19F nuclear magnetic resonance spectroscopy for the quantitative detection and classification of carbonyl groups in lignins

Article abstract of DOI:10.1021/jf980431p

A novel method that permits the quantitative detection and classification of various carbonyl groups in lignins has been developed. The proposed method was optimized with the quantitative trifluoromethylation of a series of carbonyl-containing lignin-like model compounds. This effort was followed by ¹⁹F NMR spectral analyses of the resulting fluorine derivatives allowing for a thorough understanding of their structure/¹⁹F chemical shift relationships. The various carbonyl groups present in lignins were also investigated by trifluoromethylating them in the presence of catalytic amounts of tetramethylammonium fluoride (TMAF), followed by hydrolysis with TMAF in tetrahydrofuran. By using a variety of selective reactions, it became possible to assign a number of prominent ¹⁹F NMR signals to a variety of carbonyl groups present in lignins. These studies demonstrated that the proposed method can be applied to the quantitative determination of carbonyl groups that are present in soluble native and technical lignins.

Full text of DOI:10.1021/jf980431p

190
J. Agric. Food Chem. 1999, 47, 190−201
19F Nu clea r Ma gn etic Reson a n ce Sp ectr oscop y for th e Qu a n tita tive
Detection a n d Cla ssifica tion of Ca r bon yl Gr ou p s in Lign in s
Behzad C. Ahvazi, Claudia Crestini,† and Dimitris S. Argyropoulos*
Department of Chemistry and Pulp and Paper Research Centre, McGill University,
Montreal, Quebec H3A 2A7, Canada
A novel method that permits the quantitative detection and classification of various carbonyl groups
in lignins has been developed. The proposed method was optimized with the quantitative
trifluoromethylation of a series of carbonyl-containing lignin-like model compounds. This effort
was followed by 19F NMR spectral analyses of the resulting fluorine derivatives allowing for a
thorough understanding of their structure/19F chemical shift relationships. The various carbonyl
groups present in lignins were also investigated by trifluoromethylating them in the presence of
catalytic amounts of tetramethylammonium fluoride (TMAF), followed by hydrolysis with TMAF
in tetrahydrofuran. By using a variety of selective reactions, it became possible to assign a number
of prominent 19F NMR signals to a variety of carbonyl groups present in lignins. These studies
demonstrated that the proposed method can be applied to the quantitative determination of carbonyl
groups that are present in soluble native and technical lignins.
Keyw or d s: Nuclear magnetic resonance (NMR); spectroscopy; carbonyl groups; lignins; quantitative
analysis; classification; methods
INTRODUCTION
equivocally shown to be responsible for the yellow color
of photochemically reverted papers (Argyropoulos et al.,
1994; Argyropoulos and Heitner, 1994; Lin and Kring-
stad, 1971; Forsskåhl et al., 1991; Castellan et al., 1993;
Gellerstedt and Pattersson, 1977).
Lignin is a complex phenylpropanoid biopolymer
formed by an enzyme-initiated radical polymerization
of cinnamyl alcohols (Harkin, 1956). Due to the random
nature of its formation, lignin does not possess regular-
ity in its repeating units (J ansherkar and Fiechter,
1983; Glasser and Kelley, 1987; Argyropoulos and
Menachem, 1997). This peculiarity makes the charac-
terization of its structure a challenging task.
Several methods for determining the carbonyl groups
in lignins are available (Green, 1963; Lindberg and
Misiorny, 1952; Lindberg and Theander, 1954; Heuser,
1953; Miyake, 1970). Among these, the most effective
and simple one utilizes the reaction of carbonyl groups
with hydroxylamine hydrochloride, forming an oxime
and hydrochloric acid. Subsequent titration of the
hydrochloric acid provides an estimation of the amount
of carbonyl groups in a sample (Gierer and So¨derberg,
1959; Miyake, 1970). A modification of this technique,
claiming greater reproducibility, has been described by
Zakis (1994). The modified procedure calls for the use
of triethanolamine to function as the acid acceptor
followed by a back-titration. A technique that attempts
the distinction of R-carbonyl groups from those of
conjugated aldehydes is also available (Lindberg and
Misiorny, 1952; Lindberg and Theander, 1954) and is
based on sample reduction (sodium borohydride) fol-
lowed by UV spectroscopic measurements. The latter
method requires the use of appropriate lignin model
compounds that serve as standards for determining the
changes in molar absorptivity of the absorption bands
that are caused by the reduction of a particular carbonyl
group to the corresponding benzylic alcohol.
A number of studies have demonstrated the presence
of small amounts of carbonyl groups in lignins (Adler
and Ellmer, 1948; Adler and Marton, 1959; Marton and
Adler, 1961; Gierer and So¨derberg, 1959). In particular,
milled wood lignins have been reported to contain
conjugated cinnamaldehyde structures and R-carbonyl
groups (Geiger and Fuggerer, 1979). Other investiga-
tions have shown that technical lignins contain ap-
preciable amounts of R-carbonyl groups in addition to
benzaldehyde and quinones (Lin and Dence, 1992;
Sarkanen and Ludwig, 1971). The presence of carbonyl
groups in lignins, in particular those present as o- and
p-quinonoids, quinonemethides, and other extended
conjugated enone systems, is responsible for not only
the color of lignified plant tissue (Hon and Glasser,
1979; Lebo et al., 1990) but also sentisizing centers in
the photoyellowing of lignocellulosic materials (Brunow
and Eriksson, 1971). In general, the low content of
these groups in lignins has made the elucidation of their
role rather elusive. For example, quinones, which are
present in wood and high-yield mechanical pulps (Ar-
gyropoulos et al., 1994; Argyropoulos and Heitner, 1994)
in rather low amounts, only recently have been un-
Infrared spectroscopy has been also used for investi-
gating various structures in lignin (Kolboe and Ellefsen,
1962; Faix, 1991; Hergert, 1971) including carbonyls
(Marton et al., 1961). Recently, Hortling et al. (1997)
reported a semiquantitative technique for the determi-
nation of carboxylic and nonconjugated carbonyl groups
by IR spectroscopy. However, the application of these
techniques was not widespread because their precision
* Author to whom correspondence should be addressed [fax
(514) 398-8254; e-mail dimitria@shared1.lan.ca].
† Permanent address: Department of Science and Chemical
Technology, University of Tor Vergata, Rome, Italy.
10.1021/jf980431p CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/12/1998

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 191
600 µL of TMS-CF3 was added. The mixture was cooled at 0
°C for 10 min followed by the addition of a catalytic amount
(15 mg) of TMAF acting as the initiator. The reaction mixture
continued to be stirred at 0 °C for 30 min and then at room
temperature for 24 h. The intermediate trifluoromethylated
siloxy adducts were then hydrolyzed by the addition of 50 mg
of TMAF at room temperature for 24 h in THF. After the THF
was evaporated under reduced pressure, the residue was
washed and centrifuged thoroughly by 3 × 50 mL water.
Finally, the isolated materials were dissolved in a mixture of
dioxane/water (25:5, v/v) and freeze-dried under reduced
pressure.
Sodium Borohydride Reduction. Lignin (200 mg) was
dissolved into 25 mL of a solution composed of (60:40:50, v/v)
2-methoxyethanol, 2-propanol, and water, respectively. This
was followed by the addition of 3 mL of a solution composed
of 0.01 N sodium hydroxide and 100 mg of sodium borohydride
and stirred at 40 °C for 24 h. The reaction mixture was then
acidified to pH 3-4 with dilute (10%) sulfuric acid. The
organic solvents were evaporated under reduced pressure, and
the lignin was precipitated with the addition of water. The
precipitated lignin was then washed and centrifuged three
times, dissolved in a mixture of dioxane/water (25:5, v/v), and
freeze-dried.
Dakin Reaction. Lignin (200 mg) was suspended in a
solution of 6.5 mL of n-propanol and 7.5 mL of water under
constant stirring for 15 min. To this mixture was added 5
drops of a 0.5 M sodium hydroxide solution, causing the
complete dissolution of the lignin. This was followed by the
addition of 702 µL of 30% hydrogen peroxide. After the pH of
the mixture was adjusted to 10.6 by 0.5 M NaOH, the reaction
mixture was stirred at 50 °C for 4 h. The reaction was then
neutralized by the addition of 1-2 drops of 25% sulfuric acid
to pH 4.7. After the organic solvent was evaporated under
reduced pressure, the lignin was precipitated by the addition
of water. The precipitated lignin was washed with water and
centrifuged three times with 15 mL of water. The isolated
lignin was then dissolved in a mixture of dioxane/water (25:5,
v/v) and freeze-dried.
Sodium Hydrosulfite Reduction of Lignin Model Com-
pounds. Selected lignin model compounds (200 mg) were
dissolved in 5 mL of dioxane, and a slow stream of nitrogen
was bubbled through the solution for ∼30 min. To this
solution was added 200 mg of sodium hydrosulfite (Na2S2O4)
dissolved in 5 mL of water. After various reaction times (15,
60, and 240 min), a 1 mL aliquot of the mixture was
withdrawn, acidified by the addition of 1 M HCl, and extracted
with ethyl acetate. The organic solvent was then evaporated
under reduced pressure, and the residue was analyzed by GC/
MS.
is limited and the various classes of carbonyl moieties
cannot be differentiated.
NMR is rapidly becoming a powerful analytical tool
in the hands of wood chemists aimed at providing
answers in relation to the structure of lignins. However,
the complex structure of these materials has imposed
some serious challenges and limitations, even in the
application and use of NMR (Argyropoulos, 1995).
Efforts to overcome some of the limitations imposed by
proton (Lundquist, 1979a,b, 1980) and 13C NMR spec-
troscopies (Lapierre et al., 1984; Robert and Brunow,
1984; Obst and Landucci, 1986) have prompted the
examination of other NMR-active nuclei. These efforts
have provided additional tools for obtaining fine struc-
tural details for these heterogeneous biopolymers. The
determination of a variety of labile protons in lignins
has been carried out by 29Si-, 31P- and 19F-based NMR
methods, for suitably silylated (Pan and Lachenal, 1994;
Brezny and Schraml, 1987), phosphitylated (Nieminen
et al., 1989; Archipov et al., 1991a,b; Argyropoulos et
al., 1992, 1993; Argyropoulos, 1994a,b; Fillppov et al.,
1991) and fluorinated (Manatt, 1966; Barrelle, 1995)
lignins, respectively. Furthermore, Lebo et al. (1990)
and, recently, Argyropoulos et al. (1995) have reported
the detection of o-quinones in mechanical pulps. In
particular, the latter group has actually managed to
follow their formation during the process of light-
induced yellowing using solid-state 31P NMR spectros-
copy (Argyropoulos et al., 1995).
The 19F nucleus is 100% naturally abundant, and its
high gyromagnetic ratio makes its NMR sensitivity
nearly the same as that of a proton. Its chemical shift
extends over a wide range providing adequate signal
dispersion that may reduce signal overlap and aid
interpretation. Attempts at determining the carbonyl
content of lignin by 19F NMR have actually been made
previously after p-fluorobenzylation (Barrelle, 1993) of
the lignin or its derivatization with p-fluorophenylhy-
drazine (Lachenal et al., 1995). In both cases the 19F
NMR signals overlapped over a relatively narrow range,
thus diminishing the quantitative reliability of the
techniques. Furthermore, the proposed methods were
unable to distinguish among the different classes of
carbonyl groups present in lignin.
In previous work, we developed (Ahvazi and Argy-
ropoulos, 1996a) a selective and quantitative trifluo-
romethylation reaction for tagging the carbonyl groups
in lignin model compounds, using trifluoromethyltri-
methylsilane (TMS-CF3) in the presence of tetramethyl-
ammonium fluoride (TMAF). A series of ketones, al-
dehydes, quinones, and dimeric lignin model compounds
were quantitatively trifluoromethylated, and the result-
ing fluorine derivatives were analyzed by 19F NMR. This
effort allowed for a thorough understanding of the
structure/19F chemical shift relationships (Ahvazi and
Argyropoulos, 1996b) of lignin-like moieties.
Sodium Hydrosulfite Reduction of Lignin. Lignin (200 mg)
was dissolved in 5 mL of dioxane, and a slow stream of
nitrogen was bubbled through the solution for ∼30 min. To
this mixture was added a solution composed of 200 mg of
sodium hydrosulfite in 5 mL of water, and the reaction mixture
was kept under stirring at room temperature for 1 h. The
mixture was then freeze-dried, and the residue was washed
and centrifuged three times with small aliquots of water.
Finally, the reduced lignin was dissolved in a mixture of
dioxane/water (25:5, v/v) and freeze-dried.
In str u m en ta tion . Gas Chromatography/ Mass Spectrom-
etry. GC/MS analyses were carried out on a Hewlett-Packard
5972 mass spectrometer interfaced to a Hewlett-Packard
5890A gas chromatograph with a 30 m × 0.25 mm packed
silica capillary column DB-5. The injection port temperature
was 280 °C, and the oven temperature increase profile was
from 100 to 250 °C, with a gradient of 5 °C/min.
The present effort attempts to expand the application
of quantitative trifluoromethylation to lignins, aimed at
elucidating the nature and the quantity of the various
carbonyl groups present in them.
19F NMR Spectroscopy. All spectra were recorded on a
Varian Unity 500 FT-NMR spectrometer at an operational
frequency of 470.3 MHz. The derivatized trifluoromethylated
lignin was dissolved in 800 µL of a solvent mixture composed
of pyridine and deuterated chloroform (15-20 mg/0.8 mL) at
a volume ratio of 1.6:1, v/v. The mixture was stirred with a
magnetic bar until the lignin was fully dissolved. To this
mixture was added 100 µL of an internal standard solution
EXPERIMENTAL PROCEDURES
Rea ction s. Trifluoromethylation. The following trifluo-
romethylation procedure was developed and applied to all
lignin samples. One hundred milligrams of lignin was dis-
solved, under constant stirring, in 10 mL of dry tetrahydro-
furan (THF) at room temperature. After 10 min of stirring,

192 J. Agric. Food Chem., Vol. 47, No. 1, 1999
Ahvazi et al.
(i.e., 1.69 g/mL 1.6:1, v/v, pyridine/CDCl3 of 3,3′-bis(trifluo-
romethyl)benzophenone; the chemical shift was referenced to
fluorotrichloromethane. Quantitative acquisitions were car-
ried out in 5 mm tubes at room temperature with acquisition
times of 0.64 s followed by a relaxation delay of 10 s. The
number of scans acquired was 1000 per measurement. Pulse
widths corresponding to a 45° flip angle and a line broadening
of 2 Hz were used during acquisition and processing of the
spectra.
) 6.1 Hz) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -77.901
(d, J F-H ) 6.1 Hz) ppm. MS m/z (silylated) 336 (M+), 267, 249,
225, 197, 195, 151. Anal. Calcd for C8H7F3O2: C, 50.01; H,
3.67; F, 29.66. Found: C, 50.12; H, 3.71; F, 29.72. 98% yield.
10: 1H NMR (acetone-d6) (TMS) δ 2.997 (s, b, 1H); 4.997
(m, 1H); 5.560 (m, 1H); 6.834 (m, 2H); 7.976 (s, b, 1H); 8.035
(S, b, 1H) ppm. 19F NMR (C2D6O) (CFCl3) δ -77.587 (d, J F-H
) 7.5 Hz) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -77.691
(d, J F-H ) 8.0 Hz) ppm. MS m/z (silylated) 424 (M+), 409, 383,
356, 352, 283, 247. Anal. Calcd for C8H7F3O3: C, 46.17; H,
3.39; F, 27.38. Found: C, 46.25; H, 3.41; F, 27.37. 97% yield.
11: 1H NMR (CDCl3) (TMS) δ 2.59 (s, 1H); 3.89 (s, 3H);
4.925 (q, 1H, J ) 6.78 Hz); 5.72 (s, 1H); 6.947 (d, 3H, J ) 13.93
Hz) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -77.756 (d,
J F-H ) 8.0 Hz) ppm. MS m/z 222 (M+), 205, 183, 153, 125, 93,
65, 51. Anal. Calcd for C9H9F3O3: C, 48.66; H, 4.08; F, 25.65.
Found: C, 48.70; H, 4.11; F, 25.69. 95% yield.
Two-Dimensional NMR Spectroscopy (HMQC). Such spec-
tra were acquired using 10-30 mg of sample dissolved in 0.6
mL of CDCl3, DMSO-d6, or acetone-d6 on a Varian Unity 500
NMR spectrometer using a 5 mm inverse detection probe
(DHP). The chemical shifts were referenced to Me4Si and
CFCl3 for 13C and 19F, respectively. HMQC spectra were
acquired over a 30 ppm window in F2 (19F) and 190 ppm in F1
(13C) with GARP-1 decoupling. The value of the coupling
2
constant was adjusted to the estimated value for the J C-F
12: 1H NMR (CDCl3) (TMS) δ 2.18 (s, 1H); 3.85 (s, 6H);
4.924 (q, 1H, J ) 6.58 Hz); 6.68 (s, 2H) ppm. 19F NMR (CDCl3/
pyridine) (CFCl3) δ -77.633 (d, J F-H ) 8 Hz) ppm. MS m/z
252 (M+), 205, 183, 167, 155, 140, 123, 95. Anal. Calcd for
of 30 Hz; 2K × 256 increments were acquired. After F1 zero-
filling, Fourier transformation, and squared cosine bell apodiza-
tion, the transformed data matrix was 1024 (F2) × 512 (F1)
real points.
C
10H11F3O4: C, 47.63; H, 4.40; F, 22.60. Found: C, 47.62; H,
Ch a r a ct er iza t ion . Characterization of Trifluoromethyl-
ated Model Compounds. 1:1H NMR (CDCl3) (TMS) δ 1.77 (s,
3H); 2.60 (s, 1H); 7.37-7.46 (m, 3H); 7.56-7.61 (m, 2H) ppm.
19F NMR (CDCl3/pyridine) (CFCl3) δ -80.30 (s) ppm. MS m/z
190 (M+), 151, 127, 121, 105, 91, 77, 51. Anal. Calcd for
C9H9F3O: C, 56.84; H, 4.77; F, 29.97. Found: C, 56.92; H,
4.69; F, 29.92. 96% yield.
4.42; F, 22.58. 98% yield.
13: 1H NMR (CDCl3) (TMS) δ 2.50 (s, 1H); 3.87 (s, 6H);
4.943 (q, 1H, J ) 6.64 Hz); 6.841 (t, 1H, J ) 4.88 Hz); 6.976
(d, 2H, J ) 3.42 Hz) ppm. 13C NMR (CDCl3) (TMS) δ 55.73;
72.65 (q, J C-CF ) 32.1 Hz); 110.19; 110.91; 120.29; 124.30 (q,
J C-CF ) 281.7 Hz); 126.44, 149.13; 149.99. 13C NMR (DMSO)
δ 24.05; 70.37 (q, J C-F ) 30.2 Hz); 114.89; 115.06; 118.85;
124.20 (q, J C-CF ) 282.3 Hz); 126.59; 144.95; 145.79. 19F NMR
(CDCl3/pyridine) (CFCl3) δ -77.756 (d, J F-H ) 7.5 Hz) ppm.
MS m/z 236 (M+), 219, 197, 167, 139, 124, 108, 96. Anal. Calcd
for C10H11F3O3: C, 50.85; H, 4.69; F, 24.13. Found: C, 50.92;
H, 4.66; F, 24.11. 99% yield.
2: 1H NMR (CDCl3) (TMS) δ 1.74 (s, 3H); 2.35 (s, 1H); 5.67
(s, 1H); 6.829 (d, 2H, J ) 8.78 Hz); 7.432 (d, 2H, J ) 8.58 Hz)
ppm. 13C NMR (CDCl3) (TMS) δ 23.85; 74.54 (q, J C-CF ) 29.1
Hz); 115.11; 155.83; 129.36 (q, J C-F ) 284.9 Hz); 155.83. 19F
NMR (CDCl3/pyridine) (CFCl3) δ -80.57 (s) ppm. MS m/z 206
(M+), 188, 167, 149, 137, 119, 107, 91. Anal. Calcd for
C9H9F3O2: C, 52.43; H, 4.40; F, 27.65. Found: C, 52.48; H,
4.49; F, 27.67. 98% yield.
14: 1H NMR (CDCl3) (TMS) δ 2.97 (s, 1H); 4.58-4.66 (m,
1H); 6.20 (q, 1H, J ) 9.46 Hz); 6.842 (d, 1H, J ) 15.93 Hz);
7.31-7.44 (m, 5H) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ
-78.229 (d, J F-H ) 7.5 Hz) ppm. MS m/z 202 (M+), 184, 165,
133, 115, 91, 77, 55. Anal. Calcd for C10H9F3O: C, 59.41; H,
4.49; F, 28.19. Found: C, 59.45; H, 4.52; F, 28.17. 99% yield.
15: 1H NMR (CDCl3) (TMS) δ 2.25 (s, 1H); 3.84 (s, 3H);
4.531 (t, 1H, J ) 13.19 Hz); 5.65 (s, 1H); 5.961 (q, 1H, J )
9.12 Hz); 6.689 (d, 1H, J ) 15.73 Hz); 6.82-6.88 (m, 3H) ppm.
19F NMR (CDCl3/pyridine) (CFCl3) δ -78.241 (d, J F-H ) 6.1
Hz) ppm. MS m/z 248 (M+), 219, 199, 179, 161, 147, 119, 91.
Anal. Calcd for C11H11F3O3: C, 53.23; H, 4.47; F, 22.96.
Found: C, 53.28; H, 4.44; F, 22.97. 99% yield.
16: 1H NMR (CDCl3) δ 2.42 (br, 4H); 6.19 (s, 4H,); 6.27 (s,
4H,) ppm. 1H NMR (DMSO) δ 6.17 (s, 4H); 6.86 (s, 2H,) ppm.
13C NMR (DMSO) δ 67.18 (q, J C-CF ) 29.4 Hz); 124.77 (q, J C-F
) 286.4 Hz); 129.16. 13C NMR (acetone-d6) δ 68.70 (q, J C-CF
) 30.3 Hz); 125.64 (q, J C-F ) 285.0 Hz); 130.26. 19F NMR
(CDCl3/pyridine) (CFCl3) δ -79.80 (s, major); -79.72 (s, minor)
ppm. MS m/z 179 (M+ - CF3), 159, 143, 110, 83, 69, 53. Anal.
Calcd for C8H6F6O2: C, 38.73; H, 2.44; F, 45.94. Found: C,
38.75; H, 2.49; F, 45.96. 96% yield.
17: 1H NMR (CDCl3) δ 2.02-2.03 (d, 3H, J ) 1.16 Hz); 2.30
(br, 2H); 5.98-5.99 (d, 1H, J ) 1.72 Hz); 6.24-6.25 (t, 2H, J )
2.16 Hz) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -80.25
(s); -76.04 (s) ppm. MS m/z 193 (M+ - CF3), 173, 145, 124,
69, 51. Anal. Calcd for C9H8F6O2: C, 41.24; H, 3.08; F, 43.48.
Found: C, 41.27; H, 3.15; F, 43.50. 95% yield.
18: 1H NMR (CDCl3) δ 1.37-1.41 (d, 9H, J ) 8.60 Hz);
3.19-3.30 (d, 2H, J ) 21.23 Hz); 7.52-7.60 (m, 3H); 7.89-
8.03 (m, 4H) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -77.85
(s); -77.69 (s) ppm. MS m/z 335 (M+ - CF3), 320, 266, 251,
209, 181, 152, 112. Anal. Calcd for C20H18F6O2: C, 59.41; H,
4.49; F, 28.19. Found: C, 59.38; H, 4.46; F, 28.22. 94% yield.
19: 1H NMR (CDCl3) δ 1.05 (s, 9H); 1.25 (s, 9H); 2.45 (s,
1H); 3.05 (s, 1H); 5.19 (s, 1H); 6.06 (s, 1H) ppm. 19F NMR
(CDCl3/pyridine) (CFCl3) δ -68.78 (s); -74.86 (s) ppm. MS
m/z (silylated) 504 (M+ - CF3), 489, 399, 379, 327, 285, 259,
239. Anal. Calcd for C16H22F6O2: C, 53.33; H, 6.15; F, 31.63.
Found: C, 53.40; H, 6.10; F, 31.59. 89% yield.
3: 1H NMR (CDCl3) (TMS) δ 1.74 (s, 3H); 2.90 (s, 1H); 3.879
(d, 3H, J ) 6.64 Hz); 5.84 (s, 1H); 6.86-7.13 (m, 3H) ppm. 19
F
NMR (CDCl3/pyridine) (CFCl3) δ -80.38 (s) ppm. MS m/z 236
(M+), 197, 167, 151, 124, 110, 69, 51. Anal. Calcd for
C
10H11F3O3: C, 50.85; H, 4.69; F, 24.13. Found: C, 50.91; H,
4.68; F, 24.19. 96% yield.
4: 1H NMR (CDCl3) (TMS) δ 1.73 (s, 3H); 2.54 (s, 1H); 3.879
(d, 6H, J ) 10.26 Hz); 6.78 (s, 2H); 7.22 (s, 1H) ppm. 19F NMR
(CDCl3/pyridine) (CFCl3) δ -80.28 (s) ppm. MS m/z 266 (M+),
227, 197, 181, 155, 123, 93, 69. Anal. Calcd for C11H13F3O4:
C, 49.63; H, 4.92; F, 21.41. Found: C, 49.69; H, 4.85; F, 21.35.
95% yield.
5: 1H NMR (CDCl3) (TMS) δ 1.73 (s, 3H); 2.72 (s, 1H); 3.841
(d, 6H, J ) 2.16 Hz); 6.825 (d, 1H, J ) 8.48 Hz); 7.03-7.11
(m, 2H) ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ -80.42 (s)
ppm. MS m/z 250 (M+), 211, 181, 139, 124, 107, 95, 77. Anal.
Calcd for C11H13F3O3: C, 52.80; H, 5.24; F, 22.78. Found: C,
52.85; H, 5.28; F, 22.80. 95% yield.
6: 1H NMR (CDCl3) (TMS) δ 2.93 (s, 1H); 7.34-7.38 (m,
6H); 7.47-7.52 (m, 4H) ppm. 19F NMR (CDCl3/pyridine)
(CFCl3) δ -73.39 (s) ppm. MS m/z 252 (M+), 233, 213, 183,
165, 127, 105, 77. Anal. Calcd for C11H13F3O3: C, 66.67; H,
4.40; F, 22.60. Found: C, 66.71; H, 4.32; F, 22.66. 94% yield.
7: 1H NMR (CDCl3) (TMS) δ 3.22 (s, 1H); 3.84 (s, 6H); 6.87-
7.03 (m, 4H); 7.03-7.47 (m, 4H) ppm. 19F NMR (CDCl3/
pyridine) (CFCl3) δ -73.76 (s) ppm. MS m/z 312 (M+), 273,
243, 212, 168, 135, 108, 77. Anal. Calcd for C16H15F3O3: C,
61.54; H, 4.84; F, 18.25. Found: C, 61.56; H, 4.88; F, 18.17.
96% yield.
8: 1H NMR (CDCl3) (TMS) δ 2.60 (s, 1H); 4.989 (q, 1H, J )
6.60 Hz); 7.38-7.48 (m, 5H) ppm. 19F NMR (CDCl3/pyridine)
(CFCl3) δ -78.680 (d, J F-H ) 6.1 Hz) ppm. MS m/z 176 (M+),
159, 127, 107, 89, 79, 51. Anal. Calcd for C8H7F3O: C, 54.55;
H, 4.01; F, 32.36. Found: C, 54.58; H, 3.99; F, 32.40. 97%
yield.
9: 1H NMR (acetone-d6) (TMS) δ 5.073 (q, 1H, J ) 7.34 Hz);
5.30-6.30 (s, b, 1H); 6.658 (m, 2H); 7.359 (m, 2H); 7.80-8.90
(s, b, 1H) ppm. 19F NMR (C2D6O) (CFCl3) δ -77.745 (d, J F-H

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 193
F igu r e 1. Trifluoromethylation of carbonyl-containing (including quinones) lignin-like model compounds.
20: 1H NMR (CDCl3) δ 4.75 (s, 2H); 7.25-7.34 (m, 2H);
7.41-7.55 (m, 6H); 7.67-7.72 (t, 2H, J ) 7.32 Hz) ppm. 19F
NMR (CDCl3/pyridine) (CFCl3) δ -75.66 (s) ppm. MS m/z 175
(M+ - C8H6F3O), 152, 105, 77, 51. Anal. Calcd for C16H12F6O2:
C, 54.87; H, 3.45; F, 32.54. Found: C, 54.91; H, 3.50; F, 32.58.
98% yield.
-77.65 (d, J F-H ) 7.1 Hz); -77.04 (b); -77.75 (s); -77.76 (s);
-78.13 to -78.23 (b); -78.63 (s); -84.32 (s) ppm.
Steam explosion lignin: 19F NMR (CDCl3/pyridine) (CFCl3)
δ -72.84 (s); -73.65 (s); -74.27 (s); -75.81 (s); -76.37 (b);
-77.16 (d, J F-H ) 4.2 Hz); -77.29 (d, J F-H ) 7.5 Hz); -77.32
(b); -77.54 (b); -77.64 (d, J F-H ) 8.0 Hz); -77.75 (d, J F-H
)
21: 1H NMR (CDCl3) (TMS) δ 3.63 (s, 1H); 3.78 (s, 3H);
3.863 (d, 6H, J ) 5.40 Hz), 4.07-4.12 (m, 2H); 4.449(t, 1H, J
) 1.80 Hz); 5.57 (s, 1H); 6.692 (q, 1H, J ) 3.60 Hz); 6.77-6.83
(m, 2H); 6.865 (d, 1H, J ) 5.10 Hz); 6.95-6.99 (m, 1H); 7.193
(d, 1H, J ) 4.80 Hz); 7.294 (d, 1H, J ) 1.2 Hz) ppm. 13C NMR
7.1 Hz); -78.18 (b); -78.28 (d, J F-H ) 6.6 Hz); -78.63 (s);
-78.76 (b); -79.13 (b); -79.22 (d, J F-H ) 3.8 Hz); -84.32 (s)
ppm.
Straw lignin: 19F NMR (CDCl3/pyridine) (CFCl3) δ -67.67
(b); -74.26 (s); -75.15 (b); -75.80 (s); -76.52 (b); -76.64 (s);
-77.04 (b); -77.25 (s); -77.69 (b); -78.19 (s); -79.05 (s);
-79.15 (s); -83.27 (b); -83.95 (b); -84.39 (s) ppm.
Softwood Milled wood lignin: 19F NMR (CDCl3/pyridine)
(CFCl3) δ -74.23 (s); -75.55 (b); -76.41 (b); -76.64 (b); -78.18
(b); -79.13 (d, J F-H ) 6.6 Hz); -79.21 (d, J F-H ) 6.6 Hz);
-79.80 (b); -84.20 (b) ppm.
(CDCl3/) (TMS) δ 55.85; 56.04; 56.17; 61.36; 79.59 (q, J C-CF
)
27.4 Hz); 82.64; 109.87; 110.48; 112.07; 118.25; 120.72; 121.66;
124.62; 124.97 (q, J C-F ) 286.59 Hz); 128.93; 146.18; 148.56;
148.89; 151.53 ppm. 19F NMR (CDCl3/pyridine) (CFCl3) δ
-76.03 (s) ppm. MS m/z (silylated) 402 (M+), 302, 278, 248,
235, 221, 181, 165. Anal. Calcd for C19H21F3O6: C, 56.72; H,
5.26; F, 14.16. Found: C, 56.74; H, 5.29; F, 14.22. 95% yield.
22: 1H NMR (CDCl3) (TMS) δ 3.84 (s, 3H); 3.87 (s, 3H); 3.90
(s, 3H), 3.93 (d, 1H, J ) 3 Hz); 4.24 (d, 1H, J ) 12 Hz); 4.640
(d, 1H, J ) 12 Hz); 5.865 (d, 1H, J ) 3 Hz); 6.84-7.12 (m, 6H)
ppm. 13C NMR (CDCl3) (TMS) δ 56.11; 56.19; 56.64; 71.68;
75.62 (q, J C-F ) 28.65 Hz); 108.78; 110.06; 111.44; 113.40;
120.41; 121.80; 122.95; 125.60 (q, J C-F ) 286.59 Hz); 127.88;
147.37; 148.68; 149.16; 149.95 ppm. 19F NMR (CDCl3/pyridine)
(CFCl3) δ -76.57 (s) ppm. MS m/z 372 (M+), 303, 248, 235,
217, 189, 180. Anal. Calcd for C18H19F3O5: C, 58.06; H, 5.14;
F, 15.31. Found: C, 58.21; H, 5.20; F, 15.29. 96% yield.
23: 1H NMR (CDCl3) (TMS) δ 2.25 (s, 3H); 3.75 (s, 1H); 3.79
(s, 3H), 3.88 (s, 3H); 3.935 (d, 1H, J ) 9 Hz); 4.12 (d, 1H, J )
5.99 Hz); 4.33 (s, 1H); 5.60 (s, 1H,); 5.70 (s, 1H); 6.55-6.64
(m, 3H); 6.94 (d, 1H, J ) 5.40 Hz); 7.10 (d, 1H, J ) 4.80 Hz);
7.311 (d, 1H, J ) 1 Hz) ppm. 13C NMR (CDCl3) (TMS) δ 21.19;
55.88; 55.93; 61.31; 79.85 (q, J C-CF ) 28.9 Hz); 82.64; 109.59;
112.95; 113.93; 118.43; 121.01; 122.19; 125.07 (q, J C-F ) 286.59
Hz); 128.71; 134.83; 143.81; 145.62; 146.35; 151.36 ppm. 19F
NMR (CDCl3/pyridine)(CFCl3) δ -75.93 (s) ppm. MS m/z
(silylated) 618 (M+), 480, 451, 411, 365, 343, 323, 271. Anal.
Calcd for C19H21F3O6: C, 56.72; H, 5.26; F, 14.16. Found: C,
56.78; H, 5.19; F, 14.18. 93% yield.
RESULTS AND DISCUSSION
The detailed chemical reactions used to quantitatively
trifluoromethylate the carbonyl groups (including qui-
nones) that are present in lignin are shown in Figure
1. The precise trifluoromethylation conditions used for
lignins were developed from an understanding of the
reaction details for various model compounds (Ahvazi
and Argyropoulos, 1996b).
The acquisitions of the 19F NMR spectra for all
trifluoromethylated lignins were carried out in a mix-
ture of CDCl3 and pyridine (1:1.6, v/v), due to the
relatively low solubility of lignins in common organic
solvents. The particular choice of CDCl3/pyridine (1:
1.6, v/v) was made on the basis of our previous work on
31P NMR spectra of phosphitylated lignins (Argyropou-
los et al., 1993; Argyropoulos, 1994a,b; Fillppov et al.,
1991). For this reason all 19F chemical shift values for
trifluoromethylated carbonyl-containing lignin model
compounds were recorded in CDCl3/Py (DMF and
DMSO were not suitable) (Tables 1-4).
Characterization of Trifluoromethylated Lignins. Dioxane
lignin: 19F NMR (CDCl3/pyridine) (CFCl3) δ -67.71 to -67.79
(m); -72.70 to -72.99 (b); -74.23 (s); -74.54 (s); -75.20 (b);
-75.83 (s); -76.87 to -77.11 (b); -77.78 (d, J HF ) 6.6 Hz);
-78.21 to -78.26 (b); -78.78 (d, J HF ) 7.0 Hz); -82.66 (s);
-84.25 (d, J HF ) 5.2 Hz) ppm.
The 19F NMR signals of trifluoromethylated ketones
for lignin end-groups range between -80.28 and -80.42
ppm (upfield from CFCl3), whereas those of dimeric
units are confined between -73.39 and -76.57 ppm.
Trifluoromethylated derivatives of cinnamic-like alde-
hydes appeared between -78.23 and -78.24 ppm,
whereas benzaldehyde analogues occupied the range
from -77.63 to -77.90 ppm. The latter signals ap-
peared as doublets due to the coupling of fluorine to the
adjacent proton present on the trifluoromethylated
carbon, with coupling constants ranging between 6.1
and 8.0 Hz (Ahvazi and Argyropoulos, 1996b).
Kraft lignin: 19F NMR (CDCl3/pyridine) (CFCl3) δ -73.78
(s); -74.29 (s); -74.48 (s); -74.92 (s); -76.06 (s); -76.40 (s);
-77.61 (b); -77.78 (d, J F-H ) 6.6 Hz); -79.14 (d, J F-H ) 6.6
Hz); -79.24 (d, J F-H ) 7.1 Hz); -83.42 (s); -84.22 (d, J F-H
4.2 Hz) ppm.
)
Sucrolin acid hydrolysis: 19F NMR (CDCl3/pyridine) (CFCl3)
δ -74.28 (s); -75.77 (s); -75.81 (s); -76.53 (b); -76.64 (s);
-77.65 (d, J F-H ) 6.6 Hz); -77.76 (d, J F-H ) 7.5 Hz); -77.89
(d, J F-H ) 7.0 Hz); -79.13 (d, J F-H ) 6.6 Hz); -84.32 (s);
-85.35 (b) ppm.
The 19F NMR chemical shifts of trifluoromethylated
ortho- and p-quinones were spread over a wide range,
and their position was found to be sensitive to steric
effects. The 19F NMR spectra of a trifluoromethylated
o-quinone model compound showed two signals at -68.8
and -74.9 ppm, whereas the signals of trifluoromethyl-
Alcell organosolv: 19F NMR (CDCl3/pyridine) (CFCl3) δ
-66.19 (s); -66.23 (s); -72.84 (s); -73.20 (b); -73.65 (s);
-74.27 (s); -74.81 (t, J F-H ) 13.2 Hz); -75.81 (s); -76.37 (b);
-77.29 (d, J F-H ) 7.5 Hz); -77.35 (b); -77.46 (b); -77.54 (b);

194 J. Agric. Food Chem., Vol. 47, No. 1, 1999
Ahvazi et al.
Ta ble 1. F lu or id e Ion In d u ced Tr iflu or om eth yla tion of Ca r bon yl Com p ou n d s of Keton es w ith
Tr iflu or om eth yltr im eth ylsila n e
ated p-quinone model compounds ranged from -76.0 to
-80.2 ppm (upfield from CFCl3).
NMR chemical shifts and coupling constants (J F-H). The
19F NMR signals of trifluoromethylalted aldehydes were
spread over two regions: from -77.6 to -77.9 ppm and
from -78.8 to -79.1 ppm. These regions were assigned
to benzoic and cinnamic aldehyde type structures,
respectively. The absence of the characteristic aldehydic
doublets signals from the 19F NMR spectra of some
lignins could be due to signal overlap.
Figures 2-4 show the 19F NMR spectra for a variety
of trifluoromethylated lignin samples, black spruce
milled wood lignin, residual kraft lignin, Sucrolin, Alcell
organosolv, steam explosion lignin from yellow poplar
(Andersons and Faix, 1995; Milne et al., 1992), and
milled straw lignin. These spectra contain a number
of 19F NMR signals that spread over 20 ppm, with a
number of common signals for all of the lignins.
We clarified these signal assignments by examining
13C NMR signal splitting by fluorine nuclei in two-
dimensional 19F-13C heteronuclear NMR experiments.
This is because the 13C NMR spectra for a number of
trifluoromethylated model compounds showed distinct
signals with appreciably different J -coupling constants.
More specifically, the 13C NMR chemical shifts for CF3
groups (quartet) appeared between 123 and 126 ppm,
To ensure that the trifluoromethylation reaction was
selectively carried out on the carbonyl groups in lignin,
19F NMR spectra of lignin samples before and after
reduction with sodium borohydride were acquired. As
anticipated, the 19F NMR spectra of the reduced and
trifluoromethylated lignins showed no signals (Figure
5).
Sign a l Assign m en t. Structural elucidation of sev-
eral trifluoromethylated carbonyl-containing moieties in
lignins was first carried out by comparing their 19F
NMR chemical shifts to the various trifluoromethylated
lignin model compounds. The 19F NMR spectral analy-
ses of different fluorinated lignins displayed numerous
well-resolved sharp signals ranging from -64 to -87
ppm, corresponding exactly to the region of various
trifluoromethylated lignin model compounds, allowing
for some tentative signal assignments. These assign-
ments were tentative because the 19F NMR chemical
shifts of trifluoromethylated quinones occupied a wide
range, overlapping with those of ketones (Ahvazi and
Argyropoulos, 1996b). As such, complete signal iden-
tification could not be carried out solely on the basis of
model compound chemical shift information.
1
with a J C-F coupling constant of ∼285 Hz. Further-
more, a long-range 2J C-F coupling constant was found
to be ∼30 Hz, confined (quartet) between 68 and 80
ppm, allowing for the differentiation of the ketonic from
the quinonic signals (Table 5).
In an effort to select a suitable set of parameters that
would cover all possible 19F-13C long-range coupling
constants that may be encountered in lignin, several
HMQC experiments were conducted on different car-
bonyl-containing lignin model compounds. These stud-
ies revealed that, during an HMQC experiment, minor
variations in the selected J -coupling constants could
have serious implications on cross-peak intensity. For
example, Figure 7 shows HMQC spectra of trifluoro-
methylated 3,4-dimethoxybenzaldehyde (13) with J
values 28, 30, 32, and 282 Hz. The cross-peak at -72.25
and 73.70 ppm (Figure 7, signal I) is our target signal.
However, when the J value was varied, another cross-
peak at -72.37 and 130.9 ppm (Figure 7B-D, signal
The presence of different aldehydes in trifluorometh-
ylated lignins was detected on the basis of their 19F

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 195
Ta ble 2. F lu or id e Ion In d u ced Tr iflu or om eth yla tion of Ca r bon yl Com p ou n d s of Ald eh yd es w ith
Tr iflu or om eth yltr im eth ylsila n e
Ta ble 3. F lu or id e Ion In d u ced Tr iflu or om eth yla tion of Qu in on es w ith Tr iflu or om eth yltr im eth ylsila n e
II) was apparent as a result of an isotope shift effect
due to the 13C-19F interaction. This is not surprising
because isotopic substitution causes changes in shield-
ing effects: for instance, the 19F NMR chemical shift of

196 J. Agric. Food Chem., Vol. 47, No. 1, 1999
Ahvazi et al.
Ta ble 4. F lu or id e Ion In d u ced Tr iflu or om eth yla tion of Ca r bon yl Gr ou p s w ith Tr iflu or om eth yltr im eth ylsila n e
F igu r e 3. Quantitative 19F NMR spectra of Sucrolin (top) and
milled straw (bottom) lignins.
F igu r e 2. Quantitative 19F NMR spectra of steam explosion
yellow poplar (top) and black spruce milled wood (bottom)
lignins.
tion times (24 h), were inconclusive due to the low
carbonyl contents of lignins that gave low signal-to-noise
ratios.
CF3I is shielded by 0.149 ppm more for the 13CF3I
isotopomer than in 12CF3I (Harris, 1983). This signal,
however, was easily distinguished from the primary
correlation because it was confined in the -CF3 13C
chemical shift region, and, in addition, it was slightly
shifted (∼0.1 ppm) from the parent 19F peak. Neverthe-
less, isotope shift effects in HMQC spectra of trifluo-
romethylated lignins could increase the complexity of
signal assignment. Because such potential problems
were known, a number of trifluoromethylated lignins
were subjected to HMQC experiments. The accumu-
lated spectral data, however, despite the long acquisi-
Because 2D NMR was of limited utility in aiding the
19F NMR signal assignments for trifluoromethylated
lignins, our attention was focused to the application of
selective chemical derivatization techniques. Two dif-
ferent reactions were considered, namely, the Dakin
oxidation (Bailey and Dence, 1969; Reeves and Pearl,
1965) and sodium hydrosulfite reduction (Rabjohn,
1963; Fieser and Fieser, 1967; Grundmann, 1977).
The Dakin reaction causes the selective oxidation of
various carbonyl groups present in lignins (Reeves and

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 197
F igu r e 6. 19F-13C coupling constants for different classes of
carbonyl groups.
nonphenolic benzaldehydes are converted directly to the
corresponding benzoic acids.
Therefore, a lignin sample subjected to the Dakin
reaction should be enriched in p-quinones and depleted
of aldehydes and R-carbonyls that bear free phenolic
hydroxyl groups. The total concentration of etherified
R-carbonyl structures, however, should remain the same
before and after the Dakin reaction.
F igu r e 4. Quantitative 19F NMR spectra of residual kraft
Sodium hydrosulfite is a mild reducing agent that has
been reported (Rabjohn, 1963; Rieser and Fieser, 1967;
Grundmann, 1977) to selectively reduce quinones in the
presence of aldehydes or ketones. To select the best
reaction conditions for selective reductions of lignins, a
series of exploratory experiments were carried out.
These experiments were also aimed at confirming that
cinnamyl and benzyl aldehydes as well as model R-ke-
tones would not be reduced by sodium hydrosulfite.
More specifically, di-tert-butyl-o-quinone, p-quinone,
acetovanillone, and syringaldehyde were reduced by
sodium hydrosulfite. Both o- and p-quinones were
reduced quantitatively to their corresponding alcohol in
15 min, whereas acetovanillone and syringaldehyde
were not affected, even after a 4 h reaction. The
reduction of lignin with sodium hydrosulfite was com-
plete within 1 h.
(top) and Alcell organosolv (bottom) lignins.
Figure 8 shows the 19F NMR spectra of trifluoro-
methylated samples of residual dioxane lignin before (A)
and after Dakin oxidation (B) and after sodium hydro-
sulfite reduction (C). On the basis of the above ac-
counts, and the chemical shift data of Tables 1-4, a
number of major carbonyl signals are tentatively as-
signed.
The comparison of 19F NMR spectral analyses of
trifluoromethylated dioxane lignin showed a number of
prominent signals that were significantly affected by the
Dakin and sodium hydrosulfite reactions. For example,
the intensities of signals located at -67.7, -73.0, and
-78.2 ppm in the original spectrum of the dioxane lignin
(Figure 8A) were reduced almost completely (Figure 8C)
after their reaction with sodium hydrosulfite. There-
fore, these signals were assigned to o- and p-quinones
on the basis in chemistry known to occur between
sodium hydrosulfite and quinones.
Another important signal, centered at -74.5 ppm,
which appeared consistently in all of the different
trifluoromethylated lignin samples (Figures 2-4), was
also identified. This signal, which was assumed to
represent R-carbonyl-containing â-O-4 structures or
quinones (Table 4), was found to be drastically reduced
after Dakin oxidation, whereas it remained unaffected
upon treatment with sodium hydrosulfite (Figure 8). As
such, this signal was assigned to be due exclusively to
R-carbonyl groups of â-O-4 units bearing a free phenolic
hydroxyl group para to the side chain.
F igu r e 5. 19F NMR spectra of trifluoromethylated kraft lignin
before (top) and after (bottom) reduction of carbonyls by
sodium borohydride.
Ta ble 5. 13Ca r bon NMR Ch em ica l Sh ifts a n d Cou p lin g
Con sta n ts of Som e Tr iflu or om eth yla ted Mod el
Com p ou n d s
13C NMR δ (ppm) J C-F (Hz)
compound
CF3
C-CF3
1J
2J
dimer â-O-4 (21)
125.1
125.1
125.5
130.8
79.6
79.8
75.6
74.5
72.6
68.7
285 26
285 27
285 47
290 29
282 32
285 30
dimer â-O-4 (22)
dimer â-O-4 (23)
p-hydroxyacetophenone (2)
3,4′-dimethoxybenzaldehyde (13) 124.1
p-benzoquinone (16) 123.0
Pearl, 1965). R-Carbonyl groups are oxidized to p-
quinones when a free hydroxyl group is present para
to the side chain. In contrast, when the phenolic group
is etherified, the system is totally unreactive. Further-
more, R,â-unsaturated aldehydes react with alkaline
hydrogen peroxide with the formation of the corre-
sponding benzaldehydes and benzoic acids, whereas

198 J. Agric. Food Chem., Vol. 47, No. 1, 1999
Ahvazi et al.
19F NMR signals of aldehydes, quinones, and also
R-carbonyls could be all found in this region.
The last set of 19F NMR signals in trifluoromethylated
lignin spectra located between -82 and -85 ppm
(Figure 8) were assigned to different unhindered ke-
tones. Comparison with model compound data (Table
1) allowed the assignments of two different classes of
ketones in this region. The signal that was not affected
by the Dakin oxidation appeared at -84.2 ppm and was
assigned as being due to the R-carbonyl of etherified
lignin end-groups. However, the signal at -82.7 ppm
was found to be seriously reduced by Dakin oxidation.
This signal was assigned to the ketonic structures
bearing a free phenolic hydroxyl group in the para
position of the aromatic ring such as 5-5′-biphenyl or
4-O-5′ units. Traces of acetone used to wash and dry
the glassware were found to give rise to the signal at
-82.15 ppm. Therefore, the sensitivity of this technique
dictates that when acetone is used for cleaning purposes,
it should be thoroughly removed.
Qu a n tita tive Eva lu a tion of th e Ca r bon yl Gr ou p s
in Lign in s. The quantification of the total amount of
carbonyl groups in lignin was carried out by using 3,3′-
bis(trifluoromethyl)benzophenone as an internal stan-
dard. This compound had all of the characteristics of a
reliable internal standard required for accurate mea-
surements: it is a pure crystalline solid possessing two
equivalent CF3 groups giving a sharp signal at -62.511
ppm. Its position is in the proximity of the lignin
signals, so as to allow the use of a narrow sweep width
during spectral acquisition, and at the same time does
not overlap with any of the lignin signals, allowing for
precise integrations. The use of this internal standard
permitted the quantitative determination of all carbonyl
groups present in all examined lignins. This was made
possible because adequate delay time between pulses
was used (10 s). This selection was based on detailed
measurements of the 19F spin-lattice relaxation times
for trifluoromethylated lignins and the internal stan-
dard. As anticipated, the longest T1 was that of the
internal standard.
To examine the reproducibility and quantitative reli-
ability of our measurements, several native and techni-
cal lignins were selected and their carbonyl contents
were determined after trifluoromethylation. The total
carbonyl content for each lignin sample was determined
four times, and the calculated mean values and stan-
dard deviations are shown in Table 6. Notably, the total
amount of carbonyls determined by 19F NMR was found
to be different from sample to sample with high preci-
sion. A further investigation aimed at substantiating
the present technique as an analytical tool for the
quantification of the various carbonyl groups in different
soluble lignins was conducted. In particular, the quan-
titative derivatization of carbonyls by trifluoromethyl-
ation was examined by selecting lignin samples for
which the carbonyl contens were determined according
to two different techniques, oximation and UV spec-
troscopy, during the 1991 International Round Robin
effort (Andersons and Faix, 1995; Milne et al., 1992). It
was thus possible to compare the results furnished by
quantitative 19F NMR with those produced by indepen-
dent methods in other laboratories for the same samples
as presented in Table 7.
F igu r e 7. 19F-13C HMQC spectra of trifluoromethylated 3,4-
dimethoxybenzaldehyde acquired by selecting different
values.
J
The fine structural elucidation for a number of signals
located at -75 to -79 ppm (Figure 8) was restricted
because various trifluoromethylated carbonyl signals in
lignin partially overlap in this region. For instance, the
The proximity of the two sets of data in Table 7
qualifies the 19F NMR technique as a novel and promis-
ing analytical tool for detecting and determining the

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 199
F igu r e 8. 19F NMR spectra of trifluoromethylated dioxane lignin (A), after Dakin reaction (B) and reduction (C) with sodium
hydrosulfite.
Ta ble 6. Qu a n tita tive An a lyses of Ca r bon yl Gr ou p s in Sever a l Lign in s by Usin g 19F NMR Sp ectr oscop y
CO/C9
wt %
lignin sample
xj
s
xj
s
MW
Sucrolin acid hydrolysis (bagasse)
Alcell organosolv (mixed hardwoods)
steam explosion (yellow poplar)
dioxane acidolysis
0.12 ( 0.01
0.11 ( 0.01
0.13 ( 0.01
0.15 ( 0.01
0.0097
0.0086
0.010
1.89 ( 0.2
1.56 ( 0.2
1.57 ( 0.2
2.27 ( 0.2
2.90 ( 0.2
0.15
0.13
0.18
0.15
177.4
178.5
194.8
189.4
0.0095
kraft residualc
straw
0.018 ( 0.005
0.0011
0.25 ( 0.05
0.012
201.1
a
b
xj ) mean value. s ) standard deviation. c The amount of CO/C9 is not reported because the C9 unit cannot be defined.
Ta ble 7. Deter m in a tion of Tota l Am ou n t of Ca r bon yl Gr ou p s in Lign in s by Differ en t Tech n iqu es
detection method
lignin sample
19F NMR
oximation
UV-visa
MW
formula
Sucrolin acid hydrolysis (bagasse)
Alcell organosolv (mixed hardwoods)
steam explosion (yellow poplar)
0.12 ( 0.01
0.11 ( 0.01
0.13 ( 0.01
0.12 ( 0.12
0.10 ( 0.0
0.11 ( 0.04
0.03 ( NR
0.11 ( NR
0.09 ( NR
177.4
178.5
194.8
C9H8.3O2.2(OCH3)0.83
C9H7.7O1.9(OCH3)1.04
C9H7.8O2.5(OCH3)1.25
a
Reduction with sodium borohydride followed by UV-vis [standard deviation was not reported (NR)].
most prominent carbonyl-containing groups present in
different soluble lignins.
values using independent techniques. The proximity of
these data for three lignin samples qualifies the 19F
NMR technique as a new analytical tool for detecting
and determining carbonyl groups in lignins.
Con clu sion s. The quantitative trifluoromethylation
of carbonyl groups can be applied for the detection and
quantitative determination of the various carbonyl
groups present in lignins. By applying selective reac-
tions such as borohydride and hydrosulfite reductions
and Dakin oxidation, it became possible to assign a
number of prominent 19F NMR signals in trifluorom-
ethylated lignins. The quantification of the total amount
of carbonyls can be carried out using 3,3′-bis(trifluo-
romethyl)benzophenone as an internal standard. The
total amounts of carbonyls determined according to the
proposed technique in a variety of samples were found
to be different from one another and yet close to reported
LITERATURE CITED
Adler, E.; Ellmer, L. Coniferylaldehydgruppen im Holz und
in isolierten Lignin praparaten (Coniferaldehyde groups in
wood and in isolated lignin preparations). Acta Chem.
Scand. 1948, 2, 839-840.
Adler, J .; Marton, J . Zur Kenntnis der Carbonylgruppen im
Lignin. I (On the determination of carbonyl groups in lignin).
Acta Chem. Scand. 1959, 13, 75-96.
Ahvazi, B. C.; Argyropoulos, D. S. Quantitative trifluoro-
methylation of carbonyl-containing lignin model compounds.
J . Fluorine Chem. 1996a , 78, 195-198.

200 J. Agric. Food Chem., Vol. 47, No. 1, 1999
Ahvazi et al.
Ahvazi, B. C.; Argyropoulos, D. S. 19F Nuclear magnetic
resonance spectroscopy for the elucidation of carbonyl
groups in lignins: Part 1. Model compounds. J . Agric. Food
Chem. 1996b, 44, 8, 2167-2175.
Andersons, B.; Faix, O. Carbonyl group determination of
lignins. Proceedings of the 8th International Symposium on
Wood and Pulping Chemistry, Helsinki, Finland; 1995; Vol.
I, pp 559-566 (ISBN 952-90-6479-9).
Archipov, Y.; Argyropoulos, D. S.; Bolker, H. I.; Heitner, C.
31P NMR spectroscopy in wood chemistry. Part I, Model
compounds. J . Wood Chem. Technol. 1991a , 11 (2), 137-
157.
Archipov, Y.; Argyropoulos, D. S.; Bolker, H. I.; Heitner, C.
31P NMR spectroscopy in wood chemistry. Part II, Phosphite
derivatives of carbohydrates. Carbohydr. Res. 1991b, 220,
49-61.
agent for lignin functional group analysis materials. Pre-
sented at the 7th International Conference of Young Sci-
entists, Riga, Latvia, 1991.
Forsskåhl, I.; Tylli, H.; Olkkonen, C.; J anson, J . Photochem-
istry of quinones and hydroquinones on solid matrices.
Proceedings of the International Symposium of Wood and
Pulping Chemistry, Melbourne Australia; 1991; Vol. 2, pp
325-332.
Geiger, H.; Fuggerer, H. U¨ ber den Chemismus der Wiesener-
Reaktion auf Lignin. Z. Naturforsch. 1979, 34, 1471-1472.
Gellerstedt, G.; Pattersson, E. L. Light-induced oxidation of
lignin. Part 2. The oxidative degradation of aromatic rings.
Svensk. Papperstidn. 1977, 77, 15-21.
Gierer, J .; So¨derberg, S. U¨ ber die Carbonylgruppen des
Lignins. Acta Chem. Scand. 1959, 13, 127-137.
Glasser, W. G.; Kelley, S. S. Light stability of polymers. J . Pulp
Paper Can. 1987, 8, 795-851.
Green, J . W. Determination of carbonyl group. In Methods in
Carbohydrates Chemistry; Whistler, R. I., Ed.; Academic
Press: New York, 1963; Vol. III, pp 49-54.
Grundmann, C. Chinone: p-Chinone der Benzol-und Naph-
thalin-Reihe. In Methoden der Organischen Chemie; Mu¨ller,
E., Ed.; Houben-Weyl, Georg Thieme Verlag: Stuttgart,
1977; Vol. VII/3a, pp 648-650.
Harkin, J . M. Lignin-A natural polymeric product of phenol
oxidation. In The Chemistry of Phenolic Resins; the Forma-
tion, Structure and Reactions of Phenolic Resins and Related
Products; Wiley: New York, 1956; Chapter 6, pp 243-312.
Harris, R. K. Nuclear Magnetic Resonance Spectroscopy; A
Physiochemical View; The Universities Press: Belfast,
Northern Ireland, 1983; Chapter 8, p 189.
Hergert, H. Infrared spectra. In LigninssOccurrence, Forma-
tion, Structure and Reaction; Sarkanen, K., Ludwig, C., Eds.;
Wiley-Interscience: New York, 1971; pp 267-297.
Heuser, E. Analytik von Chinonen. In Methoden der organis-
chen Chemie (Houben-Weyl) 4. Aufl., Bd II Analytische
Methoden; Mu¨ller, E., Bayer, O., Meerwein, H., Ziegler, K.,
Eds.; George Thieme: Stuttgart, Germany, 1953; pp 434-
472.
Hon, D. N. S.; Glasser, W. On possible chromophore structures
in wood and pulpssa survey of the present state of knowl-
edge. Polym. Plast. Technol. Eng. 1979, 12, 157-159.
Hortling, B.; Tamminen, T.; Kentta¨, E. Determination of
carboxyl and non-conjugated carbonyl groups in dissolved
and residual lignins by IR spectroscopy. Holzforschung
1997, 51, 405-410.
J ansherkar, H.; Fiechter, A. Lignin: biosynthesis, application,
and biodegradation. Adv. Biochem. Eng./ Biotechnol. 1983,
27, 119-178.
Kolboe, S.; Ellefsen, O. Infrared investigations of lignin; A
discussion of some recent results. Tappi 1962, 45 (2), 163-
166.
Lachenal, D.; Fernandes, J . C.; Froment, P. Behaviour of
residual lignin in kraft pulp during bleaching. J . Pulp Paper
Sci. 1995, 21, 173-177.
Lapierre, C.; Monties, B.; Guittet, E.; Lallemand, J . Y.
Quantitative measurements in hardwood lignin 13C NMR
spectra. Holzforschung 1984, 39, 367-368.
Lebo, S. E.; Lonsky, W. F.; McDonough, T. J .; Medvecz, P. J .;
Dimmel, D. R. The occurrence of light-induced formation of
ortho-quinonoid lignin structures in white spruce refiner
mechanical pulp. J . Pulp Paper Sci. 1990, 16, 5, 139-143.
Lin, S. Y.; Dence, C. W. Methods in Lignin Chemistry; Springer
Series in Wood Science; Springer-Verlag: Berlin, 1992;
Chapter 7.4, pp 446-457.
Lin, S. Y.; Kringstad, K. P. Some reactions in the photoinduced
discoloration of lignin. Norsk. Skiksind. 1971, 25, 252-256.
Lindberg, B.; Misiorny, A. Quantitative analysis of reducible
carbohydrates by means of sodium borohydride. Svensk.
Papperstidn. 1952, 55, 13-14.
Lindberg, B.; Theander, O. Quantitative determination of
carbonyl groups in oxycellulose by means of sodium boro-
hydride. Svensk. Papperstidn. 1954, 57, 83-85.
Argyropoulos, D. S. Quantitative phosphorus-31 NMR analysis
of lignins, A new tool for the lignin chemist. J . Wood Chem.
Technol. 1994a , 14 (1), 45-63.
Argyropoulos, D. S. Quantitative phosphorus-31 NMR analysis
of six soluble lignins. J . Wood Chem. Technol. 1994b, 14
(1), 65-82.
Argyropoulos, D. S. 31P NMR in Wood Chemistry, A review of
recent progress. Res. Chem. Intermed. 1995, 21 (No. 3-5),
373-395.
Argyropoulos, D. S.; Heitner, C. 31P NMR spectroscopy in wood
chemistry. Part VI. Solid state 31P NMR of trimethyl
phosphite derivatives of chromophores and carboxcylic acids
present in mechanical pulps: A method for the determina-
tion of ortho-quinones. Holzforschung Suppl. 1994, 48, 112-
116.
Argyropoulos, D. S.; Menachem, S. B. Lignin. Adv. Biochem.
Eng./ Biotechnol. 1997, 57, 127-158.
Argyropoulos, D. S.; Archipov, Y.; Bolker, H. I.; Heitner, C.
31P NMR spectroscopy in wood chemistry. Part IV, Lignin
models: Spin-lattice relaxation times and solvent effects
in 31P NMR. Holzforschung 1992, 47, 50-56.
Argyropoulos, D. S.; Bolker, H. I.; Heitner, C.; Archipov, Y.
31P NMR spectroscopy in wood chemistry. Part V, Qualita-
tive analysis of lignin functional groups. J . Wood Chem.
Technol. 1993, 13 (2), 187-212.
Argyropoulos, D. S.; Heitner, C.; Morin, F. G. 31P NMR
spectroscopy in wood chemistry. Part III. Solid state 31P
NMR of trimethyl phosphite derivatives of chromophores
in mechanical pulp. Holzforschung 1994, 46, 211-218.
Argyropoulos, D. S.; Heitner, C.; Schmidt, J . A. Observation
of quinonoid groups during the light-induced yellowing of
softwood mechanical pulp. Res. Chem. Intermed. 1995, 9
(No. 3-5), 263-274.
Bailey, C. W.; Dence, C. W. Reactions of alkaline hydrogen
peroxide with softwood lignin model compounds, spruce
milled-groundwood lignin and spruce groundwood. Tappi
1969, 52 (3), 491-500.
Barrelle, M. A new method for the quantitative 19F NMR
spectroscopic analysis of hydroxyl groups in lignins. Holz-
forschung 1993, 47, 261-267.
Barrelle, M. Improvements in the structural investigation of
lignins by 19F NMR spectroscopy. J . Wood Chem. Technol.
1995, 15 (2), 179-188.
Brezny, R.; Schraml, J . Silicon-29 NMR spectral studies of
kraft lignin and related model compounds. Holzforschung
1987, 41, 293-298.
Brunow, G.; Eriksson, B. R-carbonyl groups as sensitizers in
the photo-dehydrogenation of phenolic structures in lignin.
Acta Chem. Scand. 1971, 25, 2779-2781.
Castellan, A.; Nourmamode, A.; J eager, C.; Forsskåhl, I.
Photochemistry of lignocellulosic materials. ACS Symp. Ser.
1993, No. 531, 60-76.
Faix, O. Classification of lignins from different botanical
origins by FTIR spectroscopy, Holzforschung Suppl. 1991,
45, 21-27.
Fieser, F.; Fieser, M. Reagents for Organic Synthesis; Wiley:
New York, 1967; pp 1080-1082.
Fillppov, V.; Archipov, Y.; Argyropoulos, D. S. 4,4,5,5-tetra-
methyl-1,3,2-ioxaphospholanyl chloride-derivatization re-

19F NMR of Carbonyl Groups in Lignins
J. Agric. Food Chem., Vol. 47, No. 1, 1999 201
Lundquist, K. NMR studies of lignins. 2. Interpretation of the
1H NMR spectrum of acetylated brich lignin. Acta Chem.
Scand. 1979a , B33, 27-30.
Obst, J . R.; Landucci, L. L. Quantitative 13C NMR of lignins.
Methoxyl: aryl ratio. Holzforschung Suppl. 1986, 40, 87-
92.
Pan, X.; Lachenal, D. Structure and reactivity of spruce
mechanical pulp lignins IV: 13C NMR spectral studies of
isolated lignins. J . Wood Chem. Technol. 1994 14 (4), 483-
506.
Rabjohn, N. Organic Syntheses; revised edition of annual
volumes; - Wiley: New York, 1963; Collect. Vol. IV, pp 13-
18.
Reeves, R. H.; Pearl, I. A. Reaction products formed upon the
alkaline peroxide oxidation of lignin-related model com-
pounds. Tappi 1965, 48 (2), 121-125.
Robert, D.; Brunow, G. Quantitative estimation of hydroxyl
groups in milled wood lignin from spruce and in a dehydro-
genation polymer from coniferyl alcohol using 13C NMR
spectroscopy. Holzforschung 1984, 38, 85-90.
Sarkanen, K. V.; Ludwig, C. H. In LigninssOccurrence,
Formation, Structure and Reactions; Wiley-Interscience:
New York, 1971; pp 461-463, 518, 532.
Lundquist, K. NMR studies of lignins. 3. 1H NMR spectroscopic
data for lignin model compounds. Acta Chem. Scand. 1979b,
B33, 418-420.
Lundquist, K. NMR studies of lignins. 4. Investigation of
1
spruce lignin by H NMR spectroscopy. Acta Chem. Scand.
1980, B34, 21-26.
Manatt, S. L. Characterization of functional groups by nuclear
magnetic resonance. I. Classification of alcohols from the
fluorine-19 spectra of trifluoroacetates. J . Am. Chem. Soc.
1966, 88 (6), 1312-1324.
Marton, J .; Adler, E. Carbonyl groups in lignin III. Mild
catalytic hydrogenation of Bjo¨rkman lignin. Acta Chem.
Scand. 1961, 15, 370-383.
Marton, J .; Adler, E.; Persson, K. I. Carbonyl groups in lignin
IV*. Infrared absorption studies and examination of the
volumetric borohydride method**. Acta Chem. Scand. 1961,
15 (2), 384-392.
Milne, T. A.; Chum, H. L.; Agblevor, F.; J ohnson, D. K.
Standardized analytical methods. Biomass Bioenergy 1992,
2, 341-366.
Zakis, G. F. Functional analysis of lignins and their deriva-
tives. Tappi 1994, 61-76.
Miyake, M. Kraft cooking with addition of reducing agents.
III. Effect of yield increase by hydrazine or hydroxylamine
and its mechanism. J . J pn. Tappi 1970, 24, 4-80.
Nieminen, O. J .; Pulkkinen, E.; Rahkamaa, E. Determination
of hydroxyl groups in kraft pine lignin by Silicon-29 NMR
spectroscopy. Holzforschung 1989, 43, 303-307.
Received for review May 5, 1998. Revised manuscript received
September 14, 1998. Accepted September 22, 1998.
J F980431P