Urea-Resorcinol-Formaldehyde Adhesives

of Low Resorcinol Content

E. SCOPELITIS and A. PIZZ1*

Department of Chemistry, University of the Witwatersrand, Johannesburg, South Africa

SYNOPSIS
Low resorcinol content cold-settingurea-resorcinol-formaldehyde (URF) resins were pre-
pared. Most of these adhesives had water-resistantproperties, whereas some also presented
weather- and boil-proof adhesive characteristics and performance. URF resins of acceptable
adhesive performance were also prepared by inducing branching by introducing a certain
amount of resorcinol or of polyflavonoid mimosa tannin extract as a branching unit. Re-
sorcinol content as low as 12.7 and 12.15% on total liquid resin were obtained by
this method while still maintaining adhesive performance and water-resistant properties.
0 1993 John Wiley & Sons, Inc.

INTRODUCTION content. This is a more difficult undertaking than

for PRF resins, because of an additional effect mil-
The field of adhesives for wood has been dominated itating against such resins: Although a decrease of
during the last half-century by two main classes of resorcinol appears, in principle, possible (as for the
synthetic resins: aminoplastic and polyphenolic ad- PRF) , too low an amount of resorcinol, while pos-
hesives. The latter have been used mostly for the sibly still affording good dry strength of the bonded
production of weather- and boil-proof wood adhe- joint, may not be able to protect the urea-formal-
sives. Among phenolics, phenol-resorcinol-form- dehyde ( U F ) backbone of the resin from water de-
aldehyde (PRF) resins are the ones used as cold- generation, possibly causing catastrophic strength
setting adhesives for the preparation of weather- loss in the wet strength of the joints.
proof structural glulam and fingerjointing.’ Since
resorcinol is an expensive chemical, significant re-
ductions in the cost of such adhesives must, by ne- APPROACH
cessity, be coupled with a decrease in their resorcinol
content. Recently, a step forward has been reported2 Supposing that the average degree of polymerization
in the formulation of PRF adhesives of lower resor- of the UF resin before resorcinol addition is n,then
cinol content. This was the mistakenly called the the URF resin produced can be schematically rep-
“branching principle”’; in practice, a very rapid resented as 2*3
doubling of the average molecular mass of the liquid
resin obtained by the use of small amounts of very resorcinol -CH2-
reactive tri- or polyfunctional monomer units.3 In
( - urea -CH, -) -
R resorcinol
this article, the same approach already reported for
PRF resins has been used to prepare urea-resorci-
nol-formaldehyde (URF) resins of low resorcinol where n 2 1 in integer numbers.
If a chemical molecule capable of extensively
branching, or of rapidly doubling linearly the mo-
* To whom correspondence should be addressed. lecular weight, of the UF and URF resins is used
Journal of Applied Polymer Science, Vol. 48, 2135-2146 (1993) after the preparation of the UF resin, then a URF
0 1993John Wiley & Sons, Inc. CCC 0021-8995/93/122135-12 of lower resorcinol content will be produced. Thus,

2135
2136 SCOPELITIS AND PIZZI

resorcinol - ( - CH2-urea- ) -CH2- cantly lower than the 16-18% of today’s commercial
n
PRF resins. Many molecules can be used as addi-
branching -CH2- ( -urea -CH2- ) tional branching units? In the URF resins presented
unit in this work, the additional tridimensional branch-
-resorcinol
n
+ small amounts tridimensionally ing units used were resorcinol and a polymeric nat-
ural resorcinolic material, mimosa polyflavonoid
branched on the “branching unit” tannin,’ the latter being, on average, a pentafunc-
tional unit.
where n 2 1 in integer numbers.
The lower resorcinol requirement is achieved by
halfing the terminal resorcinol needed in the resin EXPERIMENTAL
as well as by an increase in viscosity of the resin of
the same solids content. One major expected differ- Method of Preparation of a Nonadditionally
ence from PRF adhesives, which are mostly linear, Branched URF Resin
is that in URF resins tridimensional branching is One hundred grams of a commercial UF (UF = 1 :
already present in the basic UF resin, urea being 1.5) resin was added into a 250 mL round-bottom,
effectively trifunctional (in principle, tetrafunc- three-necked flask equipped with a condenser and
tional) in its reaction with formaldehyde. That UFs a thermometer. The flask was placed in a warm wa-
are always tridimensionally branched can easily be
seen by I3C-NMR Thus,
-
ter bath ( 65°C). When the resin reached 65”C,
the appropriate amount of urea and 20 g of distilled
water was added to lower the formaldehyde/urea
CH, -resorcinol ratio. The amounts of urea are listed below. These
I amounts are for 100 g of commercial UF ( U : F
resorcinol - ( -CH, -urea -) -CH, -resorcinol
= 1 : 1.5) resin.

where n 2 1in integer numbers.

Thus, if additional branching is introduced by a U : F Ratio 1 : 1.5 1 : 1.25 1 : 1.1 1 : 1 1 : 0.8 1 : 0.5
further branching molecule, the extensively
branched URF can be represented as follows: Mass of urea
added ( g ) 0 7.7 14.1 19.2 33.3 76.4
CH, -resorcinol
I
resorcinol - ( -CH2- urea 7CH, CHZ- The reaction mixture was allowed to stir for 10 min
\ / at 65°C. During this period, pH was maintained be-
branching
molecule
tween 8 and 8.5 with 22% aqueous sodium hydroxide
solution. The pH of the reaction mixture was then
CH2- resorcinol
decreased to 5-5.3 with 30% formic acid and the
I
( - urea -CH, -) - resorcinol reaction mixture was allowed to stir at 65°C for a
further 10 min. The urea-formaldehyde solutions
where n 2 1 in integer numbers obtained were clearly indicating that the UF resins
Thus, it is clearly noticeable when comparing had not condensed to extremely large molecular
normal UFs and additionally branched UF resin that weight resins. The reaction temperature was then
for n molecules of urea used two or three molecules decreased to 55-60°C and the pH increased to 8-
of resorcinol are used in the first case, whereas only 8.5 with 22% sodium hydroxide solution. A certain
one or two molecules of resorcinol for n molecules amount of resorcinol was then added to the UF resin.
of urea are used in the second case. It has already The resorcinol added was a 50% resorcinol : 50%
been shown3 that branched PRF adhesives of low solvent (30%methanol :70% water) solution whose
resorcinol content, still presenting good perfor- -
pH had been preadjusted to 9 with a 22% sodium
mance, can be prepared. For these, the lowest per-
centage of resorcinol on liquid resin that appeared
still to give a resin consistently satisfying the re-
allowed to stir a t -
hydroxide solution. The reaction mixture was then
55-60°C for the appropriate
time. The reaction time varied according to the U :
quirements of international specifications 6,7 for F ratio used. Throughout the reaction time, the pH
close-contact, cold-setting adhesives for wood with was kept between 9 and 10 with the addition of 22%
still acceptable pot life, was 10.6%.3This is signifi- sodium hydroxide solution. While the resin was still
 UREA-RESORCINOL-FORMALDEHYDE ADHESIVES 2 137

warm, a certain amount of wattle tannin extract was Analysis of Resins

added to the resin. On cooling, the resulting resins
Resins were analyzed using IR and I3C-NMR ana-
were diluted to a usable viscosity with 30% metha-
lytical techniques. IR spectra were recorded on Jasko
nol : 70% water solvent.
FT/IR5000 and PYE UNICAM PU 9512 infrared
spectrometers. 13C-NMRspectra were recorded on
Method of Preparation of a Branched URF Resin a Bruker AC-200 spectrometer (200.13 MHz) with
the samples dissolved in D20.
A UF resin of U : F ratio 1 : 0.5 was prepared as Though UF resins are characteristically white in
described above. The appropriate amount of color, on addition of the resorcinol, they become
brancher was added as a 50% brancher :50% solvent brown. The mimosa polyflavonoid tannin extract
(30% methanol : 70% water) by mass solution. The gives them a darker brown tint. However, on addi-
reaction mixture was allowed to react under contin- tion of a hardener, both the resorcinol- and wattle
uous stirring a t 55-60°C €or a maximum period of tannin-extract branched URF resins become dark
4 h or until the reaction mixture thickened exces- brown after curing and cannot be distinguished from
sively. After addition of the brancher, the pH of the traditional PRF resins.
reaction mixture was adjusted to - 9 and main-
tained at this pH throughout the branching reaction
period with 22% sodium hydroxide solution. The RESULTS AND DISCUSSION
appropriate amount of terminal 50% resorcinol :
50% solvent (30% methanol : 70% water) solution
was added whose pH had been preadjusted to
with 22% sodium hydroxide solution. A reaction pe-
-
9
The preparation of URF resins has been reported
several times, although these resins have not been
used very much due to the fear that their durability
riod of 1or 2 h was allowed at 55-60°C. The pH of might not be as good as that of their phenolic coun-
the reaction mixture was maintained between 9 and terparts. Traditional URF resins of acceptable per-
10. While the resin was still warm, 15% wattle tan- formance (cf. results of resins 1 and 2 in Table I )
nin extract by mass of undiluted resin was added to operate a t 55-58% resin solids content. This is
thicken the resin. On cooling, the resin was diluted slightly higher than for linear PRF resins. The vis-
to a usable viscosity with 30% methanol :70% water cosity of these resins at this percentage solids con-
solution. tent is low and they cannot be further diluted with-

Table I Preparation Parameters and Results of Linear URF Resins

Resin URF No.

1 2 3 4

U : F ratio 1 : 0.5 1: 0.5 1: 0.5 1: 0.5

% Resorcinol in undiluted liquid resin 25.0 22.5 20.0 17.0
Reaction time at 60°C (min) 360 360 360 240
% Solids content 57.1 55.9 47.6 55.1
% Resorcinol in total liquid resin 17.4 16.0 12.0 12.4
Viscosity (cps) 155 395 225 500
Pot life (min) 76 80 90 75
Shelf life (days) > 40 > 40 > 40 5
% Wattle by mass in undiluted liquid resin 15 15 15 5
Strength (kN)
Dry 2.57 2.50 2.50 2.65
Soak 2.20 2.20 1.70 1.00
Boil 2.20 2.20 1.60 0.90
Wood failure (%)
Dry 70 43 27 94
Soak 75 75 42 3
Boil 87 83 43 2
 rn
Q

8E=
Table I1 Preparation Parameters and Results of Branched URF Resins

Resin URF No.

!U2
5 6 7 8 9 10 11 12 13 14 14W 15 15W 16 16W i
u
U : F ratio 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5
Liquid resin 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4 176.4
Branching resorcinol (mol) 0.041 0.082 0.160 0.250 0.330 0.029 0.072 0.140 0.210 0.082 0.045 0.160 0.087 0.029 0.016
Branching time at 60°C 240 240 40 145 130 60 60 60 60 Res. Tan. Res. Tan. Res. Tan.
(min) 240 240 240 40 25 240

Terminal resorcinol (mol) 0.074 0.74 0.74 0.74 0.74 0.60 0.60 0.60 0.60 0.74 0.74 0.74 0.74 0.60 0.60
Terminal resorcinol time at 120 120 120 120 120 60 60 60 60 120 120 120 120 60 60
6OoC(min)
% Resorcinol in liquid resin 16.6 14.9 18.3 19.6 19.8 15.5 15.5 16.6 17.8 14.9 13.4 18.3 10.7 15.5 12.7
% Solids content 56.7 48.8 56.1 56.8 54.5 59.0 56.2 55.9 56.2 48.8 56.5 56.1 51.9 59.0 57.4
PH 9.22 9.73 8.86 8.80 8.89 8.95 9.20 9.33 9.30 9.73 8.88 8.86 8.79 8.95 8.96
Viscosity (cps) 116 155 220 250 155 135 155 240 186 155 247 220 230 135 160
Pot life (min) 75 75 87 80 65 76 87 61 70 75 75 87 66 76 77
Shelf life (days) > 40 > 40 > 40 > 40 > 40 > 40 > 40 > 40 > 40 > 40 z 40 > 40 > 40 > 40 > 40
% Wattle in total liquid resin 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15
Strength (kN)
Dry 2.60 2.60 3.15 2.52 2.94 2.90 2.73 2.72 2.84 2.60 2.60 3.15 2.00 2.90 2.55
Soak 2.10 1.75 2.10 2.15 2.42 1.96 1.70 2.25 1.68 1.75 1.90 2.10 0.85 1.96 1.80
Boil 2.02 1.74 2.40 2.40 2.57 1.70 1.73 1.93 1.97 1.74 2.10 2.40 1.43 1.70 1.60
Wood failure (%)
Dry 36 70 35 16 8 46 100 94 60 70 78 35 73 46 100
Soak 76 75 65 55 53 70 90 67 22 75 77 65 13 70 77
Boil 88 95 50 46 43 70 90 70 53 95 99 50 25 70 99
 UREA-RESORCINOL-FORMALDEHYDE ADHESIVES 2139

out deterioration in their performance. Resorcinol- show that there is an optimum level of resorcinol
branched URF adhesives operate in a much wider for each type of URF resin.
percentage solids content range. For instance, resin URF resins to which resorcinol as an additional
URF 6 and resin URF 11( cf. Table 11)display good branching unit was added and URFs without an ad-
performance and operate at 48.8 and 59%resin solids ditional branching unit had pot lives longer than
content, respectively. In general, resorcinol-induced did urea-branched PRF resins3 (Fig. 1) . URF resins
branching in URF resins does not appear to have a in which mimosa polyflavonoid tannin was the ad-
significant effect on the viscosity of the resins. The ditional branching unit gave, instead, a much shorter
resins produced have a viscosity comparable to that pot life (Fig. 1). Such URF resins presented the
of “linear” URF resins. Mimosa tannin-branched same sensitivity to a pH decrease as did their PRF
URF resins are, however, slightly more viscous than counterparts Again, only acetylsalycilic acid3 was
are the corresponding resorcinol-branched URF capable of correcting the resin pH downward with
resins. This increase in viscosity could be attributed consequent lengthening of pot-life. Any other acid,
to the larger molecular weight of the wattle tannin in solution or solid, caused localized precipitation of
macromolecules that have condensed onto the URF the resin.
resin. These resins also operate in a much narrower Another problem encountered in URF resins was
range of percentage resin solids content ( 56-58% ) their stability (shelf life). Whereas traditional and
(cf. results of URF 14W and URF 16W in Ta- branched PRF resins are stable and display long
ble 11). shelf lives, this was a problem in the production of
The lowest percentages resorcinol by mass on to- URF resins. A decrease in the formaldehyde /urea
tal liquid resin that gave acceptable performance in molar ratio resulted in a more stable resin with a
“traditional” URF and resorcinol-“branched” and much longer shelf life. The most stable U : F ratio
tannin-“branched” URF resins are 16, 13.4, and was found to be 1 : 0.5 (cf. results in Table 111).
12.7%, respectively (cf. results of URF 1.6 and 16 As urea has three reactive sites, and also as a
W in Table 11).The results in Table I1 also clearly result of its reactivity, most UF resins are shown to

90

0 troditional URF resin

h
a 70 A resorcinol-bronched URF resin
m m wattle-branched URF resin
L
1

Y*i
m
3I
L
0
a

Figure 1 Resins pot lives as a function of resin glue-mix pH: ( 0 )traditional URF resin;
( A ) Resorcinol-branchedURF resin; (m)wattle tannin extract-branchedURF resin; (+)
urea-branched PRF resin.
2140 SCOPELITIS AND PIZZI

be naturally branched or potentially branched. This

d
m W is quite different to traditional PF resins that appear
to be mostly linear.” It must be stressed that what
m are referred to as traditional URF resins in this ar-
a
d
ticle are not necessarily linear: They are merely URF
resins in which additional induced branching has
m
0.1
*
m not been introduced by the addition of another
branching unit during its preparation. Also, it was
ri
m
not possible to distinguish between mono-, di-, and
0
m trisubstitution on the resorcinolic nucleus by a
0
change in the chemical shift of the resorcinolic C’s
m IS W in 13C-NMR spectroscopy. Thus, 13C-NMR spec-
0
d
d A d z % troscopy was not used to identify tridimensional
branching or linear lengthening on the resorcinolic
m
N
0
N
nuclei, but rather just on the urea molecules. The
peak that was used as an identification whether tri-
dimensional branching occurred is a t 6 = 55 ppm:
m
N
m
t-

t-
N
m
10
On this basis, URF resins to which resorcinol has
W
N not been added as a branching unit show an increase
%
of branching on the UF backbone of the resin when
total terminal resorcinol is increased. Conversely,
% m URFs in which additional branching has been in-
d
duced by introduction in the midreaction of a small
amount of resorcinol branching units show a de-
d
N
0 crease of branching with a resorcinol decrease. This
m
can be observed by the absence, or presence, and
m intensity, of the
N 3
N

I
N
- HZC-N - CHZ-
m
0

band at 6 = 55-56 pprn in I3C-NMR. Thus, in the

rl
noninduced-branching URF resins 1 and 2 (Table
N
r-
N I ) of, respectively, 25 and 22.5% terminal resorcinol
content, this I3C-NMRband is, respectively, absent
0
N and present (but small). This result could be inter-
t-
d preted as that by decreasing the percentage of re-
sorcinol more free formaldehyde and methylol
m
rl
groups are still available for reaction, increasing the
*
m
probability of UF branching.
m
In induced-branching, the reverse applies. Thus,
rl
10
m
induced-branching URF resins 35 and 38 (Table IV)
of, respectively, 25 and 22.5% total terminal resor-
r-
4
cinol content have the 13C-NMRband a t 6 = 55-56
10
N ppm present and absent, respectively. Applied re-
sults indicated that as the amount of resorcinol
brancher increased in branched URF resins the per-
centage wood failure results decreased (Table 11).
By examining the I3C-NMR spectra, it was pos-
sible to note differences that can be used to under-
stand why the applied results are such. When only
 UREA-RESORCINOL-FORMALDEHYDE ADHESIVES 2 14 1

Table IV Effect of Methanol in Branched URF Resins

Resin URF No.

35 36 37 38 39 40
~

U : F ratio 1: 0.5 1: 0.5 1: 0.5 1: 0.5 1: 0.5 1: 0.5

% Resorcinol in undiluted liquid resin 25.0 25.0 25.0 22.5 22.5 22.5
% Resorcinol in total liquid resin 17.4 17.1 16.9 16.0 16.8 15.8
Reaction time at 60°C (min) 360 360 360 360 360 360
Methanol : water ratio in solvent 30 : 70 15 : 85 45 : 55 30 : 70 15 : 85 45 : 55
% Solids content 57.1 56.2 55.3 55.9 58.8 55.0
PH 10.42 8.81 9.31 9.65 8.43 9.31
Viscosity (cps) 155 160 105 395 160 100
Pot life (min) 76 83 79 80 70 62
Shelf life (days) > 40 > 40 > 40 > 40 > 40 > 40
% Wattle by mass in undiluted liquid resin 15 15 15 15 15 15
Strength (kN)
Dry 2.57 2.62 2.60 2.50 2.23 2.61
Soak 2.20 1.51 1.24 2.20 1.03 0.89
Boil 2.20 1.66 1.18 2.20 1.60 1.27
Wood failure (%)
Dry 70 68 100 43 29 77
Soak 75 32 15 75 28 25
Boil 87 44 30 83 23 20

2.9 X lo-, mol to 7.2 X rnol of resorcinol By substituting the resorcinol brancher with mi-
brancher is used (URF 10 and 11, Table 11), tridi- mosa tannin extract, it was hoped that the percent-
mensional branching on the urea molecule appears age resorcinol by mass on total liquid resin would
to occur as the peak at 6 = 55.90-56.04 ppm is clearly decrease. It was, however, feared that the percentage
evident (Figs. 2 and 3 ) . Resin URF 11 (with 7.2 wood failure results would be lower because con-
X mol of resorcinol brancher experiences the densation of the macromolecule on the URF resin
greatest tridimensional branching on the urea mol- would imply an increase in the molecular weight of
ecule according to the 13C-NMR spectrum, yet it the resin.
also exhibits the highest percentages wood failure. The applied results showed that if the amount of
When the amount of resorcinol brancher is further resorcinol brancher did not prove to be excessive in
increased, then it appears from the 13C-NMRspectra the branched URF resin, optimum results were also
that tridimensional branching on the urea molecule achieved when an equivalent amount of tannin ex-
is eliminated as the -N- (CH,), peak a t 6 = 56 tract was used as a brancher. (cf. results in Table
ppm disappears. The amount of branching resorcinol 11).The surprising result was that the introduction
may be sufficient enough to allow the macromolecule of the tannin extract as a brancher decreased the
to lengthen linearly. This linear lengthening appears percentage of resorcinol by mass on total liquid resin
to decrease the percentages wood failure, probably without deteriorating the percentage wood failure
as a consequence of the decreased ability of the results. A possible explanation of this is the differ-
polymer to wet the substrate. The optimum amount ence in the molecular geometry created when
of brancher appears to be 7.2 X lop2 mol to 8.2 branching occurs on a small resorcinol molecule
X mol of resorcinol brancher for approximately compared with that created when it occurs on a large
15% of terminal resorcinol used by mass of liquid tannin macromolecule. When resorcinol is used as
resin. Though, as expected, linear lengthening is the a branching molecule, it has a much more rigid
favorite mechanism for decreasing the resorcinol structure than when tannin extract is used. The
required, it appears that in branched URF resins molecular geometry of branching in the latter is not
the macromolecules produced are too large. This as densely packed. The reactive sites on the wattle
causes poor specific adhesion and is manifested in tannin macromolecule are also more widely spaced
poor wood failure results. than those on resorcinol. The more widely spaced
2142 SCOPELITIS AND PIZZI

the optimum conditions to produce a branched URF

with good performance are

( i ) the use of 4.5 X mol of polyflavonoid

tannin extract as a branching unit for 176.4 g
of 1: 0.5 UF liquid resin and a 4 h branching
period at 60°C; and
(ii) 7.4 x lo-' moles of terminal resorcinol for
176.4 g of 1 : 0.5 UF liquid resin and a reaction
time of 2 h at 60°C.

The amount of methanol used in the solvent also

affected the performance of both traditional and
branched URF resins. In traditional URF resins, it
was observed that if the amount of methanol in the
solvent used was increased the percentage wood
failure in the dry test improved, but the water-re-
sistant properties of the resin deteriorated (Ta-
ble IV) .

f
t Methanol is known to be a retarder in the UF
'3, resin synthesis. Methanol retards the process
30
t whereby paraformaldehyde release formaldehyde

Figure 2 13C-NMR spectrum region 44-61 ppm of resin

URF 10; note that the 56.04 ppm band indicatesbranching.

and more flexible molecular geometry introduced

when polyflavonoid tannin is used as a brancher may
be allowing better adhesion of the resin to the wood,
resulting in higher percentages wood failure. Ordi-
narily higher molecular weights cause poorer wet-
ting-hence, poorer wood failure, but in this case, *
it appears that the molecular geometry caused by t'.
0
branching molecules has a significant effect on the In
ability of the adhesive to wet the substrate. 13C-NMR
spectroscopy (16W, Table I1 and Fig. 4 ) also indi-
cates that when tannin extract is used tridimen- 0
c",
sional branching on the urea molecule is even more In
In
encouraged than when resorcinol is used. The peak
at 6 = 55 ppm

is even more intense in these spectra. This further

supports the hypothesis that branching on the poly-
flavonoid tannin occurs more readily due to the bet-
L
ter steric availability of the reactive sites. Figure 3 13C-NMRspectrum region 44-61 ppm of resin
Both applied and analytical results showed that URF 11.
 UREA-RESORCINOL-FORMALDEHYDE ADHESIVES 2143

I I 1 L I
180 170 160 150 140 130 120 110 90 80 70 60 50 40 30 20
Figure 4 I3C-NMR spectrum of resin URF 16W wattle (mimosa) tannin extract-
branched.

that can then react with urea to form methylol

groups. As fewer reactive methylol groups results,
shorter UF polymer chains are formed. Specific
adhesion to lignocellulosic substrates of lower mo-
lecular weight resins is greater". For this reason, it
was believed that higher percentages wood failure
in the dry test would result, as indeed they did (cf.
results in Table IV) . However, it has been reportedI2
that methanol content makes a considerable differ-
ence in the water resistance of UF resins. This effect
is due to the formation of methylated UF resins,
which, on curing, remain unchanged and because of
CO
their easy solubility in water destroy the water re-
sistance of the cured, bonded joint. The same effect 2
v)
is being observed in traditional URF resins in which
45% by mass of methanol was used in the solvent
mixture (Table I V ) .
I3C-NMR spectroscopy indicated a possible ex-
planation for the above phenomenon. Resins URF
1 and 35 (Tables I and IV) are traditional URF res-
ins with 17.4% of resorcinol by mass on total liquid
!A
resin. Resin URF 37 is its corresponding resin but L
with 45% of the solvent being methanol, whereas in
the former, only 30% was methanol. The applied Figure 5 13C-NMR spectrum regions 41-61 ppm and
results indicated an increase in percentages wood 150-180 ppm of resins URF 1 and URF 35. Note absence
failure in the dry test but a weakening of the water of 55 ppm band.
2144 SCOPELITIS AND PIZZI

resistance. If one examines the 13C-NMRspectra of

these resins (Fig. 5), one notes that with an increase
in percentage methanol used in the solvent there is
an increase in the amount of monomeric urea and
methylenebisurea present. This is deduced by an in-
crease in intensity of the bands a t 6 = 164.10 ppm
( C =0 of urea) and a t 6 = 161.99 ppm ( C =O peak
of methylenebisurea) in the spectrum of URF 37
(Fig. 6 ) relative to the intensity of the same peaks
in the spectrum of URF 1and 35 (Fig. 5). The pres-
ence of this increased amount of monomeric urea
and methylenebisurea may be causing weakening in
the water-resistant properties of URF 37. Another
possible explanation is that in the presence of a
larger percentage methanol the decrease of reactivity
of the formaldehyde causes a slower degree of po-
lymerization of the resin. Though this may be caus-
ing stronger adhesion in the dry test due to better
penetration of the resin into the wood, it also may
be simultaneously weakening the cohesive forces
of the resin. This may be to the extent that if the
cured resin is exposed to severe water attack con-
ditions the resin exhibits impaired water resistance.
The percentage methanol in the solvent also affects
the results of branched URF resins. For instance,
Figure 6 13C-NMR spectrum region 149-173 ppm of resin URF 6 (Table 11) is a branched URF resin
resin URF 37. with 14.9% of resorcinol by mass on total liquid

Table V Effect of Methanol in Branched URF Resins

Resin URF No.

41 42 43 44 45 46

U : F ratio 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5 1 : 0.5

Mass of liquid resin (g) 176.4 176.0 176.4 176.4 176.4 176.4
Methanol : water ratio in solvent 30 : 70 45 : 55 30 : 70 45 : 55 30 : 70 45 : 55
Branching resorcinol (mol) 0.041 0.041 0.082 0.082 0.160 0.160
Branching time at 60°C (min) 240 240 240 240 40 50
Terminal resorcinol (mol) 0.74 0.74 0.74 0.74 0.74 0.74
Terminal resorcinol time at 60°C (min) 120 120 120 120 120 120
% Resorcinol in liquid resin 16.6 16.3 14.9 17.4 18.3 17.7
% Solids content 56.7 55.3 48.8 57.2 56.1 54.2
PH 9.22 9.16 9.73 9.20 8.86 8.90
Viscosity (cps) 116 150 155 195 220 95
Pot life (min) 75 75 75 70 87 77
Shelf life (days) > 40 > 40 > 40 > 40 > 40 > 40
% Wattle by mass in undiluted liquid resin 15 15 15 15 15 15
Strength (kN)
Dry 2.60 2.80 2.60 2.50 3.15 2.43
Soak 2.10 2.07 1.75 1.86 2.10 0.88
Boil 2.02 1.75 1.74 2.10 2.40 1.39
Wood failure (%)
Dry 36 100 70 100 35 5
Soak 76 60 75 84 65 26
Boil 88 86 95 92 50 24
 UREA-RESORCINOL-FORMALDEHYDE ADHESIVES 2 145

resin. URF 44 is its corresponding resin except that

once again 45% of the solvent is methanol instead
of only 30%.Applied results showed an increase in
percentages wood failure on dry test, but no dete-
rioration on the soak and boil tests (Tables I1 and
V, resin 6 vs. 4 4 ) . The 13C-NMR spectrum of resin
URF 44 does show an increase in the amount of
urea compared to that in resin URF 6 (Figs. 7 and
8). The relative intensity of the peak a t 6 = 164.10
ppm ( C =0 of urea) in relation to the rest of the
spectrum is much more intense in the spectrum of
URF 44 than in URF 6. However, there is no dif-
ference in relative intensity between the C =0 peak
of methylenebisurea ( 6 = 161.9 ppm) in the two
spectra. This implies that there does not appear to
be an increase in the amount of methylenebisurea
as in URF 37 (Fig. 6 ) (the traditional URF resin
with 45% methanol in the solvent). From the 13C-
NMR results, one can only conclude that too large
a percentage of methylenebisurea in the resin ap-
pears to have a weakening effect on the water-re-
sistance properties of the resin. This is perhaps the

Figure 8 I3C-NMR spectrum region 152-171 ppm of

resin URF 44.

first clear analytical indication of the water lability

of the -NH -CH2-NH -bond.

CONCLUSION

In conclusion, effects that appear to influence the

results of branched URF resins include the

( i ) amount of brancher, with amounts of resor-

cinol brancher between 0.07 and 0.08 mol giv-
ing the best balanced performance;
(ii) type of brancher, with polyflavonoid tannins
(0.045 mol) allowing greater decrease of re-
sorcinol content;
(iii) temperature of reaction, with lower reaction
temperatures (50-60°C) giving much better
1 , results;
(iv) amount of final resorcinol, with resorcinol
Figure 7 13C-NMR spectrum region 145-175 ppm of contents as low as 12-13% still giving resins
resin URF 6. of acceptable performance;
2146 SCOPELITIS AND PIZZI

( v ) amount of methanol used in solvent, with 4. J. R. Ebdon and P. E. Heaton, Polymer, 18, 971
increasing methanol increasing percentage (1977).
wood failure but progressively impairing water 5. R. M. Rammon, W. E. Johns, J. Magnuson, and
resistance of the cured linear resins, but not A. K. Dunker, J.Adhes., 19,115 (1986).
of the cured branched resins; and 6. British Standard BS 1204,1965, Parts 1and 2, Spec-
ification for Synthetic Resin Adhesives.
(vi) physical properties such as pH and viscosity. 7. South African Bureau of Standards SABS 1349-1981.
Phenolic and Aminoplastic Resin Adhesives for the
Though these effects also influenced the results of Laminating and Fingerjointing of Timber and for
branched PRF resin^,^ it appears that the range in Furniture Joinery.
which these effects overlap to produce acceptable 8. A. Pizzi, in Wood Adhesives Chemistry and Technol-
results in branched URF resins is much narrower ogy, A. Pizzi, Ed., Marcel Dekker, New York, 1983,
than in branched PRF resins. Vol. 1, Chap. 4.
9. E. A. Blommers and R. H. Moult, U.S. Pat. 4,032,515
( 1987).
10. N. J. L. Megson, Phenolic Resin Chemistry, Butter-
REFERENCES worths, London, 1958.
11. A. Pizzi, J.Adhes. Sci. Technol. 4,573 ( 1990); 4,589
1. J. M. Dinwoodie, in Wood Adhesives Chemistry and ( 1990); Holzforschung Holzverwertung, 43 ( 3 ) , 63
Technology, A. Pizzi, Ed., Marcel Dekker, New York, ( 1991).
1983, Vol. 1, Chap. 1. 12. M. Saula, in Handbook of Adhesives, 2nd ed., I. Skeist,
2. A. Pizzi, in Wood Adhesives Chemistry and Technol- Ed., Van Nostrand Reinhold, New York, Chap. 25,
ogy, A. Pizzi, Ed., Marcel Dekker, New York, 1983, 1976.
Vol. 2, Chap. 7.
3. E. Scopelitis and A. Pizzi, J. Appl. Polym. Sci., to Received April 4, 1992
appear. Accepted September 30, 1992