вход по аккаунту


The interactions between wood components and formaldehyde-based resins. I. Monofunctional resin model compounds

код для вставкиСкачать
Die Angewandte Makromoiekuiare Chemie 113 (1989) 137- 152 (Nr. 2865)
Ecole FranGaise de Papeterie (INPG), BP 65, F-38402 Saint Martin D’H&es, France
The Interactions between Wood Components and
Formaldehyde-Based Resins
1. Monofunctional Resin Model Compounds*
F. Mora, F. Pla, and A. Gandini**
(Received: 10 October 1988)
Results are presented on the chemical interactions occurring when some monofunctional model compounds simulating the structure and reactivity of thermosetting phenol- and urea-formaldehyde resins are mixed with wood components (hemicelluloses,
cellulose, and lignins). Whereas hemicelluloses clearly show a propensity to give
condensation products with these model compounds, lignins can react less readily in
some instances depending on the delignification procedure used to isolate them and
on the type of model compound. Cellulose did not react under the conditions chosen,
mostly because of its crystalline character.
Die Resultate von Untersuchungen uber die chemische Wechselwirkung zwischen
einigen monofunktionellen Modellverbindungen, die die Struktur und Reaktivitat
von warmehktbaren Phenol- und Harnstoff-Formaldehyd-Harzen simulieren, mit
Holzkomponenten (Hemicellulosen, Cellulose und Lignin) werden dargelegt. Wlhrend Hemicellulosen deutlich dazu neigen, Kondensationsprodukte mit diesen Modellverbindungen zu ergeben, reagieren Lignine in manchen Fallen weniger leicht und
abhangig davon, durch welchen ProzeR sie gewonnen worden sind, und abhangig von
der Art der Modellverbindung. Cellulose reagierte wegen ihrer Kristallinitat nicht unter den gegebenen Bedingungen.
Presented in part at the Fourth International Symposium on Wood and Pulping
Chemistry, Paris, April 27 - 30, 1987.
** Correspondence author.
0 1989 Huthig & Wepf Verlag, Base1
F. Mora, F. Pla, and A. Gandini
Formaldehyde-based resins, namely with phenol, urea, and melamine, are
widely used both as bulk materials and as fillers and adhesives. Among the
latter applications, some involve wood as the structural element of classical
composites, e. g. particle boards'.
The purpose of the present investigation was to gain a better insight into
the details of the interactions occurring when wood chips are treated with the
liquid prepolymer based on one of the combinations mentioned above and
when the thermosetting of the corresponding resins is carried out to make
the actual boards.
Three types of interactions can be envisaged between the crosslinking
polymer matrix and the wood particles: (i) purely physical adhesion based on
the thermodynamic interactions between the two types of surfaces to form a
strongly bound interface; (ii) physico-chemical bonding between specific
chemical polar functions present on each type of surface, e. g. hydrogen
bonding; (iii) chemical reactions between groups on either side of the interface producing covalent links which irreversibly join the two originally independent components. Obviously, these features are not mutually exclusive
and one could readily envisage the contribution of the three phenomena with
a different degree of importance in the making of the agglomerate.
It seemed logical to tackle this complex phenomenology starting from the
chemical aspect, viz. the possibility of reactions between wood (components)
and resin (prepolymer or its setting products). Even within this specific
framework, the variables are too numerous for a straightforward study of
the system wood-chips-plus-resin as such. Indeed, it was felt that only
through a progressive approach starting from simple combinations, one
could hope to advance and reach a better knowledge of the possible
mechanistic features pertaining to the actual composite materials.
Thus, wood was replaced by the three major components which make up
its texture, namely cellulose, lignin, and hemicelluloses. Each of these components was taken individually as a potential reactant towards the resin.
Conversely, model compounds simulating the reactive sites of specific formaldehyde resins were envisaged. This approach is more direct and explicit
than a previous investigation', based on deductive conclusions, with wood
and cellulose.
The present report gives an account of the procedures adopted for
obtaining and characterizing both the pure wood components and the pure
Wood Componentsand Formaldehyde-Based Resins
resin model compounds and of the results dealing with the interactions of
specific pairs of potential reagents, i. e. the possible sources of chemical
reaction when wood and resin are mixed and heated to set.
Isolation and Characterization of Hemicelluloses
I. NaOH Extraction
a) From an Aspen Wood:
Aspen wood contains xylans which can be extracted with alkali without previous
delignification procedures3. Thus, aspen wood was ground to sawdust and treated
first with a benzene-ethanol mixture (2: 1 v/v) and then with boiling water in order to
remove pigments and pectic polysaccharides. The wood meal was then shaken for two
days with an excess of aqueous sodium hydroxide (4.3 N) at 5°C in a nitrogen
atmosphere to avoid polysaccharides degradation by the peeling reaction. At the end
of this treatment the suspension was filtered and the filtrate poured into an excess of
methanol-acetic acid mixture (10: 1 v/v): the precipitate formed was centrifuged,
dialyzed against distilled water and freeze-dried.
Carbohydrate analyses were conducted on these hemicelluloses. After acid
hydrolysis by trifluoroacetic acid (2 N, 100°C, 4 h), the excess acid was removed by
vacuum evaporation and the resulting sugars analysed by gas-liquid chromatography
as alditol acetate derivatives. The results were typically as follows: 4% arabinose and
85% xylose; to these neutral sugars one must add 5 - 10% of 4-0-methylglucuronic
acid which is not detected by GLC under our experimental conditions. The presence
of a small amount of arabinose can be attributed to ramifications on the anhydroxylose main chains of the xylans or to contamination by pectic arabans.
b) From a Birch Kraft Pulp:
The xylans from an industrial bleached Kraft pulp were extracted at room temperature by an aqueous 4.3 N sodium hydroxide solution, as described above.
The carbohydrate analysis gave the following results: 91% xylose and traces of
F. Mora, F. Pla, and A. Gandini
11. DMSO Extraction
In order to assess the relative importance of acetyl moieties borne by in-situ xylans,
hemicelluloses contained in oak wood were extracted with DMS04 to avoid hydrolysis of these groups by the alkaline treatment. The procedure involved the same
treatment of the sawdust samples as described above to remove pigments and pectic
polysaccharides from aspen wood. This was followed by a delignification step with
sodium chlorite (solution buffered with acetic acid) for one day at 40°C. The
resulting delignified materials were stirred with DMSO at room temperature for three
days. After filtration, the extracted hemicelluloses were precipitated, separated,
purified, and dried as described above for the NaOH treatment.
Fig. 1 shows the IR spectra of these three hemicelluloses. It is clear that acetyl
groups were removed during the pulping process. These groups are still present on
xylans extracted with DMSO (see the band at 1700 cm-I). Xylans from aspen wood
represent an intermediate situation.
Fig. 1.
Wavenumber Icm”)
IR spectra of three hemicelluloses. A: extracted by 4.3 N NaOH from a birch
pulp; B: extracted by 4.3 N NaOH from an aspen wood; C: extracted by
DMSO from an oak wood.
Wood Components and Formaldehyde-Based Resins
Purijication and Characterization of Cellulose
Cotton linters were submitted to two successive 6 h extractions, one with ethanol,
the other with diethyl ether. These were followed by a 10 h treatment with a boiling
0.01 N NaOH solution (linters/soln. = 0.025 w/v). Every hour, 10% of the liquid
was removed and replaced by an equivalent volume of fresh NaOH solution. The
linters were then treated at 70°C with a sodium chlorite solution buffered at pH 4.9
(linterddry chlorite = 0.034 w/w). The cellulose samples were finally washed with
distilled water and air dried’.
Their characterization consisted in determining (i) the sugar composition (GLC)
after a 72% sulfuric acid hydrolysis6 (see above for NaOH-extracted hemicelluloses):
glucose was the only sugar detected; and (ii) the weight-average degree of polymerization by size-exclusion chromatography of the tricarbanilated derivatives’ : the values
were close to 4600.
Origin and Characterization of Lignins
Two different types of lignins were used in this study:
I. Organosolv spruce lignin, kindly provided by MD Organocell (Munich). This
material was extracted as reported’ and characterized by I3C-NMR. The results of
this structural analysis gave the following incidence of side groups per aromatic unit:
0.18 C 3 chains and 0.7 methoxy groups; the average number of unsubstituted aromatic carbons was 2.0 per ring. These data suggest that the delignification treatment was
indeed quite degradative given the important loss of propylic side chains from the
classical “phenylpropane” units.
11. Kraft pine lignin, purchased from the Westvaco company. The elemental
analysis of this material gave: C = 59.95%; H = 5.83%; S = 4.11 Vo; 0 = 30.11 9’0.
It was found that the total OH content was 1.3 groups per “phenylpropane” unit, of
which 0.5 belonged to alcoholic functions, about 0.7 to phenolic moities, and the rest
to carboxylic acid functions. This lignin contained 0.7 methoxy groups per “phenylpropane” units. Quite clearly, the reactive side groups are more abundant here than
in the organosolv materials (see above), suggesting less drastic structural modifications during delignification.
Synthesis and Characterization of Resin Model Compounds
As mentioned in the introduction it was decided to limit the functionality of these
compounds to unity for the first approach to the reactivity towards wood components and so to avoid gel formation. The following compounds were chosen as representative of phenolic and urea resins, respectively:
F. Mora, F. Pla, and A. Gandini
\ /"CH20H
(M 11U)
(M 13U)
Fig. 2.
Model compounds synthesized.
I. 2,6-Dimethyl-4-hydroxymethylphenol(DMHMP, Fig. 2). The blocking of the 2
and 6 positions of the ring leaves the 4 position as the only possible reactive site since
it is well known that the C 3 and C5 positions do not intervene in the resinification
reactions. Commercial 2,6-dimethylphenol was treated with a 30% formaldehyde
solution in the presence of NaOH. The reaction was conducted in water at room
temperature during 5 h with a molar ratio of 1 : 1:l of the two reagents and the
catalyst. The resulting mixture was neutralized with formic acid and the ensuing
precipitate filtered, recrystallized from a methanol-water mixture (1 : 1 v/v) and
vacuum dried at room temperature. The infrared spectra of the phenols before and
after methylolation showed the characteristic differences in the 3000 - 3500 cm-I
region (OH stretching), around 2950 cm-' (CH stretching) and around 1000 cm-'
(C-0 stretching), clearly showing that methylolation had indeed taken place. This
was confirmed and placed into a quantitative pattern by a 'H-NMR spectrum
(Fig. 3).
11. 1,l- and -1,3-dimethylmonomethylolureas(M11U and M13U, Fig. 2). It is
known that tetrasubstituted ureas are very difficult (even impossible) to prepare by
direct substitutions9. Thus, we decided to use 1,1-dimethylurea and 1,3-dimethylurea
as starting reagents for preparing the corresponding monomethylolated model
compounds. The syntheses were carried out by the slow addition of a 30% aqueous
solution of formaldehyde to the respective dimethylurea solutions cooled with ice
(reaction temperature: 5 to IOOC) and kept at pH slightly above 10 during the whole
1 42
Wood Components and Formaldehyde-Based Resins
1 0 0 0 3500 3000 2500 2000
W a v e numb er
( crn-'1
Fig. 3. IR and ' H-NMR spectra of 2,6-dimethyl-4-hydroxymethylphenol.
course of the reactions. The consumption of formaldehyde was followed by the
sulphite methodlo and the reactions were allowed to proceed until the concentration
of formaldehyde reached a constant value (ca. 24 h). The resulting solutions were
then neutralized with HCl and freed from the salt by treatment with specific ionexchange resins (amberlite resins IRN-78 and IR-120). The products were isolated by
vacuum evaporation of the water: the crystalline compounds obtained were analysed
by IR and 'H-NMR spectroscopy (see Fig. 4 for M13U). In both instances monomethylolation had taken place without any detectable hint of dimethylolation. Moreover, l , l -dimethylurea was found to react more slowly than its l ,3-homologue which
gave quantitative yields readily.
It is important to note that if the dimethylureas were made to react with formaldehyde at lower pH conditions and at higher temperatures, methylolation was rapidly
followed by condensation reactions and the major products were the corresponding
condensed dimers, i.e. unreactive compounds of no use to our investigation. The
structure of these products was ascertained by IR, ' H-NMR spectroscopy, and mass
spectrometry. This point is particularly relevant to our further studies since it
emphasizes that the monomethylolated dimethylureas prepared as model compounds
have a strong tendency to self-condense at pH < 8.
F. Mora, F. Pla, and A. Gandini
L n
LOO0 3500 3000 2500 2000
Fig. 4.
IR and I H-NMR spectra of 1,3-dimethyl-monomethylolurea.
The Reactivity of Model Compounds towards Wood Components
I. Xylans + substituted ureas: Xylans (0.5 g) with different degrees of acetylation
were dissolved in 10 ml of a mixture of DMSO and water (4:l v/v). 0.25 g of
monomethylolated 1 , l - or 1,3-dimethylurea were then added and the pH adjusted to
4 with acetic acid. The mixtures were stirred for 20 h at 90°C. Then the products were
washed with methanol and vacuum dried.
11. Xylans + substituted phenol: The procedure described above was adopted here
except that the pH was raised to 10 with NaOH.
111. Lignins + substituted ureas: Organosolv or Kraft lignin (1 .O g) were dissolved
in 10 ml of a DMSO-water mixture (4: 1 v/v) together with 0.25 g of monomethylolated 1 , l - or 1,3-dimethylurea. The mixtures were then brought to pH 4 and stirred
for 20 h at 90°C. At the end of this treatment the lignins were precipitated in an
excess of a 0.01 N HCl solution, washed with distilled water and vacuum dried.
IV. Lignins + substituted phenol: The procedure was the same as that employed
for the substituted ureas, but the pH was kept at 10 with NaOH.
Wood Components and Formaldehyde-Based Resins
V. Cellulose: The reactivity of this polysaccharide towards model urea and phenol
compounds was determined under the same conditions as those described for xylans.
Since cellulose is insoluble in the DMSO-water mixture, all these experiments
involved heterogeneous systems.
Analytical Techniques
Elementary analysis, GLC, H- and "C-NMR spectroscopy, continuous-wave IR
spectroscopy, as well as certain determinations of functional groups were conducted
using conventional procedures. The IR spectra dealing with the interaction
experiments and the differential scans thereby were carried out with a Bruker IFS 48
Fourier-Transform Spectrophotometer. The subtractions operated by this instrument
were of course normalized with respect to a characteristic strong and sharp peak
specific to the context under investigation.
The determination of formaldehyde during the methylolation reactions was carried
out following the classical sulphite method". The interference of bound formaldehyde was taken into account by extrapolating the results to zero reaction time.
Results and Discussion
The approach adopted to establish a first qualitative criterion as to
whether the monofunctional resins model compounds react with wood
components was to submit the isolated products resulting from the
interaction to an FT-IR analysis. The ensuing spectrum was compared with
that of the starting wood component by a normalized subtraction procedure
operated by the instrument. Significant differences arising from the
consumption of functionalities existing on the wood components and from
the presence of model compound moieties could be detected much more
accurately by this technique than by others.
The Behaviour of Xylans
The interaction of deacetylated or poorly acetylated hemicelluloses with
2,6-dimethyl-4-hydroxymethylphenolresults in condensation reactions as
suggested by the differential spectrum shown in Fig. 5B. The most relevant
features of this scan are: (i) the negative peaks at about 3400 and 1OOO cm -'
(the stretching and bending modes of pyranose-ring hydroxy groups, respec145
F. Mora, F. Pla, and A. Gandini
Wavenumber (crn-'J
Fig. 5. Differential IR spectra of A: xylans-M13U minus xylans; B: xylansDMHMP minus xylans.
tively) reflecting a decrease in OH concentration in the xylans; (ii) the positive peaks at about 3600 cm-' (phenolic OH), 1490 and 1200 cm-' (substituted aromatic ring and phenolic C-0, respectively) and the weaker absorption pattern below 1000 cm -' (CH out of plane deformation vibrations in
aromatics). Thus the occurrence of condensation reactions between the
methylol group on the phenolic model and the OH functions of xylans seems
well established.
The interaction of xylans with 1,3-dimethyl-monomethylolurea also
results in condensation reactions as indicated by the differential spectrum
given in Fig. 5A. The negative peaks are the same as above (Fig. 5B)
whereas the positive ones arise from the chemical incorporation of trisubstituted urea moieties giving a sharp NH band above 3600 cm -' and a
characteristic C = O resonance at about 1700 cm-'. Both the 1,l- and the
1,3-dimethylurea derivatives give evidence of reacting readily with hemicelluloses.
Wood Components and Formaldehyde-BasedResins
W o v e n urnb e r
1c rn-’1
Fig. 6. Differential IR spectra of A: acetylated xylans-DMHMP minus acetylated
xylans; B: acetylated xylans-Ml3U minus acetylated xylans.
This reactivity is lowered when the condensation is conducted in the
presence of “native” xylans (DMSO-extracted xylans) and particularly when
the reaction involves urea model compounds, as shown in Fig. 6 . In the case
of phenols, the spectrum shows the characteristic peaks of the model
compound. The negative acetyl peak (= 1700 cm - ’) indicates that during the
reaction at pH = 10 acetyl side groups were hydrolysed to produce new
hydroxy groups which react further with the phenolics. This sequence of
events does not occur in reactions with ureas which are conducted at pH 4
(spectra 6A and 6B).
F. Mora, F. Pla, and A. Gandini
The Behaviour of Lignins
The interaction of organosolv lignin with 2,6-dimethyl-4-hydroxymethylphenol leads to differential spectra like that shown in Fig 7C. Attachment of
the model compound to the lignin structure is clearly indicated by the pattern
of positive peaks typical of the trisubstituted phenol (Fig. 5B) and by the
negative carbonyl peak around 1670 cm -’ which suggests that the condensation reactions occur between the methylol group of the model phenol and
either the aldehyde or the carboxylic acid groups borne by the lignin, or
The differential spectra A and B in Fig. 7 relate to the products of the
interactions between dimethyl-monomethylolureas and kraft and organosolv
lignin, respectively. The intensity of both positive and negative peaks is too
low to justify any reasonable conclusion as to the occurence of condensation
reactions. It can however be argued that the lower reactivity of lignins
towards urea model compounds with respect to that encountered towards
phenolics could arise from steric factors pertaining to both reagents, i. e. the
Fig. 7.
Woven urnbe r
( c m-’l
Differential IR spectra of A: Kraft lignin-Mi 3U minus Kraft lignin; B: organosolv lignin-Mi 3U minus organosolv lignin; C: organosolv ligninDMHMP minus organosolv lignin.
Wood Components and Formaldehyde-Based Resins
rather crowded lignin structures coupled with the trisubstituted ureas.
Neither the combination lignin-model phenol nor those involving the xylans
were characterized by such a “double” steric effect. More work is needed to
gain a better understanding of these interactions.
The Behaviour of Cellulose
According to the IR spectra shown in Fig. 8 no reaction could be detected
between ureas and cellulose and only a very modest reactivity between the
phenolic model compound and cellulose. This is confirmed by the absence of
W a v e nurn ber (crn-’)
Fig. 8. Differential IR spectra of A: cellulose-DMHMP minus cellulose; B: cellulose-M13U minus cellulose.
F. Mora, F. Pla, and A. Gandini
c,. c,. c,
Fig. 9. CP/MS I3C-NMR spectra of A: standard cellulose; B: cellulose after treatment with 1,3-dirnethyl-monomethylolurea;C: cellulose after treatment
with 2,6-dimethyl-4-hydroxymethylphenol.
Wood Components and Formaldehyde-BasedResins
any detectable difference in the solid state NMR spectra of the cellulose
before and after treatment with the model compounds (Fig. 9).
The difference in reactivity of the hydroxy groups borne respectively by
hemicelluloses and cellulose cannot be attributed to the change in reaction
conditions (cellulose was not soluble in our reaction media, whereas xylans
dissolved completely), because the interaction of hemicelluloses with difunctional compounds in heterogeneous conditions gave the same positive
results“ as in the present work. It can be concluded that the poor reactivity
of cellulose is due to its crystalline structure and more generally to the low
accessibility of its OH groups.
These experiments on the potential reactivity between wood components
and formaldehyde-resin model compounds have shown that hemicelluloses
react readily with both urea-formaldehyde and phenol-formaldehyde monofunctional model compounds. The presence of acetyl side groups in xylans
reduces this reactivity, . particularly against urea-formaldehyde model
compounds. Lignins are a more difficult substrate to study, because their
reactivity depends on the mode of delignification, i. e. their structures can
vary considerably and are unclearly correlated to that existing inside the cell
wall. Cellulose is not a potential reactant towards the resin model
compounds. In general terms phenolic model compounds have a higher
reactivity towards wood components than trisubstituted monofunctional
The financial assistance of DSM Research BV is gratefully acknowledged
as well as their collaboration in taking solid state NMR spectra.
J . F. Oliver (Ed.), Adhesion in Cellulosic and Wood-Based Composites, Plenum
Press, New York 1981, p. 127
G. G. Allan, A. N. Neogi, J . Adhes. 3 (1971) 13
M. Zimbo, T. E. Timell, Sven. Papperstidn. 68 (1965) 647
B. Lindberg, K. G. Rosell, S. Svensson, Sven. Papperstidn. 76 (1973) 30
J. M. Lauriol, P. Froment, F. Pla, A. Robert, Holzforschung 41 (1987) 1 0 9
J. F. Saeman, W. E. Moore, R. L. Mitchell, M. A. Millet, Tappi J. 37 (1954) 336
D. Gast, C . Ayla, J. Puls, 2nd EC Conference “Energy from Biomass”, Applied
Science Publishers Ltd., London 1983, p. 829 - 83 1
F. Mora, F. Pla, and A. Gandini
H. Cheradame, M. Detoisien, A. Gandini, F. Pla, G. Roux, Br. Polym. J. 21
(1989) 269
B. Meyer (Ed.), Urea formaldehyde resins, Addison Wesley Publishing Company
Inc., Reading 1979, p. 92
J. F. Walker (Ed.), Formaldehyde, R. E. Krieger Publishing Company,
Huntington, N. Y. 1975, p. 486
F. Mora, F. Pla, A. Gandini, to be submitted.
Без категории
Размер файла
568 Кб
base, monofunctionalize, interactions, mode, compounds, wood, components, resins, formaldehyde
Пожаловаться на содержимое документа