The Effect of Catalyst Structure on the Synthesis
of a Dental Restorative Monomer
M. FARAHANI, A.D. JOHNSTON, and R.L. BOWEN
Paffenbarger Research Center, American Dental Association, National Institute of Standards and Technology, Gaithersburg, Maryland 20899
The addition product of 2-hydroxyethyl methacrylate (HEMA)
and pyromellitic dianhydride (PMDA), known as PMDM, is
a mixture of two structural isomers. The para PMDM iso-
mer-currently used in mediating adhesive bonding of restor-
ative materials to hard tooth tissues-is a crystalline solid. The
meta isomer is a liquid. In the synthesis of PMDM, the para
isomer, which can be purified by crystallization, is usually
present to the extent of only 50% of the product mixture. The
effect of the amine catalyst structure was studied relative to its
role in increasing the yield of the para isomer, either by a
reduction in the amount of the meta isomer or by an increase
in the extent of overall reaction. The chemical structure of the
amine catalyst had an important role in the synthesis of PMDM
and influenced the ratio of the isomers. Among aliphatic amines,
especially noteworthy as catalysts that gave excellent yields of
the para isomer in high purity were N,N-di-isopropyl-ethylam-
ine and hexamethylenetetramine.
J Dent Res 70(1):67-71, January, 1991
Introduction.
A three-step protocol for adhesive bonding of dental resins to
dentin and enamel was previously developed (Bowen et al.,
1982). The first step was the application of a mordant solution
of acidic, aqueous, metallic oxalate to the surface to remove
the smeared (disturbed) layer resulting from instrumental tech-
niques used to remove carious material. The second step was
the application of an acetone solution of an N-substituted amino
acid, NPG-GMA [the addition product of N-phenylglycine
(NPG) and glycidyl methacrylate] or NTG-GMA (the addition
product of N-tolylglycine and glycidyl methacrylate). The third
step was the application of an acetone solution of PMDM [the
addition product of two moles of 2-hydroxyethyl methacrylate
(HEMA) per mole of pyromellitic dianhydride (PMDA; 1,2,4,5-
benzene tetracarboxylic dianhydride)].
Subsequent work showed that the three-step protocol could
be reduced to a two-step procedure (Bowen et al., 1987) by
combining NPG with dilute nitric acid as the first step. NPG
in place of its adduct with glycidyl methacrylate is effective
in either protocol (Bowen, 1985). The second step currently
Received for publication May 15, 1990
Accepted for publication September 10, 1990
This work was supported, in part, by USPHS Research Grants
DE05129 and DE09322 to the American Dental Association Health
Foundation from the National Institutes of Health, National Institute
of Dental Research, and is part of the dental research program con-
ducted by NIST (National Institute of Standards and Technology, for-
merly National Bureau of Standards), in cooperation with the American
Dental Association Health Foundation. Certain commercial materials
and equipment are identified in this paper to specify the experimental
procedure. In no instance does such identification imply recommen-
dation or endorsement by NIST or the ADA Health Foundation or
that the material or equipment identified is necessarily the best avail-
able for that purpose. Contribution of the National Institute of Stan-
dards and Technology. Not subject to copyright.
remains the application of a PMDM solution to the tooth sur-
face immediately prior to placement of the dental resin or com-
posite.
In the synthesis of PMDM, the reaction product consists of
a mixture of para and meta regio-isomers, shown in Fig. 1
(Johnston and Bowen, 1987). Because of the extensive use of
these monomers, improvements in synthetic procedures were
investigated.
The overall purpose of this study was to assess the effect
of the catalyst structure on the reactivity of the HEMA with
the anhydride linkages of PMDA and the paralmeta isomer
ratio of the products. At present, thepara isomer is used most
extensively because it is a crystalline solid that is readily puri-
fied by re-crystallization. Efforts to crystallize the meta isomer
have not been successful. This study was therefore aimed at
producing a higher yield of the para PMDM monomer. The
theoretical yield of the para isomer from the addition reaction
is 50%. It was of interest to determine whether the para/meta
isomer ratio could be altered in favor of one or the other isomer
by changing the structure and basicity of the catalytic amine.
Control of the isomer ratio would allow more efficient pro-
duction of either isomer, assist industry in the current use of
thepara isomer, and permit planned studies of the comparative
effectiveness of the two isomers in adhesives and other appli-
cations.
Materials and methods.
The organic chemicals used in these experiments were all
purchased from Aldrich Chemical Co. (P.O. Box 355, Mil-
waukee, WI 53201). The 1,2,4,5-benzene tetracarboxylic
dianhydride (PMDA, 97%) was freshly sublimed before each
reaction (175°C, 26.7 Pa). The 2-hydroxyethyl methacrylate
(HEMA, 97%) was stored in a refrigerator over molecular
CH2 CH,
0+ 0O'-~r
H3C
0
OH + H
0 0~~~~~~
Amine
THF
Renux
I h
CH, 00 1 2
° ° JH,
H3C --g -( XH3
C, HOnIi..lyOO
pars-PMDM meta-PMDM
Fig. 1-The standardized reactions scheme used in this study. The two
isomers of PMDM, para and meta, differ only in the positioning of the
second HEMA group, but the para isomer is a solid. The meta isomer,
an oily product, resisted attempts to crystallize it. The temperature of the
reaction was 65°C, the boiling point of tetrahydrofuran, THF.
67
68 FARAIANI et al. J Dent Res January 1991

sieves (4A) to maintain a minimal water content. The 2,6-di- magnetic stirring bar, heating mantle, and magnetic..tirring
tert-butyl-4-methylphenol (BHT, 99 + % Gold Label) was used motor. The materials were added in the following order: PMDA,
directly from the bottle as a polymerization inhibitor. The amines 0.10 mole; HEMA, 0.22 mole; BHT (2,6-di-tert-butyl-4-
listed in the Table were used directly as received. methylphenol), 0.001 mole; THF (tetrahydrofuran), 100 mL;
The proton magnetic resonance ('H NMR) spectra of the and the amine catalysts listed in the Table, 0.02 mole. The
isomeric PMDM mixtures were recorded with a JEOL GSX solid PMDA went into solution slowly as the flask was heated.
270 MHz FT NMR spectrometer with deuterated acetone as The solution was refluxed for one h; then the reaction mixture
solvent and 0.3% tetramethylsilane (TMS) as internal standard. was assayed by 'H NMR analysis (Fig. 2). After being cooled,
Electronic integration of the spectra yielded the para/meta iso- the reaction mixture was diluted with diethyl ether (500 mL).
mer product ratios and the extent of conversion. Any PMDA that had been hydrolyzed to give the tetra-acid
The standard conditions of PMDM synthesis were the fol- was insoluble in the reaction mixture and was removed by
lowing: The reaction vessel was a 1000-mL, single-neck, round- vacuum filtration through a pad of celite on sand on a coarse-
bottom flask equipped with a drying-tube-capped condenser, fritted glass filter. The organic solution was transferred to a
TABLE
EFFECTS OF AMINE CATALYSTS ON REACTION PRODUCTS IN THE SYNTHESIS OF PMDM
Crude 'Dried' Final
Ratio Crude Ratio Melting Yield,
p/m/mono Mass p/m/mono Point, Para*
Amine pKa Ref PMDM % PMDM 0C % g
ALIPHATIC:
N,N-diisopropyl- 11.1 a 1/1.1 111 100/1 161-162 69 16.5
ethylamine 1/1.1 100/2 161-162 41 9.2
1/1 104 100/3 161-162 44 10.4
Hexamethylene- 6.3 b 1/1.0 100/3 161-162 46 10.8
tetramine 1/1 91 100/1 161-162 46 11.1
1/1 100/3 160-161 44 12.6
1,4-Diazabicyclo- 8.2, c 1/0.9 101 100/0 161-162 40 9.4
[2.2.2]-octane 4.2 1/0.6 100/3 160-162 t 4.6
1/0.6 71 100/6 156-158 38 9.1
Triethylamine 10.8 b 1/1 105 100/0 162 43 10.2
1/1 95 100/1 159-162 44 10.5
1/0.9 100/10 152-158 48 12.3
Tripropylamine 10.7 c 1/1.3 92 100/10 159-160 37 7.6
1/1.2/0.3 100/11 155-160 35 6.8
1/1 100/50 134-153 71 17.0
Tributylamine 9.9 b 1/0.8 100/13 157-161 32 8.6
1/0.6 100/6 153-159 35 10.8
BENZYLIC:
N,N-dimethyl- 8.9 b 1/0.9/0.2 100/4 160-162 46 10.4
benzylamine 1/1 98 100/3 160-162 44 10.6
1/0.7 100/9 150-157 45 12.6
Tribenzylamine 7.4 d 1/1.1/0.8 104 100/1 160-162 27 4.6
1/1/1.1 100/3/5 158-160 35 8.4
AROMATIC:
N,N,N',N'-tetra- 10.1, c 1/1.1/0.5 100/0 160-162 34 6.3
methyl-1,-4- 6.6 1/1 100/2 160-162 30 6.8
phenylenediamine
Pyridine 5.4 e 1/0.9 127 100/5 159-161 58 14.1
1/1.2/0.2 122 100/4 158-161 66 15.7
NN-dimethyl- 5.2 c 1/0.6/3.7 oil
aniline
Triphenylamine - -5.0 f 1/0.9/3.5 oil
The PMDM syntheses accomplished as part of this study are summarized here. Each amine is listed with the data pertaining to that amine. The pKa,
a reference for the pKa measurement, are in the first two columns after the amine. The para/meta/mono-addition product ratio is next. If there was no
mono-addition product (Fig. 6) observed in the reaction, then there are only two values in the ratio. The 'Dried' Crude Mass column listed the percent
yield recovered after work-up and 48-hour drying under an air stream. The final ratio reflects the integrations of the triply re-crystallized PMDM mixture,
and the next column discloses the melting point of the sample. The yield values are in percent and the grams of material collected after the reaction. Each
reaction that produced the product was repeated at least once and usually twice.
(a) Hall (1957) and Hull et al. (1969), (b) Perrin (1965), (c) Dean (1979), (d) Kudryavtseva et al. (1980), (e) Morrison and Boyd (1973), and (f) Arnet
et al. (1970).
*The yield of para PMDM is based on the weight recovered after the third re-crystallization divided by the amount of para PMDM in the crude mixture,
determined by the starting quantity of PMDA and the NMR ratios of the pyromellitic reaction products.
tSome product spilled before being weighed; mass recovered is presented in the Table, but yield percentage is unknown.
Note: The tribenzylamine was difficult to remove from the reaction mixture; therefore, the 35% and 8.6 g yield is after only two crystallizations. The
'dried' crude weight" represents the % yield of the original reaction product mixture after being washed and dried. Amounts over 100% probably
represent incomplete removal of solvents, catalyst, HEMA, etc., and amounts under 100% could represent incomplete reactions and/or losses during
washing.
Vol. 70 No. 1 BASE CATALYST STRUCTURE IN MONOMER SYNTHESIS
1000-mL separatory funnel for successive washes with 1 mol/
L aqueous HCl (3 x 150 mL), water (3 x 150 mL), and
saturated aqueous sodium chloride (3 x 150 mL). The re-
maining organic materials were poured into a large beaker
(2000 mL) and were concentrated by evaporation of the sol-
vents under a dry air stream. After 48 h, the crude solid was
again assayed by 1H NMR, dissolved completely in methanol
(130 mL), and warmed to 650C. After drop-wise addition of
water caused the solution to become cloudy, the mixture was
allowed to cool overnight at room temperature. The crude crys-
tals were filtered, subjected to 'H NMR analysis, and re-crys-
tallized from methanol/water mixtures two more times. After
the third re-crystallization, the final 'H NMR (Fig. 3) was
recorded, and the paralmeta isometer ratio was determined
(Table).
Results.
The para-PMDM isomer was a shiny, white crystalline ma-
terial with a melting-point range of 161-162TC. The melting-
point-range data for the PMDM products are listed in the Ta-
ble. A lower and wider melting-point range corresponded to a
less pure para PMDM sample.
The addition of the hydroxyl group of HEMA across the
anhydride linkages of PMDA did not occur within one h unless
a catalytic amount of tertiary amine was present. Ten mole
percent of amine was used, based on the HEMA, with lower
amine concentration giving an inadequate reaction rate under
these conditions. Twelve tertiary amines were tested (Fig. 4).
Six of these bases were aliphatic (pKa's 6.3 to 11.1), two were
benzyl amines (pKa's 7.4 and 8.9), and four were aromatic
(pKa's about - -5.0 to 10.1).
The amines in the Table are classified as aliphatic, benzylic,
or aromatic amines, listed in descending order of purity (high-
est m.p.) and yield of para PMDM. The percent yield of para
8.5 1.0 7.5 7 A 8.8 8.0 8. 58. 4.5 4.8 3. 28 LO
PPM
Fig. 2-The 1H NMR spectrum of the crude reaction mixture. The peak
at 8 0 ppm is due to the reference standard, tetramethylsilane. The peaks
near 8 8.0 and 8.3 ppm denote the presence of meta PMDM, and the peak
near 8 8.15 ppm denotes the presence of the para isomer. The two peaks
at 8 4.5 and 4.65 ppm are from the ethylene glycol chain of the reacted
HEMA molecules, and the smaller peaks at 8 3.85 and 4.25 ppm are due
to the same atoms of the unreacted HEMA. Since the HEMA was added
in 10% excess, integration of these four peaks provided the extent of
reaction. Protonated deuteroacetone solvent is represented by the complex
peak at 8 2.1 ppm. The inset shows the aromatic region in detail, with
the integrals over each peak indicating an almost 1:1 para:meta ratio. The
integrals over each signal peak represent the areas under the signal curves.
69
PMDM for each amine listed in the Table was calculated from
the para isomer component in the crude product (based on a
normalized ratio determined by 'H NMR analysis) and that
collected after the third re-crystallization.
The 'H NMR spectra were clearly diagnostic for the iso-
mers. The para isomer, 1,4-di[2-(2-methyl-2-propen-
oate)ethyl]-phthalate-2,5-dicarboxylic acid, had one resonance
(8 8.15 ppm) for the two magnetically equivalent hydrogen
atoms of the central aromatic ring (Fig. 3). The meta isomer,
1,3-di[2-(2-methyl-2-propenoate)ethyl]-phthalate-4,6-dicar-
boxylic acid, was not as rotationally symmetrical and had two
singlets corresponding to the two non-equivalent aromatic hy-
drogen atoms. These resonated upfield (8 8.0 ppm) and down-
field (8 8.30 ppm) with respect to the aromatic signal in the
desired para PMDM isomer, as shown in Fig. 2.
Discussion.
The data in the Table show the yields of para PMDM as
determined by the characteristics of the amines evaluated. The
electron density of the nitrogen controls electron donation, in-
dicated by pKa, higher values implying basicity and nitrogen's
tendency to accept a proton. Steric hindrance from physical
bulk of substituents around the nitrogen atom restricts the pos-
sible arrangements of the reactants in the transition state. Sol-
ubility and other characteristics influence ease of separation
during the purification steps.
The mechanism shown in Fig. 5 depicts the role of the amine
in the reaction of alcohols such as HEMA and anhydrides such
as PMDA (March, 1977; Butler and Gold, 1961; Fersht and
Jencks, 1970a, b). The amine reacts with the anhydride to yield
a zwitterionic intermediate (I). The positively charged N-tri-
substituted amide in structure I provides an excellent leaving
group for displacement by the nucleophilic oxygen of the HEMA.
This displacement of the amine by the alcohol results in the
zwitterionic protonated ester (II). Structure II in Fig. 5 can
8l.8 8.51.1l ulll
8.4....
8.6 8.5 ...4 8.3 8.2 8.t 8.0
7.8
7.9
1.8 18 8.5
I A-- A --i
8.5 8.0 7.5 7.0 5.5 5.0 3.5 2.5 2.0 1.5 85
6.5 6.0 4.5 4.0 3.0 1.0 0.0
PPM
Fig. 3-The spectrum of the purified para PMDM product. The strong
signals at 8 0 and near 2.1 ppm were due to the internal standard, tetra-
methylsilane, and the solvent, acetone, respectively. The two multiplets
near 8 4.5 ppm are due to the ethylene glycol chain, and the two singlets
at 8 5.6 and 6.1 ppm are due to the two terminal vinyl protons of the
methacrylate groups. The singlet near 8 8.15 is due to the two aromatic
protons of the central aromatic ring. The inset displays the aromatic region
in detail. The peaks due to the aromatic region of the meta PMDM are
absent, indicating purepara PMDM.
70 FARAHANI et al.
CH3
I :N 8 N:
(CH3CH N -CH2CH3
2
N, N-DI-IsopropyIethyIamIne
(DIIPEA)
Hexamethylene Tetramine
(HMTA)
(CH3CH2 N: (CH3CH2CH2 )mN:
Trlethylamine Tripropylamine
(TEA) (TPA)
CH3
CH2 -N
CH3
N, N-dimethylbenzylamlne Trlbenzylamine
Pyridine N, N-Dlmethylanillne
Fig. 4-The molecular structures of the amines tested in this investigation.
rapidly transfer the proton to the carboxylate group that had
been the other half of the anhydride linkage.
The mono-addition anhydride (III) can undergo the reaction
sequence again to yield either meta or para PMDM. Alterna-
tively, the addition of water can proceed slowly without amine
catalysis, and very rapidly with amine catalysis, to yield the
mono-addition product (Fig. 6). If the amine catalyst provided
sluggish support for the reaction, the aqueous work-up of the
reaction mixture yielded quantities of the mono-addition pro-
duct, in addition to the para and meta isomers. The desired
outcome was to have no mono-addition product present.
The rating of the amines was based on the final yield of the
para PMDM isomer and the ratio between the para and meta
PMDM after three re-crystallizations, as determined by 1H NMR
integration (Fig. 3; Table).
Yields of para PMDM depended on: (1) the para/meta ratio
in the crude reaction mixture (Table), (2) the addition of HEMA
to both of the anhydride linkages, and (3) the separation of all
components during washing and crystallization. When all of
these factors were favorable, the product melting-point range
was 161-162°C and the reaction yield was high.
All of the aliphatic amines were effective in promoting the
esterification, but tripropyl- and tributylamines provided mix-
tures from which it was difficult to extract the para PMDM
isomer (Table). This difficulty resulted in less pure products,
measured by 1H NMR and melting-point ranges.
The NN-dimethylbenzylamine was almost as effective as
the top three aliphatic amines. Tribenzylamine gave difficulties
during crystallization and separation of the para isomer, and
NMR spectra of the crude mixture indicated that a significant
J Dent Res January 1991
CH2CH2 \
:N-CH2CH2 N:
CH2CH2
1, 4-Dlazabicyclo-[2.2.2]-octane
(DABCO)
(CH3CH2CH2CH2 ) N:
Tributylamine
(TBA)
H3C /CH3
CH2 N N\
H3C CH3
N, N, N', N'-Tetramethyl-
phenylenedlamine
/CH3
N:
CH3 /3
Triphenylamine
quantity of the mono-addition product was formed; also, the
efficient removal of the catalyst by means of the acid washes
did not always occur.
Among the aromatic amines, N,N,N',N'-tetramethyl-1,4-
phenylene diamine was judged best, with pyridine in second
place due to difficulty in purification. With pyridine, the yield
ranked highest of all the amines assayed, but the melting-point
range did not.
The amine pKa measures the ability of the nitrogen to donate
electrons and stabilize the zwitterionic structure (I, Fig. 5).
The catalytically effective pKa range seemed to be between
5.4 (pyridine) and 11.06 (N, N-di-isopropylethylamine; Huenig
and Kiessel, 1958). When the pKa was lower, a high per-
centage of mono-addition product was formed.
With hexamethylenetetramine, DABCO, and pyridine, the
carbon atoms neighboring the nitrogen atoms are "tied back",
allowing the lone pair of electrons on each of the nitrogen
atoms to be directed outward, promoting rapid reaction with
the anhydride. The first two of these amines might most readily
partition into the aqueous HCl solutions, facilitating purifica-
tion.
Why tribenzylamine, NN-dimethylaniline, and triphenyl-
amine formed the large amounts of the mono-addition product
(Fig. 6) observed in the crude 1H NMR is not presently under-
stood. A salt of the amine and the acidic intermediate III of
Fig. 5 could effectively quench the further reaction. The aqueous
work-up of the reaction mixture would then hydrolyze the re-
maining anhydride linkages to yield the mono-addition product
predominantly. A study of the effects of amine and PMDA
stoichiometry might elucidate this and other relevant questions.
Vol. 70 No. 1 BASE CATALYST STRUCTURE IN MONOMER SYNTHESIS 71

0 0

R3N: 0 -*

I% 0

PMDA

Fig. 6-The molecular structure of the mono-addition product.

+0 0 <
I%
0 0
0 0°':
H 0
0
m fer. Reactions of Acetic Anhydride. Observation, Kinetics, Struc-
Fig. 5-The proposed mechanism of the reaction for synthesis of PMDM.
The "HOR" can represent most alcohols, but, in this case, it refers only
to HEMA. Structure III must undergo the reaction sequence one more
time to produce the isomers of PMDM.
Acknowledgment.
The financial support for the JEOL GSX 270 MHz FT NMR
spectrometer of the ADAHF Multinuclear NMR Facility pro-
vided by the Biomedical Research Support Shared Instrumen-
tation Grant iS10 RR03266 is gratefully acknowledged.
REFERENCES
ARNET, E.M.; QUIRK, R.P.; and BURKE, J.J. (1970): Weak Bases
in Strong Acids. III. Heats of Ionization of Amines in Fluorosul-
furic and Sulfuric Acids. A New General Basicity Scale, J Am
Chem Soc 92:1260-1266.
BOWEN, R.L. (1985): Bonding of Restorative Materials to Dentine:
the Present Status in the United States, Int Dent J 35:155-159.
BOWEN, R.L.; COBB, E.N.; and RAPSON, J.E. (1982): Adhesive
Bonding of Various Materials to Hard Tooth Tissues: Improve-
ment in Bond Strength to Dentin, J Dent Res 61:1070-1076.
BOWEN, R.L.; TUNG, M.S.; BLOSSER, R.L.; and ASMUSSEN,
E. (1987): Dentin and Enamel Bonding Agents, Int DentJ 37:158-
161.
H
BUTLER, A.R. and GOLD, V. (1961): The Hydrolysis of Acetic
Anhydride. Part VII. Catalysis by Pyridine and Methylpyridines
in Acetate Buffers, J Chem Soc 4362-4367.
R
DEAN, J.A., Ed. (1979): Lange's Handbook of Chemistry, 12th
ed., New York: McGraw-Hill, pp. 5-18 to 5-41.
H
FERSHT, A.R. and JENCKS, W.P. (1970a): The Acetylpyridium
Ion Intermediate in Pyridine-Catalyzed Hydrolysis and Acyl Trans-
ture-Reactivity Correlations and Effects of Concentrated Salt
Solutions, J Am Chem Soc 92:5432-5442.
FERSHT, A.R. and JENCKS, W.P. (1970b): Reactions of Nucleo-
philic Reagents with Acylating Agents of Extreme Reactivity and
Unreactivity. Correlation of Values for Attacking and Leaving
Group Variation, JAm Chem Soc 92:5442-5447.
HALL, H.K. (1957): Correlation of Base Strengths of Amines, JAm
Chem Soc 79:5441-5447.
HUENIG, S. and KIESSEL, M. (1958): Spezifische Protonacceptoren
als Hilfbasen bei Alkylierungs- und Dehydrohalogenierungsreak-
tion, Chem Ber 91:380-392.
HULL, L.A.; DAVIS, G.T.; ROSENBLATT, D.H.; and MANN,
C.K. (1969): Oxidations of Amines. VII. Chemical and Electro-
chemical Correlations, J Phys Chem 73:2142-2146.
JOHNSTON, A.D. and BOWEN, R.L. (1987): A Regioselective
Synthesis of Meta and Para Isomers of PMDM, J Dent Res 66:128,
Abstr. No. 172.
KUDRYAVTSEVA, L.A.; BEL'SKII, V.E.; and IVANOV, B.E.
(1980): Tautomerism in o-Aminomethyl Phenols, Ser Khin 4:803-
808.
MARCH, J. (1977): Advanced Organic Chemistry. Reactions,
Mechanisms, and Structure, New York: McGraw-Hill, pp. 348-
349.
MORRISON, R.T. and BOYD, R.N. (1973): Organic Chemistry,
3rd ed., Boston: Allyn and Bacon, p. 1016.
PERRIN, D.D. (1965): Dissociation Constants of Organic Bases in
Aqueous Solution, IUPAC Supplement, London: Butterworths,
pp. 28-118.