NIR-spectroscopic investigation of water sorption characteristics of dental resins and composites S. Venz* American Dental Association Health Foundation, Paffenbarger Research Center, National Institute of Standards and Technology, Gaithersburg, Maryland 20899

B. Dickens Polymers Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899 A near infrared (NIR) method using the 5200 cm-’ absorption of water has been employed to examine water absorbed in photopolymerized dental resins and composites in the form of 0.01-cm- to 0.15cm-thick specimens. The concentration, c [mol L-’1, of absorbed water in specimens of thickness t [cm] was calculated by means of Beer’s law, A = Ect. A is the NIR absorbance and E is the absorptivity of absorbed water. E depends on the environment of the water molecule and it is necessary to estimate F for water in each material. Water sorption was determined gravimetrically and correlated to the absorbance in the NIR spectrum. Once the

relationship between F and water content was known for a material, water sorption was determined rapidly on very thin specimens for faster equilibration. Where dissolution of the specimen occurred, the solubility behavior of the specimen was evaluated from a comparison of NIR and gravimetric measurements. The NIR absorptivity, E, of water absorbed in a polymeric medium was found to be inversely related to the degree of hydrophilicity and hydrogen bonding capability of the polymer. The presence of water clusters in a polyethylene oxide methacrylate polymer was inferred from convex-up curvature in the plot of E vs. water content.

IN T RODUCT ION

NIR spectrum of water

Transmission near-infrared (NIR) spectroscopy has been used in this study to determine the amount of absorbed water in several polymerized dental resins with different degrees of hydrophilicity and to provide information on the in situ environments of the absorbed water molecules. Because of its dipole moment, the OH group absorbs strongly in the IR region of the spectrum. The IR frequency of the 0 -H bond decreases with increasing hydrogen bonding for the stretching modes and increases for the bending mode. Thus, IR methods are useful to probe the environment of OH groups. ‘To whom correspondence should be addressed. This work was partially supported by a grant from the L.D. Caulk Division of Dentsply, Milford, DE 19962-0359. Journal of Biomedical Materials Research, Vol. 25, 1231-1248 (1991) Not subject to copyright in the United States. CCC 0021-9304/91/101231-18$4.00 Published by John Wiley & Sons, Inc.

VENZ AND DICKENS

1232

Most of the absorptions in the NIR region are from overtones and combination bands of hydrogenic vibrations.' An isolated water molecule has three modes of vibration: the symmetric stretch (v,),the bend (vz),and the antisymmetric stretch (v3).These three vibrations absorb strongly in the fundamental part of the IR spectrum (300 to 4000 cm-') at 3450 to 3700 cm-', 1595 to 1650 cm-', and 3550 to 3750 cm-', respectively. Overtones (such as 2vl) and combination bands (such as vz v3) of these bands are observed in the nearinfrared. Curcio and Petty' showed that liquid water absorbs at five frequencies in the NIR. Of these five, the absorption of the zl2 + v3 band at 1.94 pm (5160 cm-') is four times stronger for liquid water than the absorption of the 2vl band at 1.45 pm (6900 cm-') and 100 or more times stronger than the remaining NIR absorptions. Although the v2 v3 band contains the opposing effects of frequency shifts with hydrogen bonding from the antisymmetric stretch and the bending modes, it was used in this study because the 2vl band lies outside the wavelength range of the spectrophotometer. Water, alcohols, hydrogen-bonded ketones, and esters can all be differentiated by the v2 + v3 band and the concentration of water can be estimated in the presence of these other compounds. In some applications3of the NIR technique to determine water content, the water was extracted into a liquid to provide a known constant environment, and calibration curves were used to relate the infrared absorption to the water concentration. Water can be absorbed in great quantity without change in NIR absorptivity in a solvent such as methanol, which has a high density of OH groups. Because water in methanol is strongly hydrogen-bonded, the absorptivity of water in methanol is similar to that of pure water. Keyworth3 describes a NIR method for analysis of water in several polar hydrophilic solvents, including methanol, isopropanol, and several antifreezes. The method is rapid and provides a linear relationship between the weight percent of water in these materials and the NIR absorption. The results agree well with those from Karl Fischer titrations. Up to 60% water was determined in methanol. The position of the absorption maximum in the NIR spectrum of small quantities of water in solvents depends on the nature of the solvent. The maximum is at 1.89 pm (5291 cm-') for water in toluene, 1.91 pm (5234 cm-') for water in dioxane and polyethylene glycol, and 1.96 p m (5102 cm-') for water in ethylene glyc01.~ The vz + v3 combination band for water vapor is f ~ u n d at ~ - 1.875 ~ pm (5332 cm-'). That for liquid water is at 1.94 p m (5155 cm-'). These numbers suggest that water in toluene is isolated from other hydrogen bonding molecules, including other water molecules, whereas water in ethylene glycol is in a waterlike environment. The structure of liquid water was studied by Buijs and Choppin7using the absorption bands in the near IR region 1.1 to 1.3 pm. By measuring the absorbance of water as a function of temperature and by making some simplifying assumptions, they were able to obtain enough equations to estimate the absorptivities of nonbonded, singly bonded, and doubly bonded water

+

+

WATER IN POLYMERIZED DENTAL RESINS

1233

molecules and to estimate the abundances of these types of water molecules as a function of temperature. Similar studies on liquid water were carried out by Goldstein and Penner' over a wider temperature range with different absorption bands. In the 712 + 713 band, the absorption maxima of non-, singly, and doubly bonded water were found to be at 1.916 p m (5220 cm-I), 1.953 p m (5120 cm-'), and 2.000 pm (5000 cm-'), respectively. The abundances of the three kinds of bonded water are in fairly good agreement with the values of Buijs and Choppin7 with absorptivities at the above frequencies of 190, 55, and 0 for nonbonded water molecules; 110, 175, and 110 for singly bonded water molecules; and 41.5,44.6,46.4 for doubly bonded water molecules. The apparent absorptivity of water increases with temperature as the amount of hydrogen bonding decreases. Choppin and Downey9 discussed the frequency shifts in NIR absorbances of water in relatively nonpolar solvents and showed that the shifts can be related to the dielectric constant of the medium in those cases where the water is not hydrogen bonded. Water absorbed in CC14, CSz, CDC13, and C1CH2CH2C1was shown to be monomeric rather than hydrogen bonded. Extracting the water into an external medium thus has the advantage that the NIR absorptivity of water in the medium is known but does not allow the study of the environment of the water in the polymeric matrix, nor can incremental absorptions or desorptions easily be carried out. Also, components of the specimen other than water may dissolve in the extraction medium. To investigate the relationship between NIR absorptivity and water content and to shed some light on the state of the absorbed water, NIR analysis has been applied here to the determination of water in solid specimens. Several points are pertinent: (1)NIR spectra are much less crowded than spectra in the fundamental region because most observable NIR absorptions are vibration modes involving hydrogen. (2) NIR absorptions are much less intense than absorptions in the fundamental region. Because the attenuation of the beam is small, sample thickness is not critical. (3) The water band at 1.9 pm is one of the strongest in the NIR region. Thus, it was possible to detect water in specimens ranging from 0.5 mm to 2 mm in thickness, even in small amounts. For faster equilibration, films less than 100-pm thick can be used and spectra can be taken by means of a sandwich technique on a stack of these films. (4) The position and intensity of the NIR absorbance of a given amount of water depends on the matrix in which the water is absorbed.

Effect of water on dental composites Dental composites spend their service lives in aqueous environments and are expected to absorb some water. Absorbed water may counteract polymer-

VENZ AND DICKENS

1234

ization shrinkage by causing swelling and thus closing marginal gaps.1or11 However, water has been found to have deleterious effects. Prolonged storage of dental composites in water leads to reduction in transverse strength" and in some chemically cured Composite systems also leads to a reduction in modulus of elasticity. Soderholm et al.13 found that water has a significant weakening effect on the tensile strength of dental composites. The tensile strength could be partially recovered by redrying the specimens. Swelling of the p ~ l y m e r induces '~ stresses that, according to Kalachandra and Turner," may cause problems in bonded layers and at the dentinhestoration interface. Such plasticization of the resin rnatrix leads to softening of the composite surface and is expected to increase wear rates. The coupling agents at the fillerhesin interface were adversely affected by water in an accelerated aging test conducted by boiling dental composite specimens for 3 days." In that study, the transverse strength of the experimental composites was reduced by as much as 72% for the worst and 17% for the most effective silane coupling agent. Although silane agents can improve the performance of composites in water, they do not prevent leaching of the filler it~e1f.l~

MATERIALS

Resins and experimental composites of dental interest were selected for this study. The materials and their acronyms are shown in Table 1. The structures are given in Table 11. The compositions of the formulations are given in Table 111. No commercial restorative materials were evaluated. TABLE I Materials" and Acronyms Acronym PMMA HMDMA EBPADMA UDMA TEGDMA Bis-GMA Bis-GMA/TEGDMA NCO/TEGDMA CQ 4-EDAB MR/S C G/U/S MF/U/S

Name Poly(metliy1 methacrylate) Hexamethylene dimethacrylate Ethoxylated bisphenol A dimethacrylate Urethane dimethacrylate Triethyleneglycol dirnethacrylate 2,2-bis-[-4(-2-Hydroxy-3-methacryloxypropyloxy)-phenyl-]propane Mixture of 70% bis-GMA and 30% TEGDMA Mixture of a urethane derivative of bis-GMA with TEGDMA (proprietary) Camphorquinone Ethyl-4-dimethylaminobenzoate Milled Raysorb Glass, silanized Corning Glass unsilanized/silanized Microfine Silica, unsilanized/silanized Silane agent: 3-(trimethoxysily1)propyl methacrylate ~~

"The use of brand names in this publication is solely for the purpose of identifying the materials. No endorsement of said products by the National Institute of Standards and Technology or the American Dental Association Health Foundation is implied nor is their use here t o be used as a n indication of the efficacy of the identified products.

WATER IN POLYMERIZED DENTAL RESINS

1235

8I1 U

I

o=u

I

0

;c"

I I

U

o=u

&

3:

U

II U

6II U

0

U

I

I

u

U

8I

I -8 I

0

I

h

2

U

I

8I

0

U

I

0

I

8

I

I

f

u

8I

U

z

U

II

U

;c"

!

3:

I

0

3:

z

3:

U

I

I

3 U

o=u

I 0 I

0

I

I

0

I

s

8I 8I

3:

0

U

I 3: u I

0

uI

8 I

I

0

I

I

U $

8I 2 U

0

o=u

I I u I a I -8

r

8I

I

I

3:

U

U

0

U

I?

8I I

I

8

3:

0

I

O = U

3 U

I

o

z U

3:

I o=u u I 8 uI

I

3 U

o=u T.

8I f U

0

c

3:

U

I

0

I a I

I

I

I

I ;c" U I

I o=u I

z

2 U

2

2 U

o=u

U

I

8I

0

I

3:

s

I

o=u

o=u

rn

I o=u I 3:

II 0

rn

E

o=u

UI

V E N Z A N D DICKENS

1236

TABLE 111 Compositions of Formulations Acronym Unfilled Resins HMDMA TEGDMA EBPADMA UDMA Bis-GMA Bis-GMA/TEGDMA NCO/TEGDMA PMMA

Composition

Monomer CQ 4-EDAB

wt%

99.05 0.15 0.80

Prefabricated PMMA sheets

Filled Composite Resins NCO/T/M

NCO/TEGDMA CQ 4-EDAB MR, silanized

24.76 0.04 0.20 75.00

CG/U

NCO/TEGDMA CQ 4-EDAB CG, unsilanized

19.81 0.03 0.16 80.00

MF/U

NC O/T EGDMA CQ 4-EDAB MF, unsilanized

49.53 0.07 0.40 50.00

C G/S

NC O/T EGDMA CQ 4-EDAB CG, silanized

19.81 0.03 0.16 80.00

MF/S

NCO/TEGDMA CQ 4-EDAB MF, silanized

49.53 0.07 0.40 50.00

MET HODS

With the exception of PMMA, which was used in the form of prefabricated sheets, disk-shaped specimens of unfilled and filled polymers were prepared by photoinitiated polymerization of the formulations in Table 111. To avoid polymerization inhibition by oxygen, the specimens were prepared between glass slides separated by a spacer ring. Most specimens were cured by exposure for 1 min to an experimental light source with maximum intensity at 460 nm. The low-viscosity TEGDMA and HMDMA systems were cured by 10-min exposure to midday sunlight. Gravimetric measurements were carried out on a microgram balance. NIR spectra were measured at a resolution of 2 cm-’ on a Nicolet 7199 FT-IR Spectrometer equipped with a MCT (Hg-Cd-Te) detector and a Ge/KBr

WATER IN POLYMERIZED DENTAL RESINS

1237

(germanium on potassium bromide) beam splitter. Specimens were held between CaFz plates and were coated with paraffin oil to ensure good contact with the plates and to prevent loss of water. The specimens were cleaned afterward with hexane. To determine the absorptivity E for each polymer, five disk-shaped specimens were dried in a desiccator over silica gel at 37°C until no further change in weight and water absorbance was detected. They were then exposed to water either by immersion in water or exposure to 100% RH in a humidor at 37°C to mimic the oral environment. Since it appeared from a comparison of the NIR and gravimetric results that the specimens had lost residual monomer under these conditions, the water content of the disks was also monitored by NIR and gravimetry as the specimens were redried over silica gel. Changes in weight during drying were ascribed to changes in water concentration, c, and were related to changes in NIR absorbance, A, so that the absorptivity, E, of water in each matrix could be determined by means , t is the specimen thickness. The absorbance was of Beer’s law, A = E C ~ where estimated with the baseline technique or from the peak maximum in the difference spectrum between the dry and the water-containing spectra.

RESULTS A N D DISCUSSION

Typical spectra obtained in this study are displayed in Figure 1. These spectra are of 1-mm-thick NCO/TEGDMA specimens with various water concentrations. Figure 2 shows that water sorption can also be measured in thin polymer films of approximately 75-pm thickness sandwiched with paraffin

5 DAYS

1 DAY

DRY

INITIAL

7 65bO

6250

4

1

I

6000

5750

5500

52750

5060

47750

4200

Wavenumbers Figure 1. NIR spectra of water in NCO/TEGDMA polymer: Initial spectrum, after drying for 2 weeks in a desiccator, and after 1 and 5 days soaking in water.

VENZ AND DICKENS

1238 0.40

-7

E 0.30 i d d

P Lc 0.20 0

I

A

i

v1

P

4

0.10

0.00 1I 6500

1

6250

I

6000

I

I

5750

I

5500

5250

I

5000

I

4750

1

4500

Wavenumbers Figure 2. NIR spectrum of four stacked water-saturated 75-pm thick NCO/TEGDMA polymer films.

oil between the films. Figure 3 shows that spectra with little noise can be obtained from a 0.6-mm-thick composite resin specimen with 75 wt% fillers (NCO/T/M). The NIR absorptivity of absorbed water depends on the host polymer and whether the water is clustered in the host. Therefore, the water absorbance has to be determined experimentally over a range of sorbed water for each polymer host. The absorptivity is the slope of a plot of the water absorbance against the water concentration (Figures 4 through 9 and Table IV). The calculated values of E for water in the various polymers, assuming no curvature in the plots, are listed in the right hand column of Table IV. 90% confidence intervals are shown in Figures 4 through 9 as dotted lines. The errors involved in assuming a linear relationship between E and water content are 0.89

1

0.73

d d

P

k 0.57

0 v1

P

4

0.41

I

/

0.25 6500

I

6250

I

6000

I

5750

I

5500

1

5250

-

7

5000

4750

4500

Wavenumbers Figure 3. NIR spectrum of water in 0.6-mm-thick composite disk after 3 weeks soaking in water.

WATER IN POLYMERIZED DENTAL RESINS

2.5

1239

Absorbance/cm Sample Thlcknese

1

0

0.4

0.2

0.6

0.8

1

% Weight Change Slope = 2.15

Figure 4. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; 90% confidence intervals: dotted lines.

HMDMA Absorbance/crn SamDle Thickness 1

0.8

0.6

0.4

0.2

0

0

0.1

0.2

0.3

0.4

0.5

% Weight Change Slope

-

1.71

Figure 5. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; 90% confidence intervals: dotted lines.

0.6

VENZ AND DICKENS

1240

EBPADMA Absorbance/cm Sample Thickness 2.5 I

0

0.2

0.4

0.6

0.8

1

1.2

% Weight Change Slope = 1.66 Figure 6. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; 90% confidence intervals: dotted lines.

TEGDMA Absorbance/cm Sample Thlckness 81

0

1

2

3

4

5

% Weight Change Figure 7. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; second-degree polynomial: dotted line.

6

WATER IN POLYMERIZED DENTAL RESINS

1241

UDMA Absorbance/cm Sample Thlckness 3.5 I

0

0.5

1.5

1

2

2.5

3

% Weight Change Slope 1.10 Figure 8. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; 90% confidence intervals: dotted lines.

Bis-GMA

0

1

2

3

4

% Weight Change Slope

0

1.02

Figure 9. NIR absorbance vs water concentration in polymer disks. Linear regression: solid lines; 90% confidence intervals: dotted lines.

5

VENZ AND DICKENS

1242

Absorptivities

PMMA HMDMA EBPADMA NCO/TEGDMA TEGDMA Bis-GMA/TEGDMA UDMA Bis-GMA Liquid water

E

TABLE IV in Various Polymer Matrices Estimated from Straight Lines Fitted to Data SD

R

Density (g L-7

&

Slope

(L mo1-I cm-')

&

2.15 1.71 1.66 1.53 1.31 1.22 1.10 1.02 0.99

,999 .990 ,958 ,987 .995 .991 .990 .991 ,999

1190 1130 1180 1220 1220 1220 1220 1220 1000

3.25 2.72 2.53 2.26 1.93 1.80 1.62 1.50 0.99

0.04 0.03 0.09 0.08 0.04 0.04 0.05 0.05 0.02

Slope = absorbance/cm/wt% water; R = correlation coefficient; 100 density of matrix.

E

== slope x 18 x

generally less than the experimental errors. Because dried specimens with a small amount of residual water were used to define "zero" water absorbance and "dry" weight, the dry weights were adjusted as necessary for the lines in Figures 4 through 9 to pass through zero absorbance and zero % weight change. The F values in terms of L mol-I cm-I were obtained from the slopes in Figures 4 through 9 by applying a factor of (18 X 100/density of the polymer) as required to convert the water concentrations from wt% to mol/L. The water content of dental composites with these compositions can now be estimated using either Figures 4 through 9 as calibration charts or the absorptivities in Table IV. As mentioned in the Introduction, Goldstein and Penner8 showed that the absorptivity of the 1.9-pm band of liquid water increases with decreasing total hydrogen bonding of water. Goldstein and Penner used a multispacer technique and, following the method of Buijs and Choppin, were able to deconvolute the total hydrogen bonding into contributions from non-, singly, and doubly hydrogen-bonded water molecules by means of the profile of the absorbance band. No such deconvolution has been attempted here. Instead, the apparent absorptivity is used as an indication of the nature of the environment of the water molecules in the matrices. Thus, matrices such as bis-GMA that confer a low absorptivity on absorbed water are assumed to hydrogen bond to the absorbed water or to allow the water to cluster, depending on the availability of hydrogen bonds. Conversely, matrices such as PMMA, which confer a high absorptivity on the absorbed water molecules, are assumed not to hydrogen bond to the absorbed water molecules. Further, the absorbed water does not form clusters in these matrices, at least at the levels of absorbed water studied here. The plot of water absorbances vs. water content for TEGDMA (Fig. 7) appears to be slightly curved convex-up so that higher water contents give lower absorptivities (the slope of the curve). A second-degree polynomial was fitted to the absorbance/water content data to give a slope of

WATER IN POLYMERIZED DENTAL RESINS

1243

1.90 - 0.21 X (wt% water). A factor of 1.5 must be applied to the above coefficients to convert E from wt% to moles/liter. The E values in liter/mole/cm calculated from the slope equation range from 2.75 for very little absorbed water to 0.93 for 6 wt% water. An E value near 3 indicates isolated water molecules which are not acceptors in hydrogen bonds. A value of E = 0.99 was found (Table IV) for liquid water, which is extensively hydrogen-bonded. Thus, the range of E values found is consistent with increased hydrogen bonding of absorbed water as more water is absorbed in polymerized TEGDMA, suggesting that the water in this host preferentially forms clusters. All polymerized resins were found to contain water when stored in room air (Table V). The NIR absorptivity values of water in the various resin matrices are correlated with decreasing water content in the matrices at ambient conditions and after six months equilibration time (Table V). To a first approximation, the NIR absorptivity values provide estimates of the degree of hydrophilicity of each resin. Bis-GMA has two hydroxyl groups per molecule that can hydrogen bond to absorbed water. The polymerized bis-GMA resin absorbed the most water in molar terms and water in the bis-GMA resin has the lowest NIR absorptivity parameter, F. UDMA contains urethane groups that are hydrophilic and can hydrogen bond to absorbed water. Polymerized UDMA does not follow the same inverse correlation between E and absorbed water content as the other materials in Table V. The TEGDMA resin contains three ethylene oxide linkages that can accept hydrogen bonds but not donate them. Although EBPADMA contains approximately five ethoxylated groups, the presence of two phenyl rings in the molecule makes it more hydrophobic than TEGDMA. The only hydrophilic groups in HMDMA and PMMA are the ester groups. TABLE V Water Content of Polymers under Ambient Conditions and after 6 Months in a Humidor at 100% RH _

_

Ambient Conditions

MW of Monomer

Wt% H 2 0

SD

PMMA HMDMA EBPADMA TEGDMA Bis-GMA/TEGDMA UDMA Bis-GMA

100 278 580 286 444 470 512

0.5 0.2 0.3 1.0 1.1 0.8 1.2

0.04 0.01 0.02

100 278 580 286 444 470 512

1.80 0.52 0.97 5.45 3.75 2.67 3.60

Mol H20/Mol Monomer Unit

&

0.04

0.027 0.024 0.097 0.164 0.274 0.209 0.333

3.25 2.72 2.53 1.93 1.80 1.65 1.50

0.08 0.02 0.06 0.16 0.06 0.06 0.08

0.100 0.080 0.313 0.866 0.925 0.697 1.024

3.25 2.72 2.53 1.93 1.80 1.65 1.50

0.03

6 Months PMMA HMDMA EBPADMA TEGDMA Bis-GMA/TEGDMA UDMA Bis-GMA

VENZ AND DICKENS

1244

The water absorbance spectra (Fig. 10) show slight but distinct differences in the position of the absorbance maxima for the various polymers. Shifts to lower wavenumbers (higher wavelengths) correlate with increasing hydrogen bonding capacity of the resin matrix. The bis-GMA material forms hydrogen bonds as donor and acceptor. HMDMA and PMMA are in that sense the least capable of forming hydrogen bonds. The degree of hydrophilicity (roughly defined as the capability of forming hydrogen bonds), as estimated from the position of the maximum of the water absorbance peak, agrees with that estimated from the absorptivity in Figure 10. The finding" that the absorbance of water in PMMA is linear up to 1%water and then increases more slowly with increasing water content suggests this matrix also eventually absorbs enough water for self hydrogen-bonded water clusters to form. According to Gilbert, Pethrick, and Phillips," the interaction of water with the PMMA matrix is small but definite, but the major interaction is between water molecules themselves when enough water molecules have been absorbed. Here, PMMA was studied over the linear range, as shown by Figure 4. Figure 11 shows that water sorption is significantly less (Student's t test,p = 0.05) in a polymer made from the monomer system NCO/TEGDMA that was dried before polymerization than in a polymer made from the same monomers that were not predried. The experiment was carried out in triplicate. Drying of an already polymerized resin removes any water clusters and thus may lead to the formation of microvoids. Where microvoids exist and

PMMA 5243 cm-' HMDMA 5243 cm-'

: Q)

EBPADMA 5233 cm-'

cd P k 0 v)

TEGDMA 5220 cm-' UDMA 5213 cm-'

P

4

Bis-GMA 5214 cm-'

6500

I

6250

I

6000

I

5750

I

5500

I

5250

I

5000

I

4750

1

4500

Wavenumbers Figure 10. NIR water absorbance maxima of poIymers after 6 months in 100% RH at 37°C.

WATER I N POLYMERIZED DENTAL RESINS 3.5

1245

Weight %

O0 -

m 0 1 2

5

7

m 21

28

168

Days

-.-

Dried Reein

-I- Ambient Condition8

++Rerin from Humidor

Figure 11. Water sorption of NCO/TEGDMA: effect of water concentration in resin before polymerization on later water sorption in polymer.

sufficient water has been reabsorbed, water clusters would reform in these microvoids, as described by Gilbert et al.I9for PMMA. Polymers made from predried monomers are expected to contain fewer microvoids than polymers from nonpredried monomers and therefore reabsorb less water. Although water sorption by dental composite resins is thought to be mainly governed by the resin matrix, the filler in the composite can make a significant contribution to the amount of water absorbed. Figure 12 compares initial water desorption and subsequent sorption measurements on unfilled NCO/ TEGDMA specimens and composite samples made from 25% NCO/TEGDMA resin and 75% silanized milled Raysorb glass. The water concentration in the composite specimens, expressed in weight% water based on the resin content of the composite, is displayed in the upper two curves with asterisks for the NIR measurements and squares for the gravimetric measurements. The composite specimens are seen to have gained more water than the unfilled samples. These differences may be explained through increased stresses at the fillerhesin boundaries, which may progressively enhance diffusion'" and/or hydration of filler surfaces. Comparison of NIR data with gravimetric results for both the polymerized resin and polymerized composite specimens in Figure 12 shows greater water absorption from the NIR measurements than is obtained from the gravimetric measurements. This difference can be explained by assuming some dissolution of the specimens occurred during the sorption process. The amount of dissolution is estimated by subtraction of the NIR and gravimetric results to be 0.2%. A comparison of the weights of specimens dried before and after the sorption experiments gives a dissolution loss of 0.37% for the polymerized resin. The NIR measurements are less precise than the gravimetric measurements, as expected. Similar gravimetric measurements give

1246

VENZ AND DICKENS

Weight % Water Based On Resin Content 41 I

32-

i

* NIR Rerln

..+.

+

..n.. Gravlm. Compo8lts

NIR Compo8lts

Gravlm. Reiln

Figure 12. Comparison of NIR and gravimetric water sorption measurements; differences ascribed to dissolution of specimens.

0.18% dissolution loss for the filled resin. From the resin: filler weight ratio of 25% : 75%, this gives a dissolution of about 0.12% for the filler. The effect of silanization is shown in Figure 13, where experimental composites with unsilanized fillers (labeled /U) are compared to their silanized counterparts (labeled /S). Composites with silanized fillers pick up less water, showing, as expected, that the coupling agent between filler and organic matrix is another factor influencing water sorption in dental composite materials. Water sorption data from one week on show significant differences (Student’s t distribution, p = 0.05). The results from 1 and 2 days show either insignificant differences % Water Absorbed

M Td 6m

M 24 6m

CG/U

0wt/wt

CG/S

Composite

i d 26 6m

MF/U v/v Composite

I d 26 6m

MF/S v/v Resin

Figure 13. Effect of silanization and filler type on water sorption of composites.

WATER IN POLYMERIZED DENTAL RESINS

1247

or reversed trends, both of which may be due to diffusion kinetics, e.g., composites having lower diffusion coefficients than unfilled polymerized resins as reported by Marshall et a1.,2’ and microfine materials having much higher diffusion coefficients than conventional composites.21 CONCLUSIONS

Use of NIR spectroscopy to determine the water content of monomers, polymers and composite materials requires careful calibration of the NIR absorptivity of water in the particular material under investigation and over the range of expected water uptake. The NIR absorptivity of water in a particular medium is strongly dependent on the hydrogen bonding capability of the absorbing matrix and in some cases on the amount of water already absorbed. A number of NIR absorptivities of water in various polymerized resins of interest to dental technology have been determined. The NIR absorptivities of water are inversely related to the water sorption capacities and approach the absorptivity of 100% liquid water with increasing hydrogen bonding capability of the dental resin. Water sorption in dental resins is mostly determined by structural parameters of the resin but, to a lesser extent, also may be influenced by other factors, e.g., the filler, the coupling agents, and the existence of microvoids.

References 1.

2. 3.

4. 5. 6.

7. 8. 9.

10. 11.

W. Kaye, “Near-infrared Spectroscopy. I. Spectral identification and analytical applications,” Spectrochim. Acta, 6, 257-287 (1954). J. A. Curcio and C.C. Petty, “Near infrared absorption studies of liquid water,” J. Opt. Soc. Arner., 41, 302-304 (1951). D. A. Keyworth, “Determination of water by near infrared spectrophotometry,” Talanta, 8, 461-469 (1961). D. E. Burch, W. L. France, and D. Williams, ”Total absorptance of water vapor in the near infrared,” Appl. Opt., 2,585-589 (1963). D. M. Gates, R. F. Calfee, and D.W. Hansen, “Computed transmission spectra for 2.7-micron HzO band,” Appl. Opt., 2, 1117-1122 (1963). R. Goldstein, “Measurements of infrared absorption by water vapor at temperatures to 1000 K,“ J. Quant. Spectrosc. Radint. Transf., 4, 343-352 (1964). K. Buijs and G. R. Choppin, ”Near infrared studies of the structure of water. I. Pure water,” 1. Ckem. Pkys., 39, 2035-2041 (1963). R. Goldstein and S. S. Penner, ”The near-infrared absorption of liquid water at temperatures between 27 and 209 C,” J. Quant. Spectrosc. Radiat. Transf., 4, 441-451 (1964). G.R. Choppin and J.R. Downey, Jr., ”Near-infrared studies of the structure of water. IV. Water in relatively nonpolar solvents,” J. Chern. Phys., 56, 5899-5904 (1972). E. Asmussen and K. D. Jorgensen, “A microscopic investigation of the adaptation of some plastic filling materials to dental cavity walls,” Acta Odontol. Scand., 30, 70-75 (1972). E. K. Hansen, ”Visible light-cured composite resins: polymerization contraction, contraction pattern and hygroscopic expansion,” Scand. 1. Dent. Res., 90, 329-335 (1982).

VENZ AND DICKENS

1248 12.

13. 14. 15. 16. 17. 18. 19.

20. 21.

P. Kollmannsperger, “Biegefestigkeit von Composites nach Wasserlagerung von einem Tag bis 3 Monate,” Dtsch. zaknarztl. Z., 33,477-479 (1978). K-J. Soderholm and M. J. Roberts, ”Influence of water exposure on failure mechanisms of composite resins,” f . Dent. Res., 68, Spec. Issue Abst. No. 1296 (1989). K-J. Soderholm, ”Water sorption in a bis(GMA)/TEGDMA resin,” J. Biomed. Mater. Res., 18, 271-279 (1984). S. Kalachandra and D.T. Turner, ”Water sorption of polymethacrylate networks: Bis-GMA/TEGDM copolymers,” f . Biomed. Mater. Res., 21, 329-338 (1987). S. Venz and J. M. Antonucci, “Silanization techniques and modification of fillers for dental composite,” 1.Dent. Res., 65, Spec. Issue Abst. No. 191 (1986). K- J. Soderholm, ”Degradation of glass filler in experimental composites,” J. Dent. Res., 60, 1867-1875 (1981). E. R. S. Jones, ‘hsimple quartz infra red spectrometer for the determination of absorbed water in some polymers,” J. Sci. Instr., 30, 132-134 (1953). A. S. Gilbert, R. A. Pethrick, and D.W. Phillips, “Acousticrelaxation and infrared spectroscopic measurements of the plasticization of poly (methyl methylacrylate) by water,” f. AppE. Poolym. Sci., 21, 319-330 (1977). J. M. Marshall, G. P. Marshall, and R. F. Pinzelli, ”Diffusion of liquids into resins and composites,” Ckemteck, 426-432 (July 1983). M. Braden and R. L. Clarke, ”Water absorption characteristics of dental microfine composite filling materials. I. Proprietary materials,” Biomaterials, 5, 369-372 (1984).

Received August 1, 1990 Accepted March 12, 1991