Methylolation of Melamine with Incipient Condensation.

I. Synthesis and Characterization

V. V. Nicolau,1,2,3 M. Martinelli,4 M. C. Strumia,4 D. A. Estenoz,1 G. R. Meira1

1
INTEC (CONICET and U.N.L.), Güemes 3450, Santa Fe 3000, Argentina
2
Centro S.A., San Francisco 2400,Córdoba, Argentina
3
Departamento Quı́mica, Fac. Reg. San Fco., U.T.N., San Francisco 2400, Córdoba, Argentina
4
Departamento Qca. Orgánica, Fac. Ciencias Quı́micas, U.N.C., Córdoba 5000, Argentina

Received 25 September 2008; accepted 21 November 2008

DOI 10.1002/app.29729
Published online 2 April 2009 in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: This work investigates five reactions than that of species with two or more rings per molecule.
between melamine (M) and formaldehyde (F) at pH ¼ 9.0, The following were estimated by combining the SEC and
and with initial F : M ratios of 2. The samples were ana- volumetric measurements: number-average molecular
lyzed by size exclusion chromatography (SEC), volumetric weights, and average functionalities of reactive H’s, meth-
techniques, and (1H and 13C) NMR. Condensation was ylols, methylene bridges, and ether bridges. The dissolu-
quantified by the SEC measurements. In the experiments tion of M in water was essentially instantaneous. The
at 38 and 48
C, condensation was almost negligible, and measurements are employed in the next article of this se-
the reversibility of methylolation reactions determined that ries for adjusting a novel mathematical model. VC 2009 Wiley

equilibriums were reached in the ratios between primary, Periodicals, Inc. J Appl Polym Sci 113: 1030–1041, 2009
secondary, and tertiary amines. In the experiments at 60,
70, and 90
C, condensation was significant, but in all Key words: resins; synthesis; gel permeation chromato-
cases, the mass fraction of single-ringed species was larger graphy; NMR; molecular weight distribution

INTRODUCTION this lowers the pH and induces a premature

gelation.)
Melamine-formaldehyde (MF) resins are oligomers
In the initial methylolation stage,2 free formalde-
obtained by reaction between melamine (M) and
hyde reacts with primary or secondary amines to
formaldehyde (F) in aqueous medium. They are
produce mono- or dihydroxymethylamines:
used as adhesives, for the production of molded
articles, and for external transparent coatings in dec- NH2 þ CH2 O ! 
 NHCH2 OH (1)
orative laminates. For this last application, an a-cel-
lulose paper is impregnated with the resin solution
NHCH2 OH þ CH2 O ! 
 NðCH2 OHÞ2 (2)
in the presence of an acid, the paper is dried,
assembled with other impregnated papers, and the Equations (1) and (2) are often replaced by a
resulting ‘‘sandwich’’ is cured under pressure and at global methylolation/demethylolation1,3:
temperatures between 140 and 170
C. The degree of
crosslinking and chemical reactivity of the external H þ CH2 O ! 
 CH2 OH (3)
coating strongly depend on the molecular character-
istics of the base MF resin. In contrast, Gordon et al.2 proposed a detailed
Base MF resins are synthesized in batch stirred- methylolation/demethylolation mechanism, with
tank reactors, and with initial comonomer ratios generation of nine single-ringed methylolmelamines.
[F]
/[M]
of around 1.7.1–4 To limit condensation, The condensation reactions are3:
the pH is maintained between 9.0 and 9.5. (At pH  CH2 OH ! 
H þ  CH2 
 þ H2 O (4)
> 10, F decomposes into methanol and formic acid;
CH2 OH ! 
2 CH2 OCH2 
 þ H2 O (5)
Correspondence to: G. R. Meira (gmeira@santafe-conicet.
gov.ar). where ACH2A and ACH2OCH2A represent generic
Contract grant sponsor: Centro S.A., CONICET, U.N.L., methylene- and ether bridges, respectively. More
and SeCYT.
specifically, methylene bridges can be unsubstituted
Journal of Applied Polymer Science, Vol. 113, 1030–1041 (2009) (AHNCH2NHA), monosubstituted (AHNCH2NRA),
C 2009 Wiley Periodicals, Inc.
V or disubstituted (ARNCH2NRA). Similarly, ether
METHYLOLATION OF MELAMINE WITH INCIPIENT CONDENSATION 1031

bridges can be AHNCH2OCH2NHA, AHNCH2 with only two remaining undetected: N,N-dimethy-
OCH2NRA, or ARNCH2OCH2NRA. lolmelamine and N, N, N0 ,N0 -tetramethylolmelamine.
In water solution, formaldehyde is mainly present Analyzing the HPLC fractions with two or more az-
as methylene glycol5: ine rings per molecule by 13C-NMR, Ebdon et al.15
detected both methylene and ether bridges, and
CH2 O þ H2 O ! HOCH2 OH (6) determined the relative amounts of primary, second-
with [HOCH2OH]  [CH2O]. Concentrated F solu- ary, and tertiary amines. In addition, the ratio of sec-
tions (or ‘‘formalines’’) can polymerize into parafor- ondary to primary amines was estimated by 1H-
maldehyde and react with methanol to generate NMR.15 Analyzing resins obtained under similar
hemiformals. Methanol is normally incorporated in reaction conditions by 13C-NMR, Braun and Legra-
commercial F solutions to avoid generation of para- dic16 and De Breet et al.,12 detected ether bridges but
formaldehyde.5 In dilute solution and in the absence not methylene bridges. Taken from Refs.,11–15 Table I
of methanol, the total formaldehyde concentration is: summarizes the 13C and 1H signal assignments.
Scheepers et al.17 analyzed by 13C-NMR the con-
½FT  ¼ ½CH2 O þ ½HOCH2 OH (7) densation of low methylolmelamines at 90
C;
determining the evolutions of ANHCH2OH,
At the beginning of the F-M reaction (and unless AN(CH2OH)2, AHNCH2NHA, AHNCH2NRA, and
it were carried out in highly dilute solution), only a AHNCH2OCH2NHA. The following groups
fraction of the total M is initially dissolved.6 The fol- remained undetected: disubstituted methylene
lowing expression was proposed7 for the water solu- bridges (ARNCH2NRA), monosubstituted ether
bility of M (in mol/L) as a function of temperature bridges (AHNCH2OCH2NRA), and disubstituted
(in
K): ether bridges (ARNCH2OCH2NRA). The experi-
  ments were as follows: (a) freeze-dried mixtures of
SM ðTÞ ¼ 0:0794 101642=Tþ5:101 (8) low methylolmelamines were dissolved in DMSO-d6
(16.6% w/w); with toluene as internal standard; (b)
In our previous publication,8 a mathematical the solutions were loaded into sealed NMR tubes
model was developed for the methylolation of M in and were heated at 90
C; and (c) samples were taken
dilute solution and in the absence of condensation. It and analyzed along the curing process. The concen-
is based on the mechanism by Gordon et al.,2 and it trations of ANHCH2OH, AHNCH2NHA, and
assumes an instantaneous equilibrium for eq. (6). AHNCH2NRA were estimated from the peak areas
The 24 associated methylolation/demethylolation ki- around 64, 47, and 52 ppm, respectively. Unfortu-
netic constants were adjusted to measurements of nately, AN(CH2OH)2 and AHNCH2OCH2NHA
reactions carried out at 48
C and at pH ¼ 9.0.9 appear superimposed at 68.5 ppm (Table I). For this
Several analytical techniques have been used for reason, two indirect methods were employed for
analyzing MF resins.1,2,5,10–22 For reactions carried first determining [AHNCH2OCH2NHA], and then
out at 35, 40, and 70
C; and pH between 3 and 10.6, [AN(CH2OH)2] from its difference with the total 68.5
Okano and Ogata1 combined the results of the iodo- ppm signal. The first method is totally based on the
metric and sulfite methods5 for estimating the con- 13
C spectra, and it assumes a reaction scheme where:
centrations of FT and global methylols. These (a) unsubstituted ether bridges are produced by
methods are also applied in this work, and are reaction between two ANHCH2OH; (b) unsubsti-
described further below. Gordon et al.2 used a modi- tuted methylene bridges are produced by reaction
fied sulfite method10 for determining the evolution between ANHCH2OH and ANH2; (c) methylolations
of [FT] along a series of methylolation/deme- are neglected because the condensations were car-
thylolations. ried out in the absence of F; and (d) monosubstitued
For reactions at pH > 8 and T > 70
C, several methylene bridges are not produced by reaction
publications have analyzed the reaction products by between two ANHCH2OH groups because the areas
13
C-NMR11–17 and 1H-NMR.14,15,18 Drawbarn et al.,11 under azine carbons bearing primary and secondary
De Breet et al.,12 and Tomita and Ono13 directly ana- amines were seen to remain constant along the con-
lyzed the oligomers mixture by 13C-NMR, employ- densations. The concentration of AHNCH2OCH2
ing dimethylsulfoxide (DMSO-d6) as solvent. In NHA was obtained from:
contrast, Ebdon et al.14,15 first isolated several reac-
tion mixture components by preparative HPLC NHCH2 OH0
½
(with a water-methanol mixture as mobile phase), ¼ ½
NHCH2 OH þ ½
HNCH2 NH
and then analyzed each of the isolated components
þ 2½
HNCH2 OCH2 NH ð9Þ
by NMR. For reactions carried at 50
C, pH ¼ 8.3,
and [FT]
/[M]
between 0.5 and 16, Ebdon et al.14 where [ANHCH2OH]0 is the initial concentra-
identified seven simple-ringed methylolamelamines, tion of secondary amines, and [ANHCH2OH] and

Journal of Applied Polymer Science DOI 10.1002/app

1032 NICOLAU ET AL.

TABLE I
NMR Analysis of MF Resins: Frequency Assignments (in ppm) at Several Measurement Temperatures
Breet Tomita and Ebdon Ebdon
Dawbarn et al.11 et al.12 Ono13 et al.14 et al.15

(25
C) (70
C) (25
C) (100
C) (25
C) (25
C)

(a) 13C Spectra: Azine rings 167.4a 165.7 167.2 166.9 167.21a
167.2b 166.97b
167.0c 166.79c
– 166.79e
166.3b 166.3 166.5 166.1 166.05b
166.0c 165.78c
– 165.64e
165.8d 165.51d
– 165.48f
– 165.46g
– – 166.2 165.6 165.00e
165.2f 164.66f
165.2g 164.65g
165.2h 164.59h
Methylol groups ANHCH2OH 64.5 64.7 64.8 64.5 – 64
AN(CH2OH)2 68.7 – – – – 68–69
AN(CH2A)CH2OH – – 69.8 69.5 – –
Methylene bridges ANHCH2NHA – – 47.3 47.3 – 47
AN(CH2A)CH2NHA – – 52.2 52.1 – 52
AN(CH2A)CH2N(CH2A)A – – – – – 58
Methylene ethers of methyl groups ANHCH2OCH3 – 72.9 73.0 72.5 – 73
AN(CH2A)CH2OCH3 – 77.3 77.4 76.6 – 77
Ether bridges ANHCH2OCH2NHA – 69.9 69.8 68.6 – 68–69
AN(CH2A)CH2OCH2NHA – – 73.0 73.3 – 72
AN(CH2A)CH2OCH2N(CH2A) – – – – – –
Hemiformals of methylol groups ANHCH2OCH2OH 67.4 – 69.8 68.6 – –
AN(CH2A) CH2OCH2OH 71.6 – 73.0 73.3 – –
ANHCH2OCH2OH 85.6 – – 86.2 – –
AN(CH2A)CH2OCH2OH 85.6 – – 86.2 – –
Methanol CH3OH – 49.3 50.7 48.4 – –
Methoxy groups ACH2OCH3 – 55.0 55.3 54.5 – 55
Methylene glycol and derivatives HOCH2OH 84.4 83.8 – 82.1 – 82
HOCH2OCH2OH – 86.9 – 85.2 – –
HOCH2OCH3 – – – 89.5 – –
H(OCH2)nOCH2OCH3 – – – 93.9 – –
(b) 1H Spectra:
ANHCH2OH 7.31–746f 7–8
ANH2 – 6–7
AN(CH2OH)2 5.37f 5.4
ANHCH2OH 5.21f 5.2
AN(CH2OH)2 4.97f 5.0
ANHCH2OH 4.67f 4.7

Reproduced from Refs. 11–15.

a
Melamine.
b
Monomethylolmelamine.
c
N, N0 -dimethylolmelamine.
d
N, N0 , N00 -trimethylolmelamine.
e
N, N, N0 -trimethylolmelamine.
f
N, N, N0 , N00 -tetramethylolmelamine.
g
Pentamethylolmelamine.
h
Hexamethylolmelamine.

[AHNCH2NHA] were determined as mentioned ear- (as determined by 13C-NMR). The calculation
lier. The second method estimates [AHNCH2O assumed the presence of the following methylene
CH2NHA] from the difference between the concen- bridges: AHNCH2NHA and AHNCH2NRA; where
tration of condensation water (as determined by 1H- the last group is generated by reaction between
NMR), and the concentration of methylene bridges ANH2 and AN(CH2OH)2.

Journal of Applied Polymer Science DOI 10.1002/app

METHYLOLATION OF MELAMINE WITH INCIPIENT CONDENSATION 1033

Figure 1 NMR analysis of the initial F solution. The 1H spectrum (a), and the 13
C spectrum (b) show negligible amounts
of methanol, hemiformals, or paraformaldehyde.

Analyzing F-free solutions of single-ringed meth- (DMSO-d6 Sigma Aldrich spectroscopic grade, purity
ylolmelamines in DMSO, Duliban et al.18 employed 99.9%).
1
H-NMR for investigating the intramolecular rear- Prior to each reaction, the F solutions were pre-
rangements that take place when the solutions are pared by depolymerization of solid paraformalde-
heated or stored for several months at ambient hyde. To this effect, the polymer was washed with
temperature. deionized water, filtered under vacuum, and dried
On several occasions,13,19–22 size exclusion chroma- for a few minutes at 50
C. Then, 53 g of the solid
tography (SEC) has been employed for analyzing was loaded in 100 mL of pure water, and the depo-
MF resins. In spite of the low solubility of MF resins lymerization was carried out until completion, at
in tetrahydrofurane (THF), this solvent was still 70
C and pH ¼ 10.0. The initial F concentrations
employed (in combination with polyvinyl acetate gel were determined via the sulfite method,5 and they
columns) by first transforming the original resin into all resulted between 7 and 8 mol/L. The initial F sol-
the trimethylsilated derivative.19 Direct SEC fractio- utions were also analyzed by NMR (Fig. 1). The 1H-
nations of MF resins have been carried out with the NMR spectrum [Fig. 1(a)] shows a minimal contami-
following systems: (a) dimethylformamide (DMF) nation by methanol (see small peak at 3.47 ppm cor-
solvent combined with alkylated dextran gel responding to the three methyl H’s of methanol).
columns (Sephadex LH-20)20–22; and (b) dioxane The 13C-NMR spectrum [Fig. 1(b)] exhibits a large
with a full set of six poly(styrene-divinylbenzene) peak at 82.1 ppm (due to methylene glycol carbons),
columns.13 but not observed are methylene signals of parafor-
In this article, five reactions between M and F at maldehyde at 85–90 ppm and hemiformal methoxy
pH ¼ 9.0 and at temperatures between 38 and 90
C signals at 55 ppm (Table I). Thus, the F solutions
are analyzed by volumetric techniques, SEC, and were considered essentially free of impurities.
NMR.
Kinetics of melamine dissolution
MAIN EXPERIMENTS
Three experiments were carried out to estimate the
The following chemicals were employed: (a) mela- rate of dissolution of M in water at 50
C and pH
mine (purity 99.8%, from Agrolinz Melamine Inter- ¼ 9.0  0.5. In all cases, 12 g M were added into 135
national GmbH, Austria); (b) paraformaldehyde g of water, and the pH was adjusted with the NaOH
(Cicarelli, Argentina); (c) a 2N solution of NaOH solution. The dissolution periods of Exps 1–3 were
(Anedra proanalysis, purity 97%); (d) a 1N solution 20 s, 15 min, and 30 min, respectively. After such
of Na2SO3 (Cicarelli proanalysis, purity 98%); (e) 0.1 periods, the mixtures were filtered under vacuum
and 2N solutions of HCl (Cicarelli proanalysis, conc. with Filtrak paper N
42, and dried until constant
36–38.5%); (f) a 0.1N iodine solution (Anedra proa- weight for 2 h at 120
C. The concentrations of dis-
nalysis); (h) Na2S2O3.5H2O (British Drug Houses, solved M were obtained from the difference between
proanalysis); (g) DMF (Sintorgan spectroscopic the initial and undissolved M masses. For Exps 1–3,
grade); and (h) deuterated dimethylsulfoxide the dissolved M concentrations resulted 1.170, 1.092,

Journal of Applied Polymer Science DOI 10.1002/app

1034 NICOLAU ET AL.

and 1.128 g M/100 g H2O, respectively. These MEASUREMENTS AND DISCUSSION

(essentially constant) values are close to the esti-
SEC measurements
mate via eq. (8) at 50
C of 1.041 g M/100 g H2O.
Thus, the dissolution of M was considered as The chromatograph was a Waters Breeze fitted with
essentially instantaneous with respect to the rates a W2414 differential refractometer and a Shodex KD
of reaction. 802.5 column. The carrier solvent was DMF at
1 mL/min, and the system was operated at 40
C.
The dry samples were dissolved in DMF (0.1% w/
w), and the injection volumes were 80 lL. Unfortu-
Main reactions
nately, a direct molar mass calibration was impossi-
Five reactions were carried out at 38, 48, 60, 70, and ble, due to the lack of melamine-like standards.
90
C (Table II). In all cases, the pH was adjusted to Also, indirect calibrations with PEG standards pro-
9.0  0.5 by addition of the NaOH solution, and the vided inconclusive results, due to the large struc-
initial comonomers ratio was fixed at [FT]
/[MT]
tural differences between the linear standards and
% 2 (a typical industrial value). However, the abso- the ringed oligomers.
lute values of [FT]
and [MT]
were about 10 times Consider the chromatograms of Figure 3. In all
lower than in industry (Table II); and this was to cases, the low-conversion samples exhibit single
reduce the amount of initial undissolved M, while peaks (R1) that are indicative of single-ringed spe-
simultaneously increasing the reaction times. Table cies. The final samples of Exps 1 and 2 exhibit small
II also presents the initial concentrations of dissolved second peaks (R2), that correspond to double-ringed
M ([Md]
), as estimated through eq. (8). According species [Fig. 3(a,b)]. The final samples of Exps 3–5 also
to such estimates, Exps. 1–4 were all initially hetero- contain triple-and-higher ringed molecules (R3) [Fig.
geneous, whereas Exp. 5 was homogeneous through- 3(c–e)]. The mass fractions (wi) of species with 1-, 2-,
out. Errors are to be expected however, because and 3-or-more azine rings per molecule were deter-
eq. (8) does not include the effect of the other reac- mined from the areas under R1, R2, and R3 (Table III).
tion components (i.e.: F, methylene glycol, and meth-
ylolated species). Volumetric measurements
The reactor was a 2 L stainless-steel Parr 4522M
fitted with a stirrer, an electric heating jacket, an The concentration of total formaldehyde was deter-
internal cooling coil, a temperature indicator, a mined via the sodium sulfite method.5 First, the total
manometer, and valves for feeding the reagents formaldehyde contained in 5 mL of reaction mixture
and for extracting the samples by internal pressur- was consumed with an excess (25 mL) of a cold 1N
ization with Argon. First, the reactor was charged sodium sulfite solution through:
with water and M in powder under stirring, and
the pH was adjusted to 9.0. Then, the mixture was CH2 O þ Na2 SO3 þ H2 O ! HCHðNaSO3 ÞOH
heated to the final desired reaction temperature, ð10Þ
and the reaction was started by loading the F so-
lution at pH ¼ 9.0, and preheated at the same The technique provides [FT] rather than [CH2O],
temperature. because all the consumed CH2O is almost instanta-
The reaction samples were rapidly cooled in ice neously replaced via eq. (6).23 Then, [FT] was quanti-
water, and the following were measured: (a) [FT] by fied by direct titration of the generated NaOH with
the sodium sulfite method5; and (b) [FT]þ[ACH2OH] the 0.1N HCl solution in the presence of thymolphta-
by the iodometric method5 (see Table II and Fig. 2). lein. During this operation, the product of eq. (10)
Some of the reaction samples were additionally ana- was maintained cold to prevent hydrolysis of the
lyzed by: (a) SEC (see mass chromatograms in Fig. methylolmelamines in acid medium.
3, and numerical estimates in Table III); and (b) 1H The added concentrations of FT and ACH2OH
and 13C-NMR (see spectra of Exp. 3 in Fig. 4, and were determined via the iodometric method.5 First,
1
H-NMR estimates in Table III). The reaction sam- sodium hypoiodite NaIO was in situ generated, as
ples taken for the instrumental measurements were follows:
frozen and lyophilized (to eliminate water and F),
and the dry powders were freeze-stored until their 6 NaOH þ 3 I2 ! 3 NaI þ 3 NaIO þ 3 H2 O (11)
analyses. Table III presents the sample masses before
and the generated NaIO consumes both F and the
and after lyophilization. The final sample of Exp. 5
methylol groups through:
(taken at 1535 min) was discarded, because an insol-
uble condensate was observed over the stirrer and
CH2 O þ NaIO þ NaOH ! HCOONa þ NaI
internal coil. Instead, the sample at t ¼ 1350 min
was adopted as the final sample of Exp. 5. þ H2 O ð12Þ

Journal of Applied Polymer Science DOI 10.1002/app

 TABLE II
Experiments 1–5: Reaction Conditions and Volumetric Measurements. All the Reactions Were Carried Out at:
pH 5 9.0 6 0.5; [FT]
/[MT]
5 2.0; and [H2O] 5 53.5 mol/L
Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5

(a) Reaction Conditions

T ¼ 38  1
C T ¼ 48  1
C T ¼ 60  1
C T ¼ 70  1
C T ¼ 90  1
C
[FT]
¼ 0.409 mol/L [FT]
¼ 0.399 mol/L [FT]
¼ 0.407 mol/L [FT]
¼ 0.421 mol/L [FT]
¼ 0.405 mol/L
[MT]
¼ 0.202 mol/L [MT]
¼ 0.200 mol/L [MT]
¼ 0.201 mol/L [MT]
¼ 0.202 mol/L [MT]
¼ 0.202 mol/L
[Md]
¼ 0.0526 mol/La [Md]
¼ 0.0768 mol/La [Md]
¼ 0.118 mol/La [Md]
¼ 0.164 mol/La [Md]
¼0.300 mol/La

[FT] þ [FT] þ [FT] þ [FT] þ [FT] þ

Time [FT] [ACH2OH] Time [FT] [ACH2OH] Time [FT] [ACH2OH] Time [FT] [ACH2OH] Time [FT] [ACH2OH]
(min.) (mol/L] (mol/L) (min.) (mol/L) (mol/L) (min.) (mol/L) (mol/L) (min.) (mol/L) (mol/L) (min.) (mol/L) (mol/L)

b) Volumetric Measurements
0 0.409 0 0.399 0 0.407 0 0.421 0 0.405
2 0.374 0.394 2 0.330 2 0.354 2 0.343 0.410 2 0.262 0.377
20 0.365 0.399 20 0.309 19b 0.191 0.410 20b 0.154 0.412 19b 0.150 0.413
58 0.331 0.397 61 0.187 0.397 37 0.125 0.425 37 0.125 0.415 38 0.135 0.407
90 0.313 0.407 95 0.137 0.382 59 0.110 0.410 55 0.118 74 0.139 0.402
131b 0.268 0.402 129b 0.107 0.387 69 0.107 0.410 65 0.118 0.415 93 0.132 0.402
174 0.234 0.402 212 0.0786 0.397 139 0.105 0.410 92 0.116 0.405 1350b 0.0711 0.163
METHYLOLATION OF MELAMINE WITH INCIPIENT CONDENSATION

1470b 0.0674 0.405 1370b 0.0749 0.402 1370b 0.0992 0.410 134 0.116 1535 –c –c
1560 0.0636 0.402 1440 0.073 0.407 1510 0.105 0.405 153 0.116 0.410
2820 0.0636 0.405 1520 0.0749 0.407 1680 0.0955 0.394 1360b 0.110 0.364
4260 0.0562 0.407 1605 0.0749 0.407 2825 0.0936 0.384 1425 0.110 0.359
4375b 0.0562 0.397 2833 0.0655 0.402 2970 0.0917 1500 0.107 0.354
2880 0.0711 0.405 3030b 0.0880 0.379 1595 0.0992 0.349
3025b 0.0711 0.397 1905b 0.103 0.351
a
Initial conc. of dissolved M according to eq. (8).
b
Sample further analyzed by NMR and SEC (Table 3).
c
Discarded value because of the presence of an insoluble condensate.
1035

Journal of Applied Polymer Science DOI 10.1002/app

1036 NICOLAU ET AL.

Figure 2 Experiments 1–5: time evolutions of [FT] (n) and [FT] þ [ACH2OH] (*). Note the two different scales employed
in the time axes.

2
CH2 OH þ 3 NaIO þ 2 NaOH
! 2 CðOHÞONa þ 3NaI þ 3H2 O ð13Þ

Then, the excess of hypoiodite is decomposed into

sodium iodure (NaI) and sodium iodate (NaIO3)
under acidic conditions:
3 NaIO ! 2 NaI þ NaIO3 (14)

and the produced NaI regenerates free I2 through:

NaIO3 þ 5 NaI þ 6 HCl ! 3 I2 þ 6 NaCl þ 3 H2 O
(15)

Finally, the free I2 is titrated with the Na2S2O3 so-

lution:
2 Na2 S2 O3 þ I2 ! Na2 S4 O6 þ 2 NaI (16)

and ([FT] þ [ACH2OH]) is determined by back-titra-

tion; that is, by subtraction of the required titration
volume from the titration volume employed for a
blank reagent without sample.5 The experimental
procedure was as follows. First, 1 mL of sample,
25 mL of a 0.1N iodine solution, and 4 mL of 2N
NaOH were loaded into an Erlenmeyer. The mixture
was maintained at room temperature for 10 min,
Figure 3 Experiments 1–5: size exclusion chromatograms
and 4.5 mL of the 2N HCl solution were added to
of samples obtained at three reaction times. Peaks R1, R2, liberate the generated iodine. The liberated iodine
and R3 indicate single-, double-, and triple-or-more ringed was titrated with a 0.1N sodium thiosulfate solution
species. (prepared from the corresponding salt); and 2 mL of

Journal of Applied Polymer Science DOI 10.1002/app

 TABLE III
Experiments 1–5: Estimates of the Instrumental Measurements at Three Reaction Times
1
SEC H NMR Titration only Titration/SEC
Sample
d d d e e
Time volume Lyoph. Mn f met fH Mn f met f He f MB e f EB e

b ½
NHCH2 OH ½
NðCH2 OHÞ2  ½
NðCH2 OHÞ2 
Exp N (min) (VL)a (L) Mass (L) (g) w1 w2b w3b ½
NH2  ½
NCH2 OH ½
NCH2 OH (g/mol) (#/molec.) (#/molec.) (g/mol) (#/molec.) (#/molec.) (#/molec.) (#/molec.)

1 131 0.00791 0.244 1 – – –c –c –c 147 0.666 5.33 145 0.628 5.37 – –

1470 0.00791 0.296 1 – – 1.78 0.0204 0.0209 177 1.67 4.33 172 1.52 4.48 – –
4375 0.00786 0.288 0.938 0.062 – 1.73 – – – – – 184 1.57 4.42 0.0293 0.00438
2 129 0.00772 0.287 1 – – 1.47 0.0184 0.0200 170 1.40 4.60 162 1.21 4.79 – –
1370 0.00768 0.314 0.971 0.029 – 1.64 – – – – – 170 1.30 4.70 0.0113 0.00420
3025 0.00776 0.265 0.938 0.062 – 1.68 – – – – – 186 1.69 4.36 0.0300 0.00370
3 19 0.00974 0.346 1 – – 1.60 0.0177 0.0182 158 1.09 4.91 154 0.945 5.06 – –
1370 0.00977 0.385 0.921 0.065 0.014 1.65 – – – – – 179 1.27 4.75 0.0333 0.0130
3030 0.00946 0.357 0.790 0.138 0.071 1.79 – – – – – 208 1.41 4.88 0.103 0.0363
4 20 0.00788 0.316 1 – – 1.58 0.0174 0.0185 166 1.28 4.72 156 0.994 5.01 – –
1360 0.00745 0.273 0.689 0.181 0.130 1.81 – – – – – 228 1.11 5.18 0.155 0.0728
1905 0.00731 0.245 0.722 0.176 0.102 1.76 – – – – – 222 1.39 5.01 0.141 0.0547
5 19 0.00780 0.253 1 – – 1.83 0.0182 0.0202 164 1.30 4.70 166 1.33 4.67 – –
1350 0.00773 0.074 0.757 0.131 0.112 –c –c –c – – – 239 1.69 4.38 0.163 0.00860
a
Estimated from the sample mass, assuming a solution density of 1 g/mL.
b
Mass fractions of single-, double-, and triple-ringed molecules (see Fig. 3).
c
Discarded estimate.
d
Estimated through eqs. (18–20).
e
Estimated through eqs. (41–45).
1038 NICOLAU ET AL.

obtained from [FT] and the global initial concentra-

tion of M ([MT]
) through:
 
MM ½MT 0 þ MF ½FT 0  ½FT 
Mn ¼ (18)
½MT 0

where MM (¼ 126 g/mol) and MF (¼ 30 g/mol) are

the molar masses of M and F. In addition, the aver-
age methylol and hydrogen functionalities of single-
ringed molecules (f meth and f H , respectively) were
found from [ACH2OH] and [MT]
, as follows:
½CH2 OH
f meth ¼ (19)
½MT 0

½HT 
fH ¼ (20)
½MT 0

with:

½HT  ¼ 6½MT 0  ½CH2 OH (21)

Figure 4 Samples of Exp. 3, at three reaction times: (a) where [AHT] is the total concentration of unreacted
1
H spectra; and (b) 13C-NMR spectra. In the 1H spectra, H’s.
peaks 3, 4, 5, and 6 are potentially contaminated by (meth-
ylene or ether) bridge signals.
NMR measurements
The NMR spectrophotometer was a 400 MHz
Brücker (Avance II), and the measurements were
a soluble starch indicator were added after the solu-
tion turned into a pale yellow. taken at ambient temperature. The samples were
The results are in Table II and in Figure 2. As prepared by dissolving 35 mg of lyophilized resin in
expected, the rate of consumption of FT increases 0.7 mL of DMSO-d6. For Exp. 3, Figure 4 presents
with the temperature. In Exps. 1 and 2 with negligi- the 1H and 13C spectra at three reaction times. The
ble condensation, all the reacted CH2O is trans- second analytical sample of Exp. 5 taken at 1350 min
formed into ACH2OH, and ([FT] þ [ACH2OH]) presented dissolution problems. This produced low-
remains essentially constant [Fig. 2(a,b)]. The follow- resolution 1H spectra, and therefore, their estimates
ing can be written: were discarded (Table III). 1H-NMR estimates of the
first sample of Exp. 1 were also discarded because of
technical errors.


½FT  ¼ ½FT  þ ½CH2 OH þ ½CH2  The 1H spectra provided quantitative information
þ 2½CH2 OCH2  ð17Þ (Table III). Special software (HNMR predictor 5.0 by
ACD labs) was used to simulate the spectrum of
protons in (methylene and ether) bridges. Double-
where [FT]
is total initial formaldehyde. In the ringed molecules with several types of methylene or
absence of condensation, eq. (17) reduces to: [FT]
ether bridges were simulated; and the results were
¼ [FT] þ [ACH2OH]; and an equilibrium is reached as follows: (a) proton signals from un-, mono-, and
due to the reversibility of methylolation reactions. disubstituted methylene bridges appear in the
In the presence of condensation, [ACH2A] þ 2 ranges ([5.65–5.0] and [5.0–4.80]), ([5.60–5.05] and
[ACH2OCH2A] builds up at the expense of [ACH2 [5.05–4.85]), and [5.60–5.0]; respectively; and (b) pro-
OH]; and ([FT] þ [ACH2OH]) falls along the reaction ton signals from un-, mono-, and disubstituted ether
[Fig. 2(c–e)]. bridges appear in the ranges [5.15–4.60], ([5.45–4.75]
In the absence of condensation, several average and [4.20–4.0]), and ([5.60–4.90] and [4.20–3.95]),
properties were calculated from the estimates of [FT] respectively.
and [ACH2OH]; with [ACH2OH] obtained from the The 1H spectra [Fig. 4(a)] do not show signals in
difference between ([FT] þ [ACH2OH]) and [FT]. The 4.20–3.95 ppm, and this suggests negligible amounts
number-average molecular weight of the single- of mono- and disubstituted ether bridges. The con-
ringed molecule mixture (that includes M) is centrations of ANHCH2OH, ANH2, AN(CH2OH)2,

Journal of Applied Polymer Science DOI 10.1002/app

METHYLOLATION OF MELAMINE WITH INCIPIENT CONDENSATION 1039

ANHCH2OH, AN(CH2OH)2, and ANHCH2OH were combined for estimating M n , f meth , f H , f MB , and f EB
quantified from the areas Ai (i ¼ 1, . . . , 6) [see (where f MB and f EB are, respectively, the average
Fig. 4(a)]; and to this effect the peaks were inte- numbers of methylene and ether bridges per mole-
grated in the ranges [7.65–7.00], [6.55–6.00], [5.50– cule). The assumptions were as follows: (1) negligi-
5.30], [5.30–5.05], [5.05–4.91], and [4.80–4.60], respec- ble concentration of species with more than three
tively. Even though bridge signals were not expected rings per molecule; (2) only unsubstituted (methyl-
to affect the signals from ANHCH2OH and ANH2, ene and ether) bridges are present; (3) the H’s of
they could however interfere with those of unsubstituted (methylene and ether) bridges remain
AN(CH2OH)2, ANHCH2OH, ANHCH2OH, and nonreactive; (4) the average ratios of methylene to
AN(CH2OH)2. ether bridges are identical in double- or triple-ringed
Following Ebdon et al.,15 the ratio of secondary to molecules; and (5) the average number of methyols
primary amines was estimated through: per molecule are proportional to the number of fea-
sible H atoms per molecule (i.e.: six in single-ringed
½NHCH2 OH A1 molecules, eight in double-ringed molecules, and 10
¼ (22)
½NH2  ðA2 =2Þ in triple-ringed molecules). Hypothesis N
4) can be
written:
Consider the results for [ANHCH2OH]/[ANH2] in
Table III. The following can be noted: (a) the final f MB2 f MB3
values remain essentially constant in the last two ¼ (25)
f EB2 f EB3
samples of Exps. 1 and 2 with negligible condensa-
tion; and (b) the value of the third sample of Exp. 4 where f EB2 , f EB3 , f MB2 , and f MB3 are the average ether
is lower than that of the second sample; possibly and methylene bridge functionalities in the fractions
due to condensation reactions and/or to the genera- of double- and triple-ringed molecules. Hypothesis
tion of tertiary amines. N
5 implies that:
For the samples with negligible condensation, the
ratio of tertiary to secondary amines was independ- 6½R1 
½CH2 OH ¼ ½R1  f meth1 (26)
ently estimated through: 6½R1  þ 8½R2  þ 10½R3 
½NðCH2 OHÞ2  1 A3
¼ (23) 8½R2 
½NHCH2 OH 2 A4 ½CH2 OH ¼ ½R2  f meth2 (27)
6½R1  þ 8½R2  þ 10½R3 
and
½NðCH2 OHÞ2  1 A5 10½R3 
¼ (24) ½CH2 OH ¼ ½R3  f meth3 (28)
½NHCH2 OH 2 A6 6½R1  þ 8½R2  þ 10½R3 

The results of eqs. (23) and (24) are also shown in where [Ri] (i ¼ 1, 2, 3) are the molar concentrations
Table III. As expected, the estimates are reasonably of single-, double-, and triple- ringed molecules.
close to each other. These concentrations were obtained through:
Consider the 13C spectra of Figure 4(b). The peaks
at 167 and 64 ppm correspond to Carbon atoms con- Lwi
½Ri  ¼ ; ði ¼ 1; 2; 3Þ (29)
tained in azine rings and secondary amines, respec- Mni VL
tively. Not observed were methylene bridge signals
at 47 ppm nor ether bridge (or secondary methylol) where wi (i ¼ 1, 2, 3) are the mass fractions of spe-
signals at 69 ppm. Thus, 13C-NMR did not provide cies with 1-, 2-, and 3-or-more rings per molecule as
information on the chemical nature and/or on the determined by SEC; L is the total mass of lyophi-
concentration of condensation bridges. Possible rea- lized sample; and VL is the corresponding sample
sons for the absence of bridge signals are: (1) the volume (Table III).
measurements were taken at ambient temperature, Call M ni , f methi , and f Hi (i ¼ 1, 2, 3) the average
while higher resolutions are expected at 90– molecular weights and functionalities of species
100
C13,17; and (2) the sample preparation did not with 1-, 2-, and 3- rings per molecule. These aver-
include a preconcentration stage by HPLC of the ages are interrelated through:
double- and higher-ringed species.14,15
Mn1 ½g=mol ¼ 120 þ 31f meth1 þ f H1 (30)
Estimates obtained by combination of the
volumetric and SEC measurements Mn2 ½g=mol ¼ 240 þ 31 f meth2 þ f H2
 
For the more general case of non-negligible conden- þ 16 1  f EB2 þ 46 f EB2 ð31Þ
sation, volumetric and SEC measurements were

Journal of Applied Polymer Science DOI 10.1002/app

1040 NICOLAU ET AL.

 
Mn3 ½g=mol ¼ 360 þ 31 f meth3 þ f H3 þ 32 1  f EB2 Mn
Meff ru ¼ (47)
rn
þ 92 f EB2 ð32Þ

f meth2 ¼ 2 f meth1  ð1 þ f EB2 Þ (33) Accurate estimates of [ACH2OH] are vital for ac-
ceptable errors in the derived averages. Errors of
5% in [ACH2OH] induce the following relative
f meth3 ¼ 3 f meth1  2ð1 þ f EB2 Þ (34) errors: 1.73 and 2.21% in Mn , 2.32 and 3.57% in f H ,
6.18 and 7.02% in f meth , 3.36 and 6.80% in f MB , 23.15
and 120.25% in f EB . Thus, errors in [ACH2OH]
f H1 ¼ 6  f meth1 (35) strongly propagate into f EB . Other sources of errors
may be due to the model hypotheses. For example,
f H2 ¼ 8  f meth2 (36) an important fraction of species with four or more
rings per molecule would induce errors by defect in
the number-average molecular weights and the
f H3 ¼ 10  f meth3 (37) bridge functionalities.
Consider the results of Table III. In the absence of
where 31, 16, 46, and 120 are, respectively, the mo- condensation, Mn , f meth , and f H obtained through
lecular weights of ACH2OH, AHCH2HA, AHCH2O eqs. (41)–(43) are close to the direct titration esti-
CH2HA, and C3N6 (i.e.: M without the six amine mates via eqs. (18)–(20). In the first sample of Exp. 1,
H’s). Replacing eqs. (33)–(37) into eqs. (30)–(32), one methylolation was incomplete because neither f meth
obtains: (% 0.63) nor f H (% 5.37) had reached their final
  quasi-equilibrium values. In Exps 1 and 2, the final
Mn1 g=mol ¼ 126 þ 30f meth1 (38)
product essentially consisted of single-ringed mole-
  cules of the following average characteristics: f meth
Mn2 g=mol ¼ 234 þ 60 f meth1 (39) % 1.6; f H % 4.4; f meth þ f H % 6; f MB % 0.03; f EB
% 0.004; rn % 1.03; and Meff ru % 180 g/mol (where
  Meff ru is the ‘‘effective’’ repetitive unit molar mass).
Mn3 g=mol ¼ 342 þ 90 f meth1 (40) The results of Exps 1 and 2 are similar due to the
common initial F : M ratios. In Exps 4 and 5 with
From the measurements of [ACH2OH], wi (i ¼ 1,
condensation, the final values of f meth and f H are
2, 3), L, and VL, eqs. (26), (27), (29), and (33)–(40)
similar to those of Exps 1 and 2; while higher values
enable to estimate f methi , f Hi , Mni , [Ri] (i ¼ 1, 2, 3),
are observed for f MB ; f EB ; rn ; and Meff ru .
and f EB2 . The global averages are given by:

Mn ¼ w1 Mn1 þ w2 Mn2 þ w3 Mn3 (41)

CONCLUSIONS
f ½R1  þ f meth2 ½R2  þ f meth3 ½R3 
f meth ¼ meth1 (42) The dissolution of M in water is almost instantane-
½R1  þ ½R2  þ ½R3  ous with respect to the reaction rates, and SEC
proved efficient for quantifying the mass fractions of
f H1 ½R1  þ f H2 ½R2  þ f H3 ½R3  single-, doubled, and triple-or-more-ringed species.
fH ¼ (43) Condensation is negligible at the beginning of the
½R1  þ ½R2  þ ½R3 
reactions, and under such conditions, 1H-NMR
proved appropriate for estimating the relative
f MB2 ½R2  þ f MB3 ½R3  amounts of primary, secondary, and tertiary amines.
f MB ¼ (44)
½R1  þ ½R2  þ ½R3  With negligible condensation, the volumetric techni-
ques provide relatively direct estimates of the aver-
age molar masses and functionalities of the single-
f EB2 ½R2  þ f EB3 ½R3  ringed mixture.
f EB ¼ (45)
½R1  þ ½R2  þ ½R3  Even though 13C-NMR proved inadequate for
detecting bridge formation, the combination of volu-
Finally, the number average chain length rn , and
metric and SEC measurements produced (gross and
the molecular weight of the ‘‘effective’’ repetitive
indirect) estimates of the average molar masses and
unit Meff ru , are given by:
bridge functionalities.
The given set of measurements is employed in the
second part of this series for adjusting a novel math-
rn ¼ f MB þ f EB þ 1 (46) ematical model that estimates the distributions of

Journal of Applied Polymer Science DOI 10.1002/app

METHYLOLATION OF MELAMINE WITH INCIPIENT CONDENSATION 1041

molecular weights and functionalities of the evolv- 9. Tomita, B. J Polym Sci 1977, 15, 2347.
ing reaction mixture. 10. De Jong, J. I.; de Jonge, J. Rec Trav Chim 1952, 71, 890.
11. Drawbarn, M.; Ebdon, J. R.; Hewitt, S. J.; Hunt, J. E. B.; Wil-
liams, I. E.; Westwood, A. R. Polymer 1978, 19, 1309.
The authors thank Marcelo Brandolini and José L. Castañeda 12. De Breet, J.; Dankelman, W.; Huysmans, W. G. B. De Wit
(INTEC) for help in chromatographic analyses. Angew Makromol Chem 1977, 62, 7.
13. Tomita, B.; Ono, H. J. Polym Sci Polym Chem 1979, 17, 3205.
14. Ebdon, J. R.; Hunt, B. J.; O’Rourke, W. T. S. Br Polym J 1987,
References 19, 197.
15. Ebdon, J. R.; Hunt, B. J.; O’Rourke, W. T. S.; Parkin, J. Br
1. Okano, M.; Ogata, Y. J Am Chem Soc 1952, 74, 5728. Polym J 1988, 20, 327.
2. Gordon, M.; Halliwell, A.; Wilson, T. J Appl Polym Sci 1966, 16. Braun, D.; Legradic, V. Angew Makromol Chem 1974, 36, 41.
10, 1153. 17. Scheepers, M. L.; Adriaenses, P. J.; Gelan, J. M.; Carleer, R. A.;
3. Nastke, R.; Dietrich, K.; Reinisch, G.; Rafler, G. J. Macromol Vanderzande, D. J.; De Vries, N. K.; Brandts, P. M. J Polym
Sci-Chem 1986, A23, 5, 579. Sci Polym Chem 1995, 33, 915.
4. Kumar, A.; Katiyar, V. Macromolecules 1990, 23, 3729. 18. Duliban, J.; Galina, H.; Lubczak, J. Appl Spectr 1996, 50,
5. Walker, J. F. Formaldehyde, 3rd ed.; R. E. Krieger Pub Co 528.
Huntington: New York, 1964. 19. Braun, D.; Legradic, V. Angew Makromol Chem 1973, 34, 35.
6. Wirpsza, Z.; Kucharski, M.; Lubczak, J. J Appl Polym Sci 1998, 20. Tsuge, M.; Miyabayashi, T.; Tanaka, S. Jpn Anal 1974, 23,
67, 1039. 1141.
7. MelamineChemicals. Melamine Crystal: Properties, Chemistry, 21. Tsuge, M.; Miyabayashi, T.; Tanaka, S. Jpn Anal 1974, 23,
Analytical Methods, Toxicology; Melamine Chemicals Inc.: 1146.
Donalsonville, USA, 1980. 22. Braun, D.; Pandojo, W.; Fresenius, Z. Anal 1979, 294, 375.
8. Nicolau, V.; Estenoz, D.; Meira, G. J Appl Polym Sci 2007, 106, 23. Winkelman, J. G. M. Absorption of Formaldehyde in Water,
2978. PhD Thesis; University of Groningen: Netherlands, 2003.

Journal of Applied Polymer Science DOI 10.1002/app