Characterization of Low Molecular Weight Polyvinyl Acetate

The thermal and dilute solution properties of a series of oligomeric polyvinyl acetate (PVAc) ... polyvinyl acetate, even though many companies manufa...
0 downloads 0 Views 437KB Size
9.1 g of azine was stirred with 50 ml of benzene and 25 ml of saturated aqueous ammonia at 70 OC, where benzene was used to avoid crystallization of azine (mp 164 OC) and also to produce a difference in density between the organic and aqueous layers as the densities of benzophenone and azine are close to unity. Zinc and copper were extracted immediately and almost quantitatively. However, the extraction from actual reaction mixtures with similar composition was slow especially for copper. After six extractions with fresh aqueous ammonia (30 min per batch), about 14% Cu was found in the organic layer. Similar behavior was also observed in the extraction with aqueous ammonium chloride. Another troublesome problem in the use of aqueous ammonium chloride is the regeneration of recovered catalyst, which precipitates from the salt solution as a complex mass including ammonium chloride. For example, a precipitate, CuC12-2NH4Cl-2H20 or a mixture with CuC12.2H20, was obtained from cupric chloride and aqueous ammonium chloride. In the case of cuprous chloride with oxygen, much more complex behavior was observed. It is, therefore, undesirable to use aqueous ammonium chloride for recycling of the catalyst. Benzophenone azine is slightly soluble in ethanol and soluble in benzene; benzophenone is soluble and the catalysts are slightly soluble, in both solvents. Thus azine was isolated from the reaction mixture and the catalysts could be recycled as shown in the following example. (Zinc chloride is soluble in ethanol but precipitates from ethanol in the presence of ammonia or imine. Cuprous chloride is insoluble both in ethanol and benzene, but some slightly soluble complexes are formed in the presence of imine.) A mixture of ammonia and oxygen (NH3/02 = 4:l) was passed through a mixture containing 18.2 g of benzophenone,

0.3 g of zinc chloride, and 0.5 g of cuprous chloride a t 200 "C under 5 atm pressure at a rate of 10 1.kfor 2 h. This condition results in a 49% one-pass yield of azine. The reaction mixture was diluted with 200 ml of ethanol to precipitate azine and catalyst. The precipitated catalyst was separated from the azine by washing with 150 ml of benzene to dissolve the azine. Unreacted benzophenone, intermediate imine, and a part of the catalyst remained in ethanol solution. The ethanol was evaporated off under vacuum, further benzophenone was added to make up for loss, and the recovered catalyst powder was returned. The procedure was repeated four times, and 36.5 g of azine was obtained from a total charge of 45.7 g of benzophenone. The overall yield of azine was 81%.

Literature Cited Abendroth, H. J., Henrich, G., Angew. Chem., 71, 283 (1959). Abendroth, H. J., Henrich. G., German Patent, 1 082 889 (June 9. 1960). Chem. Eng. News, 5 2 , (Sept 16, 1974). Hayashi, H., Watanabe, T.,Nippon Kagaku Kaishi, 858 (1973a). Hayashi, H., Nishi, H., Abe, T..Nippon Kagaku Kaishi, 1392 (197313). Hayashi, H., Nishi, H., Kawasaki, K., Nippon Kagaku Kaishi, 1949 ( 1 9 7 3 ~ ) . Hayashi, H.. Kawasaki, K., Murata, T.,Chem. Lett., 89 (1974a). Hayashi, H., Kawasaki, K., Murata. T., Chem. Lett., 1079 (1974b). Hayashi. H., Yokose, K., Nippon Kagaku Kaishi, 2216 ( 1 9 7 4 ~ ) . Hayashi. H., Kawasaki, K., Okazaki, T., Nippon Kagaku Kaishi, 252 (1975a). Hayashi, H., Yuki Gosei Kagaku Kyokaishi, 33,451 (1975b). Hayashi, H., Kawasaki, K., Fujii, M., Kainoh. A., Okazaki, T.. J. Catal., 41, 367 (1976a). Hayashi, H., Kawasaki, K., Mori, M., Chem. Lett., 205 (1976b). Kiyoura. R., Hori, S., "Ryusan, Shosan, Ensan (Sulfuric Acid, Nitric Acid and Hydrochloric Acid)", p 161, Nikkan Kogyo Shinbun, 1965. Layer, R . W., Chem. Rev., 63, 489 (1963). Meyer, PI., British Patent, 843 587 (Aug 4, 1960). Meyer, R., Pillon, D., U S Patent, 2 870 206 (Jan 20, 1959). Ugine-Kuhlmann, British Patent, 1 358 389 (July 3, 1974).

Received for revieto April 8, 1976 Accepted J u n e 12,1976

Characterization of Low Molecular Weight Polyvinyl Acetate Ronald P. D'Amelia' and Harry Jacin Life Savers, Inc., Research and Development Division, Port Chester, New York 10573

The thermal and dilute solution properties of a series of oligomeric polyvinyl acetate (PVAc) resins were studied using a variety of techniques. These contribute to an understandingof the masticatory behavhr of chewing gum. The dilute solution study provided information on intrinsic viscosity, number (M,) and weight (Mw)average molecular weight, and molecular weight distribution. The glass transition temperature ( T,) and thermal stability were determined. The samples were also analyzed, using standard titrimetric techniques for moisture and residual acetic acid. The Mark-Houwink expression for oligomeric PVAc in toluene was derived. The T, generally increased from 15.0 to 28.4 O C as the molecular weight increased, but no definite relationship could be established. The thermal stability of the oligomeric PVAc resins was found to be lower in 02 than in N2. Acetic acid was the major decomposition product in both atmospheres.

Introduction The literature is rich in articles describing the technology, properties, polymerization, and uses of predominately high molecular weight polyvinyl acetate (PVAc) (Powers, 1943; Schildknecht, 1952; Skeist, 1962). However, very little work has been reported for oligomeric or low molecular weight polyvinyl acetate, even though many companies manufacture and use this type of resin. The low molecular weight resin is used extensively in the manufacture of chewing gum. It is mixed and blended with

other ingredients such as waxes, fillers, elastomers, and other resins to make what is called by the industry "chewing gum base". The chewing gum base is the residue left in the mouth after all the H20 soluble components of chewing gum are extracted. Since polyvinyl acetate is generally the major constituent of the chewing gum base, its properties are a major contributory factor to the textural properties usually attributed to chewing gum. For example, if the PVAc level is kept constant, and all other ingredients are the same, a high molecular weight Ind. Eng. Chem., Prod. Res. Dev., Vol. 15, No. 4, 1976

303

Table I. Low Molecular Weight Polyvinyl Acetate Samples Sample

Origin

Color

A B C D E F G

USA. U.S.A. U.S.A.

Pale yellow Water white Water white Water white Water white Pale yellow Yellow tinge

Japan France Mexico

U.S.A.

resin will give a hard chew while a low molecular weight sample will give a soft chew. In this paper we would like to report on the methods used and the results obtained on examining the thermal and solution properties of a series of oligomeric PVAc resins.

Experimental Section Materials. The PVAc samples included in this study are all commercial resins and are listed in Table I. About 60% of the samples studied came from U S . manufacturers while the other 40% came from foreign sources. The sample colors ranged from clear water white to pale yellow. All samples exhibited a noticeable “cold flow” property. The solvents used in the dilute solution studies were all ACS Reagent Grade quality. Vapor Pressure Osmometry. The number average mowas determined from data obtained on lecular weight (gn) a Hewlett-Packard vapor pressure osmometer (VPO) Model 302 B. The instrument was calibrated with ultrapure biphenyl using the constant “K” method (Hewlett-PackardInc., 1962). The calibration and sample measurements were performed at 37.5 “C in toluene using a 2 min reading interval. The molar constant K obtained was 13.33 X 103 uV-l./mol. This constant was used to determine the of the PVAc samples according to the following relationship: = K/(u/c),,o. The value of ( U / C ) , + ~ for the samples was determined by the same procedure used to evaluate the molar constant K . Gel Permeation Chromatography. Gel permeation chromatography data were obtained using a Waters GPC/ ALC 301 instrument. The instrument was equipped with dual detectors, a differential refractometer, and a uv photometer, operating in series during the experiment. The differential uv detector was operated a t a fixed wavelength of 254 nm. The measurements were made at ambient temperatures with a 2 ml/min flow rate. Tetrahydrofuran, stabilized with 0.025% butylated hydroxytoluene was used as the carrier solvent. Six 3-ft stainless steel columns packed with Styragel having exclusion limits of 2 X lo3-7 X lo2A, 7 X 102-3.5 X lo28,, 100-80 A, 100-80 8,, 80-50 A, 80-50 8, were used for the separation. A standard siphon having a volume of 5.3 ml was employed as a counter for the elution volume. The column set was calibrated using narrow molecular weight distribution polystyrene standards obtained from Waters Associate and the Pressure Chemical Co. Polypropylene glycol standards from Waters Associate were also used for the calibration. A plot of log molecular weight vs. elution count gave a calibration curve as seen-in Figure 1.The number average and weight average (M,) molecular weights of the samples were determined using standard methods (Cues, 1966). All solutions were filtered, prior to injection, through a Waters Filtration Assembly. Capillary Viscometry. Viscometry measurements were made in toluene at 25.2 “C using a calibrated Cannon-Ubbelohde low shear dilution viscometer (size 50). The toluene efflux time was greater than 100 s; therefore no kinetic energy

a,, a,

(a,)

304

Ind. Eng. Chem., Prod. Res. Dev., Vol. 15, No. 4, 1976

18

19

20

2i

2i

23

2i

25

26

QPC Elution Count

Figure 1. Gel permeation chromatograph calibration curve.

corrections were made on the observed data. The temperature was controlled using a Cannon constant-temperature bath. Triplicate readings were taken a t each concentration using a stopwatch accurate to 0.2 s. Efflux times were reproducible to 0.2 s. The intrinsic viscosity [7]was determined by extrapolating to zero concentration a plot of reduced specific viscosity (asp/,) vs. concentration, c , (g/dl). Differential Scanning Calorimetry. A Perkin-Elmer differential scanning calorimeter (DSC) Model 1B was used to determine the glass transition temperatures (T,) of the PVAc resins. All samples were subjected to thermal cycling to remove any thermal history. A programmed heating rate of 10 “C/min was used for both calibration and measurement. A temperature calibration was established between the actual sample temperature and the dial reading using high purity standards of known melting point. The glass transition temperature was taken at the point of deflection from the apparent baseline. Thermogravimetry.The thermal stabilities of the samples were studied by the simultaneous operation of the PerkinElmer DSC-1B containing a thermal conductivity gas detector (EGD) and a Perkin-Elmer thermal gravimetric system-1 (TSG) containing a Cahn RG electrobalance. The DSC temperature programming unit was used to provide power for both the DSC and TGS furnaces. A temperature calibration was performed which provided temperature correspondence between the DSC and TGS units. The samples were studied under N2 and 0 2 atmospheres using a 10 “C/min scanning rate. Residuals. Moisture and acetic acid were determined by standard published methods. These analyses were routinely performed to support the manufacturers’ specification data. Computational analysis of the VPO, GPC, and instrinsic viscosity data was done on a Wang Model 700 programmed advance calculator.

Table 11. The Number Average Molecular Weight of the PVAc Samples as Determined by Vapor Pressure Osmometry Samole

Q"

Source

A B C D E F G

2720 5362 6318 5554 5127 5127 4096

U.S.A. U.S.A. USA. Japan France Mexico U.S.A.

4 5 6 7 8 911llYU14E16171819X2122~2425262728

GPC Elution Count

Figure 2. A typical GPC chromatogram. Simultaneousrecording of differential R.I. and uv detectors.

Table 111. A Summary of the GPC Results

A B C D

E F G

No. R.I.

Correspond.

1 1 1 1 1 1 1

Yes No No No No

Yes Yes

an

an

2839 5310 5980 5586 5178 4688 3902

2720 5362 6318 5554 5127 5127 4096

7465 11 303

12413 11923 11346 10557 8613

2.6 2.1 2.1 2.1 2.2 2.3 2.2

Table IV. The Intrinsic Viscosity of the PVAc Samples Sample

[vl, dllg

A B

0.066 0.103 0.112 0.103

C D

'E F

0.100

5

2.5

0.100 0.086

G

10

?5

Log Molecular Weight (X10 - 3 )

Figure 3. The Mark-Houwink intrinsic viscosity-molecular weight relationship. Table V. Summary of the Glass Transition Temperature and Residual Data

T,, "C

Water,

Sample

%

Acetic acid, %

A B C D E F G

15.0 22.8 28.4 26.6 27.4 20.0 26.4

0.94 1.32 1.31 0.90

0.079 0.009 0.015 0.005

-

-

-

-

0.34

0.021

Results a n d Discussion The number average molecular weights of the samples, as determined by vapor pressure osmometry, are summarized in Table 11. Straight lines with positive slopes were obtained in the u I c vs. c plots for each of the PVAc resins studied. The positive slope indicated that toluene can be considered a good solvent for PVAc a t this temperature and for this molecular weight range. The for most of the samples was between 5000 and 6000. However, sample A was significantly lower. A typical GPC chromatogram is shown in Figure 2. The chromatogram shows the simultaneous recording of the output from the differential refractometer (AR.1.) and the uv photometer. The chromatograms of all the PVAc resins exhibited slightly skewed R.I. peaks, similar to the R.I. peak

E f f1 u e n t

N2

4 1

-

Atmosphere

an

256

287

318

Temperature

354

382

393

438

('c)

Figure 4. A typical thermal stability profile. Ind. Eng. Chem., Prod. Res. Dev., Vol. 15, No. 4, 1976

305

90.

100,

:i

1 = Sample A 2 = Sample B 3 = Sample G

10

1

-

201

220

239

237

.267

286

305

Temperature

324

333

354

3b9

599

("C)

Figure 6. Thermograms of several PVAc samples in N2 atmosphere.

shown in Figure 2. A summary of the GPC data, including the Enand E, number of components detected by the differential refractometer and uv photometer, and the polydispersity index for each sample is given in Table 111.The GPC data again reveal that sample A has a much lower molecular weight than the other samples. Anothernoteworthy fact is the excellent correspondence between the M , determined by GPC with the R, determined by VPO. The close agreement between these two different methods attest to the reliability of the GPC calibration and data analysis. Several samples had corresponding uv and R.I. peaks. This correspondence indicates the presence of uv absorbing components, most likely carbon-carbon double bonds, along the main chain of the polymer. The samples that had uv absorption also had a yellow color. All samples have a polydispersity index slightly greater than 2. The distribution of these samples can generally be characterized as having a maximum occurring near M , and having an irregular contour on the low molecular weight side of the distribution curve. The intrinsic viscosity of the PVAc samples in toluene a t 25.2 "C is shown in Table IV. Again the intrinsic viscosity data corroborates the VPO and GPC data. For example, sample A has the lowest intrinsic viscosity, and as seen before the lowest molecular weight. The [VI-molecular weight relationships, 306

Ind. Eng. Chem., Prod. Res. Dev., Vol. 15, No. 4, 1976

using the number average determined by VPO and the weight average obtained from GPC, are plotted in Figure 3. The Mark-Houwink intrinsic viscosity-molecular weight expressions, using the slope and intercepts obtained from Figure 3, are given below.

The exponent in the previous equations indicates that PVAc favors an extended conformation in toluene. It also shows that there exists a favorable solvent-solute interaction between the polymer and toluene. This fact is also supported by the positive slopes in the VPO plots. The value of "a" reported in the literature for PVAc (Brandrup and Immergut, 1967) in toluene at 25 "C in the molecular weight range of 4-5 X lo4 mwu is 0.53. Apparently toluene is a good solvent for low molecular weight polyvinyl acetate but is a poor solvent for the high molecular weights. The results for the glass transition temperature and residuals are seen on Table V. It can be seen that the glass transition temperature of A is considerably lower than the other samples. This difference can be attributed to the lower molecular weight of A. The glass transition temperature generally

increased with increasing molecular weight but no definite relationship could be established. The glass transition temperatures of all the samples are around room temperature, supporting the cold flow behavior observed for these resins on standing. The residual data shows that sample A has a large amount of residual acetic acid. This may be caused by its low molecular weight. A typical thermal stability profile is seen in Figure 4. It shows the simultaneous recording of the weight loss and effluent gas as the sample temperature is increased. All samples started to loss weight around 287 "C. From 287 to 375', approximately 70%of the sample was degraded with acetic acid being the major decomposition product. Based on the repeat unit weight of a linear PVAc molecule, 68% of the sample is made up of the acetate moiety. Therefore the rapid weight loss between 287 and 375" corresponds to the complete deacetylation of the oligomer. These results confirm the data reported by Bataille and Van (1975).Since the residue is highly colored and hence presumably conjugated, the decomposition reaction can be visualized according to the mechanism proposed by Grassie (1952,1953). The percent weight loss of sample A in N2 and 0 2 is shown in Figure 5. It shows that sample A is less thermally stable in 0 2 than in N2. Oxygen accelerated the rate of degradation possibly by producing peroxides or hydroperoxides which when dissociated into free radicals produce an accelerated decomposition. This same phenomenon was observed for all samples.

The percent weight loss thermograms of samples A, B, and

G are seen in Figure 6. The thermogram of sample A shows it to be less stable than samples B and G. This lower stability of A can probably be attributed to its lower molecular weight and broader distribution.

Acknowledgment The authors are grateful to the Life Savers Corporation for permission to publish this paper. Literature Cited Bataille, P., Van, B. T., J. TherrnalAnal., 8, 141 (1975). Brandrup. J.. Immergut, E. ti., "Polymer Handbook", 1st ed, pp IV-17, Interscience, New York, N.Y.. 1967. Cazes, J., J. Chern. Educ., 43, A567 (1966). Grassie. N., Trans. Faraday Soc., 48, 379 (1952); 49, 835 (1953). Hewlett-Packard Inc., Vapor Pressure Osmometer, Technical Bulletin #18 (1962). Powers, P. O., "Synthetic Resins and Rubbers", pp 115-123, Wiley, New York, N.Y.. 1943. Schildknecht, C. A., "Vinyl and Related Polymers", pp 323-384, Wiley, New York, N.Y., 1952. Skeist. I., "Handbook of Adhesives", pp 349-382, Reinhold. New York, N.Y., 1962.

Receiued for review July 16,1976 Accepted August 23, 1976 Presented at the Division of Organic Coatings and Plastics Chemistry, 169th National Meeting of the American Chemical Society, Philadelphia, Pa., 1975.

Ind. Eng. Chem., Prod. Res. Dev., Vol. 15, No. 4, 1976

307