United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 2771
Research and Development
EPA/600/S8-86/037 Feb. 1987
&EPA Project Summary
Sampling and Analysis for
High-Molecular-Weight Polar
Organic Compounds
Ruby H. James, Robert E. Adams, and Herbert C. Miller
This report gives results of prelimi-
nary investigations into the determina-
tion of high-molecular-weight polar
organic compounds from wood com-
bustion residues. K is intended as a ref-
erence to be used by laboratories that
are developing methods for the identifi-
cation and quantification of high-
molecular-weight compounds.
This Project Summary was devel-
oped by EPA's Air and Energy Engineer-
ing Research Laboratory, Research Tri-
angle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report or-
dering information at back).
Introduction
No universal analytical method cur-
rently exists in the field of trace organic
analysis that covers the more polar, less
volatile components of environmental
samples. While capillary gas-
chromatographic methods may be suf-
ficient for substances that are relatively
volatile, there remain many large and
polar molecules that have increasing
environmental significance. With in-
creasing molecular weight, the number
of isomeric compounds to be resolved
also increases proportionally. The
methods for their determination have
not yet been fully developed. Although
high-performance liquid chromatogra-
phy (HPLC) is capable of handling such
large and polar molecules, improve-
ments in the applications of this tech-
nique will be required for the character-
ization of high-molecular-weight
fractions.
The proposed HPLC method is capa-
ble of giving functional-group class sep-
aration for semipreparative sample
fractionation, and it may prove to be a
useful analytical technique for the more
polar constituents (pyrrolic, -OH,
-COOH, and -NH2 substituted). But
these compounds constitute only about
40% (w/w) of the sample analyzed in
this study. About 36% of the sample
was judged to contain less polar com-
pounds (carbonyl, nitro, and aza com-
pounds). The remaining 24% were non-
polar compounds (hexane extract-
ables). Unfortunately, the method did
not distinguish between the alkyl- and
benzyl-substituted proton-donating
compounds and the nonionizable com-
pounds. It appears that the higher the
molecular weight, the greater the alkyl
substitution; or the more ring conjuga-
tion present, the more inadequate the
method becomes as an analytical tech-
nique.
The objective of this task was to de-
velop methods for determining high-
molecular-weight polar organic com-
pounds derived from coal, synfuel, and
wood combustion. The existing
methodology was reviewed, and the an-
alytical methodology was developed
leading to the separation of as many of
the high-molecular-weight species as
possible in residues from wood com-
bustion.
The proposed method was expected
to be based on initial separation by liq-
uid chromatography, which included
such methods as liquid-solid chro-
matography, gel-permeation chro-
matography, or reversed-phase liquid
chromatography. The use of the pro-
-------
posed method resulted in major sub-
fractions and, ultimately, fractions con-
taining fewer compounds. Qualitative
analysis was expected to be based on a
combination of techniques.
Fast atom bombardment/mass spec-
trometry (FAB/MS) and direct-insertion-
probe/mass spectrometry (DIP/MS)
were used to characterize fractions. The
samples were also screened by Fourier-
transform infrared spectroscopy (FTIR).
Open-Column Liquid
Chromatography
Initial separation of the sample into
functional-group classes was attempted
using a silica-gel open-column liquid
chromatography procedure. The four
fractionation schemes progress from
nonpolar (hexane) solvents to very
polar (methanol and dimethyl su If ox-
ide) solvents in attempts to separate the
woodburn samples into functional
group classes. Methylene chloride was
a moderately polar solvent used in com-
bination with the nonpolar and very
polar solvents.
HPLC Instrumentation
The HPLC system (Model 1090A) from
Hewlett-Packard (Palo Alto, CA) con-
sisted of a binary-gradient solvent-
delivery system, diode-array detector,
automatic liquid sampler, and a
variable-column injector.
HPLC Chromatographic
Methods
1. Column: Waters (j,-Styragel, 100
A, 300 x 7.8-mm ID
Solvent: Tetrahydrofuran
Flow: 1 mL/min, isocratic
2. Column: ASI Ultragel, 100 A,
250 x 7.8-mm ID
Solvent: Tetrahydrofuran
Flow: 1 mL/min, isocratic
3. Column 1: Waters jt-Styragel, 100
A, 300 x 7.8-mm ID
Column 2: AS1 Ultragel, 100 A,
250 x 7.8-mm ID
Solvent: Tetrahydrofuran
Flow: 1 mL/min, isocratic
4. Column: Whatman Partisil,
ODS-3, 10 fjim,
250 x 4.6-mm ID
Solvent A: Methanol
Solvent B: Methylene Chloride
Flow: 1 mL/min, gradient
Program: 0% B to 100% B in
40 min.
5. Column:
Solvent:
Flow:
6. Column:
Solvent A
Solvent B:
Flow:
7. Column:
Solvent A:
Solvent B:
Flow:
Program:
Waters ft-Bondapak-
NH2, 10 (Jim, 300 x 3.9-
mm ID
n-Heptane
0.5 mL/min, isocratic
Waters (i-Porasil-Si,
10 (Jim, 300 x 3.9-mm
ID
Carbon Tetrachloride
Dimethyl Sulfoxide
1 mL/min, isocratic at
0.1%, 1.0%, 10%, or
25% B
Waters jji-Porasil-Si,
10 |xm, 300 x 3.9-mm
ID
Carbon Tetrachloride
Dimethyl Sulfoxide
1 mL/min gradient
0% B to 1% B in 10
min, then to 10% B in
10 min
8. Column: Chromanetics
Spherisorb-Si, 5 i^m,
250 x 4.6-mm ID
Solvent A: Carbon Tetrachloride
Solvent B: Dimethyl Sulfoxide
Flow: 1 mL/min, isocratic
Method
option:
A = 25% B
B = 10%B
C= 1.0% B
D = 0.1% B
Initial Sample Characterization
Four woodburn samples were re-
ceived from EPA for the analysis of
high-molecular-weight polar organic
compounds:
243B 104.08 mg
243C 100.13 mg
233C 99.65 mg
237D 99.34 mg
Samples 243B and 243C appeared to be
duplicates, as did Samples 233C and
237D. All were black, tarry substances
coated on the walls of the sample vials.
The original samples were submitted
for EI/MS, FAB/MS, and FTIR character-
zation.
FTIR samples were dissolved in
methanol and analyzed as a KBr pellet.
The spectra obtained on a Nicolet MX-IE
in this manner were not intense enough
to identify functional groups. The sam-
ples should be examined using a higher
concentration in the KBr pellet or as a
film on a salt plate. However, Samples
243B and 243C appear to have similar
spectra. Samples 233C and 237D also—
have similar FTIR spectra. •
The El mass spectra obtained with a
Varian Mat 311A mass spectrometer
showed peaks only below m/z 500. The
EI/MS obtained in this manner would
not be useful for analysis of higher-
molecular-weight polar compounds.
FAB/MS was obtained with a Varian
Mat 311A mass spectrometer. Peaks up
to m/z =1500 were observed in the FAB/
MS analysis; however, these spectra
were very complex. Some representa-
tive peaks for the original samples from
the FAB/MS analysis are listed in
Table 1.
Open-Column Liquid
Chromatography
The spectra obtained by FTIR, EI/MS,
and FAB/MS of the samples indicated
many compounds of different molecu-
lar weights and functional groups.
Open-column liquid chromatography
may be used to separate materials ac-
cording to functional groups. Several
fractionation schemes were developed
in attempts to separate the woodburn
samples into less complex fractions.
The open-column liquid chromatog-
raphy fractionation schemes yielded
residues yellow to brown in color. The
FAB/MS of these residues contained
predominantly low-mass ions. How-
ever, some peaks were evident up to
m/z 1500. Table 2 gives representative
m/z values for three polar fractions of
Sample 243B.
Direct-Insertion-Probe/Mass
Spectrometry Results from
Scheme 4
An unfractionated portion of Sample
233C and Fractions 1, 2, 3, 4, 5, and 7 ol
LC fractionation Scheme 4 were ana-
lyzed using direct-insertion-probe/mass
spectrometry (DIP/MS). All of the Frac
tion 6 sample was placed in glycerol foi
FAB/MS and thus was not analyzed b\
DIP/MS. Between 10 and 50 |xg of eacl-
sample, excluding solvent, was placec
in capillary-glass sampling tubes, th«
solvent evaporated, and the residut
analyzed. Representative m/z values fo
the original sample from the DIP/MJ
analysis and the individual fractions an
summarized in Table 3. The peaks at m/,
167 and 181 may be indicative of carba
zole and carbazole derivatives. Thesi
polar polynuclear aromatic hydrocar
bons (PAHs) may be present in the sam
pie. Further work will better characteriz
the fractions and will possibly identif
more compounds.
-------
^Table 1. Representative Significant m/z Values and Approximate Relative
P FAB/MS Analysis"
Sample 243B Sample 233C Sample 243C
m/z m/z m/z
300(101" 304(15) 387(5)
304(10) 308(15) 398(5)
316(5) 310(12) 973(7)
335(5) 412(7)
360(5) 414(5)
380(4) 425(5)
393(4) 480(4)
395(4) 503(3)
398(4) 550(3)
400(4) 592(3)
412(3) 642(2)
415(3)
AtOI'l}
*t IO|O/
436(3)
"Many peaks were seen at relative intensities of 2 to 3% above m/z 1000.
Mpprox/mate percent relative intensities given in parentheses.
Intensities by
Sample 237D
m/z
300(15)
310(15)
314(15)
325(10)
338(10)
342(10)
350(8)
353(8)
365(8)
381(8)
398(8)
413(8)
.JO/VC)
fji/|3/
438(5)
497(3)
Table 2. Representative Significant m/z Values for Approximate Relative Intensities by Pos-
itive and Negative FAB/MS Analysis for Sample 243B
Fraction 5 Fraction 6
Positive Negative Positive Negative
343"(10)a 297"(10) 429t>(15) 297»(8)
345b(30) 299b(35) 431 "(30) 294b(8)
347b(15) 301b(25) 433"(20) 301»(8)
435h(5) 392(5) 521»(7)
437"(12) 416(4) 523b(10)
439"<10) 498(3) 525"(8)
444(7) 530(5)
446(7) 540(5)
461(7) 545(5)
485(5) 556(5)
506(5) 558(5)
525(5) 560(5)
667(4) 562(5)
722(3) 582(5)
740(3) 584(4)
853(3) 588(4)
647(4)
U~Tf \*Tf
674(3)
704(3)
752(3)
894(3)
1002(3)
Fraction 7
Positive
431(70)
445(50)
463(30)
486(20)
523(20)
548(15)
585(15)
598(10)
650(8)
684(8)
715(7)
795(5)
803(4)
820(4)
846(4)
873(4)
890(4)
897(4)
925(3)
1010(3)
"Approximate percent relative intensities given in parentheses.
bThese m/z's have been identified as abietic acid, or derivatives ofabietic acid (mot wt 302.4),
a component of wood rosin.
HPLC Fractionation
The first approach was to fractionate
the sample by molecular weight on a
gel-permeation column (Waters ^-
Styragel, 100 A, 300 x 7.8-mm ID) and
then to analyze each subfraction on a
normal-phase column (Waters n,-
Porasil-Si, 10 (un, 300 x 3.9-mm ID). Be-
cause one gel-permeation column did
not provide adequate separation of the
original sample matrix for fractionation,
a second GPC column (AS1 ultragel, 100
A, 250 x 7.8-mm ID) was used in series
with the (ji-Styragel column. Comparing
the original sample with standard mix-
tures of polystyrene at known
molecular-weight ranges showed that
the sample was composed of polystyre-
nes with molecular weights ranging
from ROD tn 1flnf)
It \Jltl *JW t.\f 1 wvw.
The Waters jx-Porasil column was
used to investigate the elution order
with selected model compounds
(pyridine, phenol, aniline, quinoline,
carbazole, and dibenzo(a,j)acridine) at
various ratios of the additive (DMSO) to
the eluent base (CCI4) with only mar-
ginal success. Peaks within similar com-
pound classes were not resolved ade-
quately. In addition, the less polar
groups overlapped badly. An attempt
was made to develop a long gradient
program, adjusting the amount of addi-
tive to base eluent. But as the amount of
additive was increased, the amount of
background sharply increased. Even
though the UV cutoff of DMSO is 268
nm, measurement at 276 to 300 nm is
drastically affected, especially above
the 5% level of DMSO.
The proposed chromatographic sys-
tem (Method 8) is capable of giving suf-
ficient class separation for semiprepara-
tive sample fractionation. And it may
prove to be a useful analytical technique
for the more polar constituents
(pyrrolic, -OH, -COOH, and -NH2 sub-
stituted). But these compounds consti-
tute only about 40% (w/w) of the wood-
burn Sample 233C. About 36% of the
sample was judged to contain less
polar, nonionizable compounds (car-
bonyl, nitro, and aza compounds). And
the remaining 24% were nonpolar com-
pounds (hexane extractables). Unfortu-
nately, it was impossible to distinguish
between the alkyl- and benzyl-
substituted proton-donating com-
pounds and the nonionizable com-
pounds. As a result, the higher the
molecular weight, the greater the alkyl
substitution, or the more ring conjuga-
tion present, the more inadequate the
-------
method becomes as an analytical tech-
nique.
However, the two methods (open-
column fractionation followed by
normal-phase chromatography) pro-
vide a good basis for the analysis of the
more polar constituents in environmen-
tal samples. More success is expected
in the resolution and identification
when the less polar low-molecular-
weight portions of the sample are re-
moved prior to the proposed separation
techniques.
The two methods seem to be a good
starting point for the fractionation and
analysis of the more polar constituents.
But more work is needed to handle the
less polar portion of the sample.
Conclusions
Separation by open-column liquid
chromatography is not entirely
satisfactory. Low-molecular-weight
compounds need to be removed before
separation of the woodburn sample into
functional groups.
FAB/MS studies show that the spectra
contain ions at almost every m/z to
1500, but the interpretation of the spec-
Table 3.
Original
Sample
m/z
167(100)*
181(45)
189(45)
368(10)
446(10)
Representative m/z Values for DIP/MS Analysis for Woodburn Sample 233C ^
Fraction 1 Fraction 2 Fraction 3 Fraction 4 Fraction 5 Fraction 7
m/z
167(100)
181(50)
194(50)
197(50)
m/z m/z m/z
163(72) 167(100) 167(100)
167(70) 181(50) 181(50)
181(100) 402(2) 402(5)
197(48)
213(30)
256(25)
m/z
163(80)
167(100)
181(50)
197(40)
239(30)
m/z
161(80)
178(100)
273(10)
329(20)
346(10)
514(5)
"Approximate percent relative intensities given in parentheses.
tra is complicated by the relatively high
abundance of low-molecular-weight
compounds.
DIP/MS analyses indicate the pres-
ence of polar PAHs, but give little infor-
mation above m/z 500.
The proposed HPLC (Method 8) is ca-
pable of giving sufficient class separa-
tion for semipreparative sample frac-
tionation. It may prove to be a useful
analytical technique for the more polar
constituents (pyrrolic, -OH, -COOH,
and -NH2 substituted). But these com-
pounds constitute only about 40% (w/w)
of the woodburn Sample 233C. About
36% of the sample was judged to con-
tain less polar compounds (carbonyl,
nitro, and aza compounds).
/?. James. R. Adams, and H. Miller are with Southern Research Institute,
Birmingham, AL 35255-5305,
Merrill D. Jackson is the EPA Project Officer (see below).
The complete report, entitled "Sampling and Analysis for High-Molecular-
Weight Polar Organic Compounds," (Order No. PB 87-119 434/AS; Cost:
$13.95, subject to change} will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650 ^v ftftv
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
United States
Environmental Protection
Agency "
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S8-86/037
0000329 PS
U S ENVI8 PROTECTION ftGENCT
y^iM^ISSS-lTJEET
CHICAGO 1*- 60604
------- |