DETERMINATION OF PARAFORMALDEHYDE

REACTIVITY AND ITS RELATIONSHIP TO

PRF RESIN GELATION
Paul R.Steiner
Associate Professor
Department of Harvesting and Wood Science, Faculty of Forestry
University of British Columbia
Vancouver, BC V6T 124 CANADA
(Received October 1990)

ABSTRACT
A titration procedure is described for determining paraformaldehyde reactivity in aqueous solution
using common chemical reagents of Na2S0, and NaH,PO,. This reactivity is a measure of the mono-
meric formaldehyde released by the depolymerization of paraformaldehyde. Under certain experi-
mental conditions, a curvelinear relationship was established between paraformaldehyde titration
values and gelation times in a PFR resin. By utilizing a stepwise version of this titration method, the
total reactivity distribution within paraformaldehyde samples could be determined and compared.
Keywords: Paraformaldehyde, PRF, titration, gelation reactivity.

INTRODUCTION tivity because this test condition closely par-

Paraformaldehyde is a mixture of polymeth- allels conditions used during bond formation.
ylene glycols containing 8 to 100 formaldehyde However, gelation evaluations can be time-
units per molecule (Walker 1964a). Commer- consuming and require the use of a test resin
cial grade powders are commonly used in the system that exhibits a characteristic, time-de-
hardener component of phenol-resorcinol- pendent curing property. This property is de-
formaldehyde (PRF) adhesives. The reactivity termined primarily by the manner in which
of paraformaldehyde can vary considerably the resin was synthesized. A more universal
depending on how it was manufactured and indication of hardener reactivity should result
the age of the material. When hardener and from a direct measure of paraformaldehyde
resin components are mixed, the paraformal- decomposition to formaldehyde in aqueous
dehyde breaks down to release monomeric solution.
formaldehyde, which rapidly reacts with the Since cleavage of the formaldehyde polymer
PRF polymer to first cause gelation and ulti- is believed to occur primarily at the end of the
mately resin cure. The amount of paraformal- paraformaldehyde polymer chain, depolymer-
dehyde present and its rate of decomposition ization will depend upon the number and ac-
have a major effect on PRF cure speed. In cessibility of end groups. Consequently, para-
addition, recent concerns about formaldehyde formaldehyde materials of similar molecular
emissions make control of paraformaldehyde weight may release formaldehyde at signifi-
decomposition an important parameter during cantly different rates, depending upon whether
manufacture and use of glued-wood-products. the polymer chain end groups are exposed out-
The room temperature curing capabilities of side or inside the polymer chain bundles. The
PRF adhesives make them the adhesive of rate of formaldehyde release should also be
choice for bonding timber laminates of large dependent upon molecular weight, particle size,
dimension (Chow 1977). With these adhe- temperature, and pH.
sives, gelation time at moderate temperatures This paper reports on a procedure for rap-
provides a reliable indicator of system reac- idly determining formaldehyde release into
Woodand Fiber Scrence, 24(1), 1992, pp. 73-78
O 1992 by the Society of Wood Science and Technology
74 WOOD AND FIBER SCIENCE, JANUARY 1992, V. 24(1)

aqueous solution and relates this data to gela- Gelatin time determination
tion time measurements in a PRF resin sys- Three paraformaldehyde samples, A, B (both
tem. powders), and C (flakes) were obtained as re-
agent grade material from BDH Chemicals,
EXPERIMENTAL
Fisher Scientific and BDH Chemicals, respec-
The pH change in all experiments was mon- tively. Specifications for this grade of material
itored on an Accumet model 750 pH meter indicated a minimum of 97% paraformalde-
with attached chart recorder. Measurements hyde content. A fourth sample, D, was a hard-
were performed at ambient temperature (22 zk ener mixture (FM 282) supplied by Borden
2 C) in an unthermostated beaker. Reaction Chemicals and used to cure PRF resins. Anal-
mixtures were magnetically stirred. Reagent ysis of total formaldehyde, by the standard
grade sodium sulfite (Na,SO,) (Fisher Scien- Na2S0, method (Walker 1964b) showed this
tific) and reagent grade sodium dihydrogen- hardener, D, to contain 35% formaldehyde,
phosphate (NaH,PO,) (Fisher Scientific) and with the rest being walnut shell filler. Parafor-
distilled water were used throughout. The PRF maldehyde with higher levels of polymeriza-
resin used was LT-75, a 50% solids solution tion were prepared by heating powder A in a
supplied by Borden Chemicals. closed container at 92 C. Samples 2A and 3A
were produced by heating for 9 and 16 days,
Titration of a formaldehyde solution respectively. During this heating period, the
against aqueous sodium sulfite samples were monitored daily, and the reac-
A 0.10 M HCHO solution was prepared by tivity was found to cycle between low and high
diluting 8.1 ml of a 37% formalin solution titration values in an unpredictable manner.
(containing 0.2% methanol) to one liter. This Addition of a small quantity of CaC1, into the
was titrated against 25 ml of 1.0 M Na2S0, or powder during heating helped to stabilize this
a 25 ml solution containing 1.0 M Na,SO, and fluctuation. The CaC1, likely acts as a dessicant
0.001 M of NaH,P04. and catalyst for this reaction. Samples 4A, 5A,
6A, and 7A were produced by heating a mix-
Standard titration procedures ture of sample A with 0.1% CaC1, at 67 C for
0.5, 1, 3 and 6 h, respectively.
A standard Na,SO,/NaH,PO, solution was Gelation times were determined by quickly
prepared by dissolving 63.02 g of anhydrous mixing 1.58 g of paraformaldehyde with 25 g
Na,SO, and 13.8 g of NaH,PO, H,O in dis- of LT 75 and placing the mixture in a 50-ml
tilled water to make a final volume of 1,000 test tube that was held in a water bath at either
ml. A 50-ml portion of the sulfite/phosphate 30 or 45 C. The measurement was accom-
solution was pipetted into a 400-ml beaker. plished by manually moving a stirring rod
The stirred solution was diluted with 175 ml through the mixture using an up and down
of distilled water and 25 ml of methanol. The motion (about 5 cycles per minute). The period
solution was vigorously stirred and 0.500 g of between when the test was initiated and when
paraformaldehyde quickly added while a timer the rod could no longer be moved in the mix-
and recorder were simultaneously started. It is ture was the gelation time. Duplicate samples
important that the paraformaldehyde be were run for each mixture. When the FM 282
quickly dispersed into the reaction medium. hardener was evaluated, 4.5 g of this material
The presence of methanol was found to elim- (equivalent to 1.58 g ofparaformaldehyde) was
inate the formation of lumps during the dis- used for testing gelation.
persion process. The reaction pH was moni-
tored, and the timer was stopped after a rapid Stepwise titration of paraformaldehyde
pH change to 1 1 (endpoint) was obtained. Two A 7-ml rapid filling pipette with a reservoir
replicates were titrated for each sample. containing a 0.48 M solution of NaH2P04was
 Ste~ner-PARAFORMALDEHYDE REACTIVITY A N D PRF GELATION 75

I ; ! : : I ; I I : ! I I ; : I ! : I I
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 18 17 I8 0 10 20 30 40 50 80 70 80
volume of 0.1M HCHO (ml) Tim (sec)
FIG.1. Titration of 0.10 M HCHO solution against A, FIG.2. Change in pH with time resulting from addition
25 ml of l .O M Na2S0, and B, 25 ml of both 1.OM Na2S0, of A, 0.500 g paraformaldehyde (BDH type) or B, 0.500
and 0.001 M NaHZPO,. g paraformaldehyde (heated 9 days at 92 C) to a 50-ml
aqueous solution of 0.025 M Na2S0, and 0.005 M
NaH2P0,.
positioned above a 400-ml beaker containing
25 ml methanol and 225 ml of 0.10 M Na,SO,.
A 7-ml volume was added from the pipette. Na2S03 + HCHO + H,O
- HOCH,SO,Na
- + NaOH
The pH electrodes were immersed, and the
mixture was vigorously stirred. A 0.5-g sample with stoichiometric generation of NaOH re-
of paraformaldehyde was added at once, and sulting in a pH increase during the reaction.
at the same time, the chart recorder was start- Titration of hydroxide by acid allows deter-
ed. Reaction was continued until the pH mination of the original formaldehyde con-
reached a value of 10.5, whereupon another tent. To utilize this technique for estimating
7-ml volume of NaH,PO, solution was rapidly paraformaldehyde reactivity, the presence of
added from the pipette. This caused the pH to high alkalinity in the solution must be con-
return to about 8, after which the reaction con- trolled since the rate of paraformaldehyde de-
tinued until the pH again reached 10.5 and composition increases rapidly in this environ-
another 7-ml portion of NaH,P04 solution was ment. By incorporating an appropriate weak
added. A total of four 7-ml portions were add- acid into the mixture, the titration can be per-
ed for each paraformaldehyde tested. Since each formed in situ and the time taken to consume
7-ml addition represents 20% ofthe total para- the acid is related to the breakdown of para-
formaldehyde present, up to 80% of the poly- formaldehyde.
mer reactivity was measured. The typical behavior of pH during the titra-
RESULTS AND DISCUSSION
tion of aqueous formaldehyde solution against
excess sodium sulfite is shown in curve A of
Reaction of formaldehyde and Fig. 1. A limiting pH of about 1 1.4 is reached
paraformaldehyde well before the stoichiometric endpoint of 250
In reactions where the only source of form- ml of 0.1 M HCHO for 25 ml of 10 M Na,SO,.
aldehyde is paraformaldehyde, the reaction rate By incorporation of the acid salt NaH2P04to
will be limited by the rate at which formal- consume the generated base, curve B of Fig. 1
dehyde is generated from its polymeric form. is obtained.
The sodium sulfite method has long been used These reactions proceed rapidly when mo-
as a quantitative analytical procedure for de- nomeric formaldehyde is present in aqueous
termining free formaldehyde in aqueous so- solution. However, if the formaldehyde is some
lution. This process is described by the equa- condensed polymeric form such as paraform-
tion: aldehyde, its rate of reaction may be found by
76 WOOD A N D FIBER SCIENCE, JANUARY 1992, V. 24(1)

TABLE I. Eflect ofparticlesize on titration times ofpara-

formaldehyde.
Tttrat~ontime (sec)
Sample
Mesh slrc range I 2

0
I
0 M ) 1 M ) 1 W Z W 2 W 9 0 0 9 M ) ~
TITRATION TIME (sec)

FIG. 3. Gelation time vs. titration time for parafor-

determining the time needed for a stoichio- maldehyde samples.
metric excess of this polymer to consume a
given amount of dihydrogen phosphate in the Particle size eflects
presence of sulfite. To ensure that a large pH Since this reaction involves a solid-liquid
change occurs at some stage, a deficiency of phase interaction, the accessibility of reagents
acid relative to both paraformaldehyde and to reactive sites on the polymer is expected to
sulfite is necessary. The experimental results be an important factor influencing titration
for two paraformaldehyde samples indicating times. Consequently, the mixture must be well
the change in pH with time are shown in Fig. stirred during reaction. In addition, the effect
2 . The combination of sulfite and dihydrogen of particle surface area on accessibility must
phosphate acts as a buffer until the latter is be considered. Table 1 shows the influence of
consumed, whereupon the pH rises steeply. paraformaldehyde particle mesh size on titra-
The generation of hydroxide will continue un- tion times. Reactivity increases by an order of
til one of the reactants, either sulfite or para- magnitude as accessible surface is increased by
formaldehyde, is depleted. As long as the sul- converting paraflake from the largest to small-
fite concentration is in excess, titration time est particle size. A standardized particle size
will depend primarily on the quantity of para- range of >80 to 200 mesh was chosen for this
formaldehyde present and the reactivity of the study.
polymeric form.
Correlation of titration times with
gelation times
2.
TABLE Titration times and gelation times for parqfor- After some initial experimentation with a
maldeh~desamples.
range of reactant concentrations, a standard
Para-
farmalde- Gclat~ontime* ( m ~ n )
condition was established that provided some
hyde
sample
Tltration time*
(set) 30 C 45 C
spread in titration results with various para-
formaldehyde samples. Table 2 lists titration
times and gelation times for three commercial
paraformaldehyde samples (A, B, and C ) and
heated samples made from them (2A-7A). Fig-
ure 3 indicates the relationship found between
room temperature titration times and gelation
times at 45 C. The graph displays two distinct
regions. Between gelation times of 35 and 85
minutes, an almost linear relationship exists
*Duplicate tests carried out on each sample. between titration and gelation times (r2= 0.96
 Steiner-PARAFORMALDEHYDE REACTIVITY AND PRF GELATION 77

" , ' [ I TIME (we)

11.5 1E4, I

"":~"':"":"":"":"":
0 10 20 30 40 50 00

Time (sec) 0 20 40 80 80

FIG. 4. Stepwise titration to constant pH for FM282 PARAFORYALDEHYDE PORTION REACTED (X)
hardener. Arrows represent times where additional FIG. 5. Relationship between titration times for suc-
NaH,PO, solution, molar equivalent to 20°/0 HCHO, is cessive stages in the stepwise titration sequence and the
added. proportion of paraformaldehyde reacted. Samples 1A, 2A,
and 3A are, respectively, a paraformaldehyde powder
heated for 0, 9, and 16 days at 92 C. Samples C and D
from linear regression analysis in this region). are, respectively, a flake paraformaldehyde and FM 282
Above 120 sec titration time, where the para- hardener (Borden Chemicals).
formaldehyde samples have been more exten-
sively heat treated, the gelation times remain wise titration precludes conventional kinetic
almost invariant. It should be recognized that analysis, useful information for comparing dif-
the shape of the curve in Fig. 3 is highly de- ferent paraformaldehyde forms can be ob-
pendent on titration conditions chosen. In this tained. Figure 5 shows time intervals mea-
case, about 30% of the paraformaldehyde sam- sured for sequence titrations involving 20,40,
ple needs to be reacted to generate a large 60, and 80% of the total of different parafor-
change in pH. maldehyde samples. The shape of these re-
sponses can in part be explained in terms of
Uniformity of paraformaldehyde samples the reactivity distribution in each sample.
Since a given paraformaldehyde sample is a The short time intervals needed to consume
mixture of polymer chains, nonuniformity in up to 80% of the flake sample (C) indicate the
reactivity is likely to occur. A stepwise titration rapid breakdown of this polymer into highly
method was utilized in an attempt to delineate reactive monomeric formaldehyde. In con-
this distribution in reactivity. The technique trast, only the first 20% of the paraformalde-
involves successive measurement of titration hyde sample heated for 9 days (2A) depoly-
times for each fractional component of the merized at a similar rate, while the fraction
sample. The reactivity is measured for the first between 40 and 60% yielded monomeric form-
20% of the sample, then the second 20%, and aldehyde about 30 times slower than the flake
so on until the entire sample is consumed. The sample. This result indicates that heat treat-
result, as shown in Fig. 4, is a sawtooth shape ment converts some of the lower molecular
titration curve with the distance between two weight polymers into higher molecular weight
successive peaks being the time needed to con- ones.
sume 20% of the paraformaldehyde. It is ev- A consequence of these differences in de-
ident from this figure that the time interval is composition rate for portions of a sample is
increasing, and hence the rate of polymer that it is possible to make one paraformalde-
breakdown is decreasing for each successive hyde appear more reactive for gelation than
increment of paraformaldehyde reacted. another by using an excess in the reaction with
While the very nature of this method of step- PRF resin. Since gelation precedes cure, the
78 WOOD AND FIBER SCIENCE, JANUARY 1992, V. 24(1)

fraction of the total paraformaldehyde con- need careful consideration when utilizing this
trolling gelation time will be a function of resin technique.
formulation. Information about these frac-
ACKNOWLEDGMENTS
tional reactivity differences between samples
provides a useful guide for controlling both The author is indebted to J. T. Malito, A.
adhesive pot life and presence of excess para- Sidhu, and H. S. Fraser for technical assistance
formaldehyde, which may result in prevent- and to Weyerhaeuser for partial funding sup-
able formaldehyde emissions both during and port for this project.
after resin cure.
REFERENCES
CONCLUSION CHOW,S. 1977. A curing study of phenol-resorcinol-
The reaction of paraformaldehyde with sul- formaldehyde resins using infrared and thermal analy-
sis. Holzforschung 3 1(6):200-205.
fite can be monitored by in situ titration with
WALKER,J. F. 1964a. Paraformaldehyde. Pages 145-1 58
a weak acid. This technique allows a rapid and in Formaldehyde. ACS Monogram Series 159. Ameri-
convenient method for determining total and can Chemical Society, Washington, D.C.
fractional paraformaldehyde reactivity. Titra- -. 1964b. Quantitative analysis. Pages 486-488 in
tion results can, in part, be correlated with Formaldehyde. ACS monogram series 159. American
Chemical Society, Washington, D.C.
gelation times. Temperature, pH, and para-
formaldehyde particle size are parameters that