Heat Capacities of Polyethylene IV. High Molecular Weight Linear Polyethylene

A high molecular weight linear polyethylene sample has been studied by adiabatic calorimetry from 10 to 380 K. Two broad temperature regions of unusual spontaneous temperature drift have been observed. The phenomena occurring around 240 K are similar to that observed in other polyethylene samples studied in this series, and are presumed to be caused by the relaxational processes in the amorphous phase. The weak exothermic behavior occurring around 160 K is presumed to be caused by the stabilization of the quenched sample.


Introduction
The heat capacity be havior of a sample of high mol ec ular we ight linear polyethylene has bee n studied rece ntly by Beatty and Karasz [1 , 2].' Hi gh molec ular weight polye thyle ne is more difficult to crys tallize than lowe r molec ular weight polyethyle ne. In creased amorphou s co nte nt s hould provide a stronger indication of the exis te nce of a glass tran sition in partially crystalline line ar polyethylene. A region of rather abrupt c han ge in the heat capacity differences between that high molec ular weight sa mple and a linear polyethylene sample studied by Dainton et aL [3] was noted. For the quenched hi gh molec ular weight sample, there was an 8 p ercent in crease in the heat capacity occurring in the temperature region 130 to 190 K. For the annealed sample, there was about 5 percent increase. At higher te mperatures a bend in the heat capacity curve was noted similar to Dainton's data. The region near 150 K was assigned as the glas s transition te mperat ure of linear polyet hylene. Spontaneous temperature drifts under adiabatic conditions, indicative of relaxational processes, were not noted in detaiL A similar high molecular weight lin ear polyethylene sample from the same manufac turer was s tudied in this laboratory in the te mperature range from 10 to 380 K. De tailed s pon tan eous temperature drifts were carefully observed on th e sample under various th ermal treatments. The results and co nclusions from thi s I Fi gures in brac ke ts ind ica te th e lit e rature re fere nces at the e nd of thi s paper. 51 s tudy are so mewhat different from the previously mentioned lite rature.

Calorimetry
H eat capacity meas ure me nts on the hi gh molec ular weight lin ear polyethylene sample were made with the same vacuum adiabatic calorim eter [4] that was used in th e previous studi es [5,6] on polyethylene samples derived from Standard Reference Materials 1475 and 1476.

Materials
A sample of high molecular weight polye thyle ne was furnished by Dr. R. J. Schaffhauser of the Plastics Division, Allied Chemical Corp.,2 bearing the designation 260-100. This resin was at one time also designated as AC8X, as was Beatty's sample [2]. It was produced by Ziegler-type vapor phase polymerization. Its molecular weight is estimated by the manufacturer as 8-13 X 1()6. It has an ash content of 0.04 percent and contains no additives. The molecular weights are estimated in the order of 2.7-3 X 10 6 by both viscometry and light scattering measurements from the Polymer Characterization Section of the National Bureau of Standards. Whether the d. iscrepancy in th e molecular weights is caused by th e degradation of th e polymer 1 Co mme rc ial mat erial s a re ide nti fie d in thi s pa per to adequately spec ify the expe rime ntal procedure. Such identifi ca tion does not impl y recomme nd a tion or e nd orse me nt by th e Na li onal Bureau of Standa rd s. at high temperatures or by differences in the procedures to estimate the molecular weight was not d etermined.
The sample was received in the form of a powder. Attempts to fuse the powder together under its own weight in vacuum at ISO °C were not successful due to the high viscosity of the melt. The heating produced 52 a sintered product. In order to avoid degradation of the high molecular weight material at high pressure and temperature, molding was not tried_ The material was pressed into pellets of about 1.27 cm in diameter and about I cm in height in a hydraulic press at room temperature_ 73.891 g (mass in vacuum) of the material was loaded in the sample container for at 3.8 cm Hg (5.1 kPa) at room te mperature was sealed in to improve thermal conduction within the s ample containe r.

Results
The results of the heat cap acity meas ureme nts are tabulated in table 1 and shown graphic ally in fi gure l. Table 1 is separa ted into vario us sec tions accordin g to the the rmal tre atme nt th e sa mple received . Within the section , the data are arr anged in the order of increasing initial te mpe rature of a series of heat capacity m eas ure ments. Th e seri es are numbe red in chronological se que nce throughout table 1 in orde r to facilita te the tracin g of the therm al history of the s ample . Quenc hed samples were produced by admitting helium gas to the cryostat while the assembly was s ubmerged in liquid nitrogen. A cooling rate in the order of 4-5 K min -1 was achieved. Slow-cooling (rate annealing) rates of 0.5-1 K h -1 were often used in conjunction with isothermal annealing (soak annealing) at a partic ular temperature for a period of days.
The temperature increment for a heat capacity determination may be infe rred from the differences in the mean temperatures of the adjacent determinations within the series. Curvature corrections have been added to correct for the effect of the finite temperature increment of a determination. The precision of the measurement above 25 K is in the order of 0.05 per· cent. Below 25 K, the precision gradually changes to about 1 percent at 5 K. The accuracy over most of the temperature range of the measurement is comparable to the precision as indicate d by heat capacity measure· ments on a Calorimetry Conference standard sample of sapphire [4].

Discussion
The crystallinity of 45 percent for this powdery sample was estimated from heat of fusion measurements using a dynamic scanning calorimeter. Heat of fusion of a . pressure crystallized linear polyethylene (96% crystallinity) [6] was used as the reference.
Although the high molecular weight sample has lower crystallinity than the SRM 1475 in the condition as received (71 % crystallinity) [5], it shows a somewhat lower heat capacity over the temperature range 90 to 310 K. Only at lower and higher temperatures does this sample exhibit the expected higher heat capacity. The heat capacity of this sample is very similar to that of a sample of SRM 1475 slowly cooled from the meit (88% crystallinity) [6] in the temperature range 100 to 220 K.
In a plot of . heat capacity differences, figure 2, between the heat capacity of this sample and that extrapolated for crystalline linear polyethylene [6], the changes in the 150 and 240 K regions are not as large as that for the 71 percent crystallinity sample. Beatty's data [2] are also included in figure 2. A relatively rapid heat capacity increase over that of the crystalline linear polyethylene occurs in the tempera-ture region of 120 to 170 K. Above 200 K, the scattering of the data causes diffic ulties in th e conclusion of either the existence or non exi ste nce of a subtle he at capacity change. DT A or DSC studi es on some other high molecular weight polye thylen e samples sug· gested the occurrence of a ste p change in the he at capacity around 150 K [7 , 8] and indicated a s Uf· prisingly constant heat capacity difference be tween . . . . 6 the high molecular weight sample and a high crystal· linity sample [8].
The lack of a strong glass·like feature in the heat cap acity behavior of this high molecular weight rna· te ri al is in parallel to a recent thermal expansivity study on the same ma terials. Low temperature thermal expansions of polye thyle ne samples have been studied on strips of co mpression molded thin film s [9]. The  samples included the two Standard Reference Materials 1475 and 1476, used in a previous calorimetric study [5], and the high molecular weight sample used in the present calorimetric work. In all three samples, double peaks in the derivatives of the thermal expansion coefficients at about 110 and 150 K were observed. These peaks are stronger in linear polyethylene than in branched polyethylene. Weaker peaks were observed in the temperature range 180 to 200 K. Above 200 K, there were strong peaks in the branched polyethylene, a weak one in the high 100 200 decay of the temperature of the thermomete r until a thermal equilibrium is reached between the sample and the sample container assembly. The decay con· stant depends on the construction of the sample container, the thermal conductivity of the sample and the amount of helium gas sealed in the sample container. The decay constant is about 50 s for this sample container [12] at high temperatures and less at lower temperatures. Therefore up to 10 to 15 minutes may be required for the sample to reach a temperature distribution which is uniform to within 10-4 K 300 400 T. K Although the appearance of a discontinuity in the heat capacity behavior is perhaps a necessary condition for a glass transition, it is not a sufficient condition, A broad distribution in the relaxation times may cause the glass transformation to occur over a wide temperature range, thus making a weak heat capacity discontinuity in partially crystalline material even more difficult to assess. Since the glass transition is kinetic in nature, relaxational phenomena should also be observed in the glass transition region. One highly sensitive method to detect the thermal relaxation is the method of observing the spontaneous temperature changes of the sample under adiabatic conditions [10]. The sign and the magnitude of the temperature drift as a function of time and temperature may be correlated with different thermal treatments and histories. Both the discontinuity in the heat capacity behavior and the adiabatic temp~rature drift are commonly observed in glass transition regions of various materials and also in glass-like transitions in crystals [11].
The drift observations are shown in figure 3. These are the drift rates observed at 20 minutes after the heater energy has been turned off. Because of the geometric arrangement of various components in the sample container [4], the temperature of the ther· mometer and the heater is slightly higher than that of the sample during the heating period. Therefore, when the energy is turned off, there is an exponential The drift behavior in the 240 K region for the high molecular weight material occurs in similar tempera· ture range and magnitude as that observed for the asreceived SRM 1475 [6,10] and is presumed to be caused by the relaxational behavior in the amorphous phase. For the quenched sample, i. e. , cooling rate is greater than the rate of heat capacity determination of 5-10 K h -1, the drift rate is generally positive (exothermic) with a broad peak around 230 K There is also a weaker peak at around 160 K When the sample was heated to 370 K for the first time, large exothermic behavior showed up above 330 K, indicating the onset of further crystallization processes. For the sample either slowcooled continuously to 100 K or annealed at 225 and 140 K, there is an endothermic peak only around 245 K. No endothermic behavior was observed at temperatures below 210 K for the sample either slow cooled or annealed. Thus the we aker exothermic peak occurring around 160 K for quenched samples is presumed due to some stabilization processes such as the relief of strain.
In the broad temperature region near 240 K, the drift behavior as observed in figure 3 and in other linear polyethylene samples [6,10] seems to indicate the existence of double peaks. Special thermal treatments were performed in attem,pts to resolve these peaks. Figure 4 shows the results of the drift observations for two combinations of quenching, annealing and slow cooling treatments. The sample was first annealed at 250 K for 1 day and then quenched from 250 K Upon heatin g th e sample, it s howed exothermi c be havior, similar to that of a quenched sa mple, at te mperatures below 250 K. Above 250 K, e ndoth ermic e ffects simi· lar to an annealed sample were seen.
The sample was quickly cooled (5 K min -1) to 225 K and held there for 4 days, until the drift decreased from an initial value of greater than 1 mK min -1 to less than 0.05 mK min -I. It was then slowly cooled (2.5 K h -I ) to 143 K and held there for 2 days. An e ndoth er mic peak was observed at about 245 K. Above 245 K, a small exothermi c pea k appeared near 270 K. Ap· parently annealing at 225 K did not relax the configura· tions associated with the higher temperature peak.
The two regions around 240 and 270 K may be caused by different types of amorphous relaxation mechanisms, such as from the loose ends of the poly· mer molecules or from the molecular segments with both ends confined in crystalline regions. However, th e mechanisms responsible for the relaxational behavior observed here have not been determined.
The author wishes to thank R. G. Christensen for the molecular weight determinations and C. H. Pearson for the assistance in calorimetric measurements.