sample. Resolution of rhamnose is
achieved while maintaining the resolution of xylose and mannose. Simplified sample
pretreatment allows a ca. five-fold increase in sample through-put compared with gas
chromatography of derivatized sugars or to partition chromatography. Run times are less
than half those of the widely used hydroxide reverse gradient method for HPAEC/PAD
analysis of wood sugars. Long-term system performance data indicate that the method is
highly precise and robust. The acetate loading method affords better precision than those
of other HPAEC/PAD methods and of gas chromatographic analysis of alditol acetate
derivatives by Tappi Method T249 cm-85
235
Copyright © 1998 by Marcel Dekker, Inc.
www.dekker.com
236
DAVIS
INTRODUCTION
The compositional analysis of lignocellulosics involves acid hydrolysis, typically with
sulfuric acid,
1,2
followed by quantitation of the resulting neutral sugar monomers in the
hydrolysates by chromatographic means. Alternatively, enzymatic hydrolysis can be
employed for the compositional analysis of soluble polysaccharides
3
or in cases wherein
the action of enzymes on soluble or insoluble substrates is of interest. Three methods are
commonly employed for the chromatographic analysis: (1) gas chromatography (GC) of
alditol acetate,
2
trimethylsilyl, or per-
O
-acetylated aldononitrile derivatives,’ (2) partition
chromatography (PC) on cation exchange resins with refractive index detection
5-7
or
(3) high pH anion-exchange chromatography with pulsed amperometric detection
(HPAEC/PAD).
3,8-15
Although results obtained using GC
14
or PC
9
methodologies are comparable with
HPAEC/PAD results, the former techniques require multiple sample preparation steps that
decrease their efficiency. For sulfuric acid hydrolysates, both methods require
neutralization and sulfate removal, typically using barium hydroxide and centrifugation.
Subsequent steps for GC analysis typically include reduction with sodium borohydride
and derivatization, with attendant concentration and sample clean-up procedures, and
finally sample filtration. Subsequent steps for PC analysis include ash removal,
concentration, and filtration steps.
In laboratories utilizing HPAEC/PAD, sample preparation varies widely. Some
analysts assume it is necessary to neutralize acid hydrolysates
14,16,17
and remove
sulfate
8,15,18
prior to chromatographic analysis, and others
9-12
omit these steps. In addition,
solid-phase extraction (SPE) is essential for some types of samples
14
and is widely
8,10,15,16
(although not universally)
3,11-13
employed. A major limitation of HPAEC/PAD is the lengthy
chromatographic run time, which including column conditioning and re-equilibration steps,
is typically 60 min or more.
9-11
Shorter run times have recently been reported,‘* but efforts
to
reproduce this method have not been successful.
13
A second limitation is the difficulty
of resolving the trio of minor hemicellulosic sugars, arabinose, galactose, and rhamnose,
while maintaining resolution of the xylose-mannose pair.
10,13,16,17
A third limitation is poor
quantitation of low amounts of mannose, because of the common tendency of late eluting
peaks to tail excessively.
11,13
This report describes the operational parameters in use at the USDA Forest Service,
Forest Products Laboratory (FPL), which have increased the efficiency and precision of
COMPOSITIONAL CARBOHYDRATE ANALYSIS
237
HPAEC/PAD for the routine analysis of a diverse stream of woods, wood products, and
other lignocellulosic samples. Following filtration, sulfuric acid or enzyme hydrolysates are
injected directly with no additional sample preparation steps. Matrix hydrophobic
components are removed by in-line solid-phase extraction (SPE). A reverse gradient
method wherein acetate loading is performed during column conditioning is used to
achieve near baseline resolution of the five classic wood sugars and rhamnose. Sugars
are eluted within 10 min, with a total run time of 27.5 min. The described method was
developed and implemented in June 1994, and has been used to analyze more than
4,000 samples. System performance data collected as part of the FPL quality control
program is presented to document the precision and robustness of the method.
EXPERIMENTAL
Materials
NaC
2
H
3
O
2
· 3H
2
O (Ultrapure grade) and carbonate-free NaOH were obtained from
J.T. Baker.
(1)
Water (>17.8 mOhm/cm) was obtained from a point of use polishing system
(Barnstead/Thermolyne, Dubuque, IA). Concentrated H
2
SO
4
was obtained from EM
Science (Gibbstown, NJ). Wood sugar standards and fucose (internal standard) were
obtained from Aldrich Chemical Co. (Milwaukee, WI). The wood sugars are arabinose,
galactose, rhamnose, glucose, xylose, and mannose. Purities, as certified by certificates
of analysis from the vendor, were all in excess of 97.5%. Standard concentrations were
determined gravimetrically, with no correction for impurities. Samples of aspen (
Populus
tremuloides, Michx.) and loblolly (Pinus taeda, L.), provided by Dr. Masood Ahktar of FPL,
were from freshly felled trees that were debarked, chipped, and the chips stored frozen.
Kenaf (Hibiscus cannabinus, L.) bast was provided by James Han of FPL. The pulp
sample, a bleached kraft softwood pulp, was provided by David Bormett of FPL. A large
supply of each sample was air dried, milled to pass a 1.00-mm screen, mixed well, and
(1)
The use of trade or firm names is for information only and does not imply endorsement by the
U.S. Department of Agriculture of any product or service.
238
DAVIS
stored in sealed polyethylene bags at room temperature for use as quality control
performance samples.
Hydrolysis Conditions
Hydrolysis in H
2
SO
4
was carried out essentially as described elsewhere.
1,19
Briefly,
samples were milled to pass a 1.00-mm screen and vacuum dried at 45°C. Primary
hydrolysis of 40-60 mg subsamples was performed with 1.00 mL 72% (w/w) H
2
SO
4
for
1 hr at 30°C. Hydrolysates were diluted to 4% (w/w) H
2
SO
4
with distilled water, fucose
added
as an internal standard, and a secondary hydrolysis performed for 1 hr at 120°C.
To correct for sugar degradation during secondary hydrolysis, system calibration was
based upon a standard mixture of sugars treated in parallel with each batch of samples.
Losses during primary hydrolysis are minimal and are ignored.’ Following filtration
through 0.45 µm Teflon syringe filters (National Scientific, Lawrenceville, GA), 5 µL
samples of hydrolysates were injected directly onto the chromatographic system with no
additional treatment. In some cases, parallel lignin determinations were performed. In
these cases, 80-140 mg samples were hydrolyzed, hydrolysates diluted to 100 mL with
H
2
O, and 10-15 µL portions injected.
Chromatoqraphic Analysis
Sugar contents of hydrolysates were determined by HPAEC/PAD. The
chromatographic system consisted of a 738-autosampler (Alcott Chromatography,
Norcross, GA), a GPM-1 or a GP40 gradient high pressure pump (Dionex Corp.,
Sunnyvale, CA), and a pulsed amperometric detector (PAD) (Dionex). The entire flowpath
of this system was metal free with the exception of the injection needle (titanium alloy)
and
injection valve (Hastelloy C). Hydrophobic materials were removed by in-line SPE,
employing an NG1 guard column (Dionex) plumbed between the injection valve and
Carbo-Pac columns. A time-programmed valve switching event diverted flow around the
NG1 column and autosampler 1 min after sample injection. The NG1 guard column was
washed with methanol and re-equilibrated with H
2
O after every ca. 100 injections. Sugar
COMPOSITIONAL CARBOHYDRATE ANALYSIS
239
separation was achieved with Carbo-Pac PA1 guard and analytical columns (Dionex)
connected in series. Eluent flow rate was 1.2 mL/min and the temperature was 22°C. The
reverse gradient method consisted of elution with H
2
O for 11 min, followed by a 1-min
ramp to 170 mM NaC
2
H
3
O
2
in 200 mM NaOH, which was maintained for 5 min, then a
1 min return to the original H
2
O eluent 9.5 min prior to the next injection. These conditions
were employed to elute neutral sugars, condition the column, and re-equilibrate the
column, respectively. The time of equilibration is a critical method parameter, therefore,
data from the first run of the day were not used. Sugars were detected by their oxidation
at a gold electrode surface of the PAD. PAD settings were E1 = 0.1 V, E2 = 0.9 V, and
E3 = -0.6 V for durations of 300, 120, and 300 msec, respectively. Output range was set
to 10,000 nA/V. To facilitate the pH sensitive oxidation of carbohydrates, 0.30 M NaOH
was added to the post-column eluent stream at a flow rate of ca. 0.3 mL/min. Detector
output was digitized by a Series 900 analog interface (PE Nelson, Cupertino, CA) at a
sampling rate of one point/sec and stored and analyzed using Apex ver. 3.14 software
(Autochrome, Inc., Milford, MA).
Analysis of Performance Data
System performance (HPAEC/PAD) was monitored by the performance of sugar
standards. Standard mixtures, whose sugar concentrations were similar to those of typical
sample hydrolysates (Tables 1 and 2) were subjected to secondary hydrolysis in parallel
with each hydrolysis batch (typically 18 to 24 samples) and intermittently injected
(typically 6 to 8 repetitions) during HPAEC/PAD analysis of each batch. Because the
measure of interest in this analysis is the mass of the anhydrous sugar unit as it exists in
the sample, all concentrations are given in terms of the anhydrous weight equivalent of
the free sugar (e.g., gravimetric mass of glucose is adjusted by 162/180).
Chromatographic performance was assessed by the standard deviations (SD
S
) of
retention times and resolution factors among all standard injections within a period during
which the system configuration was not changed. Resolution factors were calculated as
the ratio of two times the difference in component retention times with the sum of their
peak widths at baseline, or 2 (RT
2
- RT
1
) / (W
1
+ W
2
). System noise was calculated as
three times the SD around the chromatographic baseline, as determined by regression