JOURNAL OF RESEARCH of the Noti onal Bureau of Standards
Volume 82, No. I , July- August 1977
Heat Capacity and Thermodynamic Properties of
Poly(Vinyl Chloride)
Shu-Sing Chang
Institute for Materials Research , National Bureau of Standards, Wash ing ton, D.C. 2 0234
(May 22, 1977)
He at capac ities, C p , of three diffe re nt sampl es of po ly( vin yl c hl o ride) , PVC, have been d ete rmined fro m 6 to
375 K b y adiaba tic ca lorimetry.These three sa mples we re de ri ved from e ith er bul k· or s uspe ns ion- po lyme ri za tion
processes and \...'e re measured e ith e r as rece ived or aft e r pe ll eting unde r p ressure. The heat ca pac iti es of Ih (-'
sampl es are alm ost ide nti ca l if the the rmal a nd pressure hi stori es a re th e sa me. Belo w the g lass tra ns iti on
te mpe ra ture, 1'., of about 355 K, C p of PVC was found to be exce pt iona ll y line a r ove r a wide te mpe ra ture ra nge. C p
of a nnea led PVC Illay be re present ed by (1 0 + 0 .166 1') J K- 1 mol- I to with in I pe rce nt o f the meas ured va lues
fro m 80 to 3 40 K. Appro xim ate ly 200 J mol- I of e ne rgy we re stored in th e sa mpl es b y the pe ll e ting prucesses. The
stored e nergies begin to re lease at about 3 0 to 40 K below the glass tra ns ition te mperature . T. for powde ry or
re laxed sampl es occurs a round 352 to 356 K fo r the sus pe ns ion-polyme ri ze d PV C sa mpl e a nd 348 to 35 1 K for
bulk·polyme rized sa mpl e.
Key wo rds: Ca lorime try; e ntha lpy re laxa ti on; glass tra ns itio n; heat ca pac it y; po lyme r; puly{v in yl c hloride); pressure
effects; the rmodyna mi c propel1i es .
1.
of these two sampl es are essentiall y the sa me if the ir the rma l
a nd mec hanical hi stori es a re s imilar. Both pe lleted samples
s howed phe nomena assoc iated with the re lease of the stored
mec hanical ene rgy when compa red with a third sa mpl e whic h
cons isted of unpell eted bu lk-polymerized PVC powde r.
Introduction
Poly{vinyl c hl orid e), PVC, is one of the most importa nt a nd
widely used polyme rs . Of all the thermoplasti cs, its produ ction is second only to that of polye thylene. Heat capacity
behaviors of PVC have bee n reported in more tha n twenty
pape rs. Most ofthe papers prior to 1968 we re coll ected in a
review article [1] , t othe rs may be found among re fe re nces [222]. Except re fe rences [2-5], all othe rs re ported the ir results
in graphical re presentation onl y. Refe re nce [2] gave a table of
smoothed values a t 25 K int e rval s de rived from meas urements conducted between 60 a nd 300 K. Refe rence [3] used
linear equations for the temperature range between 250 and
330 K. Referen ces [4-5] re ported mean values over wide
temperature regions.
The agreement among diffe rent investi gations is usually
very poor. This poor agreement is often attributed to the
differences in the samples studied rathe r than the differences
in the experimental technique s and practices employed by
different laboratories. It is the purpose of this work to establish more representative thermodynamic prope rties on this
industrially important polymeric material.
Three PVC samples were studied in this work by precis ion
adiabatic calorimetry. These samples include one sampl e
consistin g of suspension polymerized PV C and two sampl es
consisting of bulk-polymerized PVc.
The suspension polymerized PVC sample contained a
sma ll amount of poly{vinyl alcohol) as the suspension age nt,
which was incorporated on the surface of the PVC granules .
A c hemicall y purer PVC sample mad e by bulk-polymerization process was also studied. Both samples were pressed into
pellets before the measurement. The heat capacity behavio rs
I
2.
2.1.
Experime ntal Detai l
Ca lorimetric Techn ique
Heat capac iti es of PVC samples we re mea sured in a
vacuum adiabati c calorime ter desc ribed previously [23]. The
heat capacity of th e suspens ion polymerized sampl e was
measured manually, whe reas those of th e two bulk-pol ymerized samples were measured automatically. A mini computer
was used for both data acqui sition and ex perimental control.
Both temperature and energy were mea sured using a double
s ix-dial pote ntiomete r, whi ch had been converted from manual into programmabl e mode of ope rati on [24]. Alth ough th e
intermittent heating method is used , continu ous c alorim etri c
measurements over a prolonged peri od may be made without
being interrupted thu s avoiding modifi cati on of sample th ermal hi story. The mini compute r was also used to d ecide wh en
a stead y state of tempe rature uniformit y throu ghout the sa mple was reached , and th en to adju st a nd apply a measured
amount of e nergy to the sample as required. Deta iled description of the automated calorime ter measurement syste m
and the measuring procedures is given elsewhe re [25].
2.2.
Materials
Three calorimetric samples of PVC have been studi ed in
this work .
Sample S: - This sample cons isted of pelle ts made by
pressing a commercially available suspe nsion-polymerized
Figures in brackets indica te the literature references at the end of this paper.
9
powder. The powder was supplied to us by Edward A. Collins
of the B. F. Goodrich Chemical Corporation with the designation Geon 103EP. 2 Its number-, viscosity-, weight- and zaverage molecular weights Mn: Mv: Mw: Mz = 64300:
129000: 142000: 254000 were obtained by gel-permeation
chromatography in tetrahydrofuran at 40°C. The intrinsic
viscosity [7]] is 0.91 dL g- l in cyclohexane at 30°C. The
sample has been used for the ASTM D20-70-04 round robin
program. The pellets, about 1.27 cm diam and 1 cm height,
were produced in a pellet-mold in a hydraulic press at room
temperature with a pressure of about 500 MPa. This sample
contained about 0.03 percent of poly(vinyl alcohol).
Sample B: - This sample was obtained by pelleting a powder bulk-polymerized sample. The powder was supplied to us
by Marvin R. Frederick of the B. F. Goodrich Company. It
bears the designation Geon 80X5 and has an intrinsic viscosity of about 0.75 dL g- l and Mw of 80,500 measured by gelpermeation chromatography. A pelleting pressure, just
enough to press the powder into pellet, of about 140 MPa was
used.
Pelleting of powdery samples is often employed in calorimetry to facilitate th e thermal conduction through contacts and
to reduce the adsorption of heat conductin g helium gas at
lower temperatures, th ereby reducing the time required for
the establishmen t of thermal equilibrium.
Sample P: - This is the starting material for sample B, the
original bulk-polymerized sample in the powder form as
received.
Neither bulk- and suspe nsion-polymerized samples contain additives other than that mentioned for sample Sand
have been pumped in vacuum to remove unreacted monomers.
In each of the sa mple loadings, the sample container was
cooled with a water jacket while the top was sealed with
50:50 indium-tin solder, so that the thermal hi story of th e
samples would not be altered before the calorimetri c measurement. Helium gas, at a press ure of about 4 kPa and a
mass equal to that used when th e heat capacity of the empty
sample container was measured, was sealed with th e pelle ted
samples Sand B. Approximately 50 kPa pressure of helium
gas was seal ed with the powder sample P. Summary of th e
calorimetric samples is li sted in table 1.
1_
TABLE
Designa-
Pulymeriza-
tion
tion
1.
Surnrnary of pvc samples
Fonn
Pelleting
pressure, MPa
Sample
mass, g
PHe
at
295 K
kPa
S
Suspension
Pellet
500
66.337
3.8
B
Bulk
Pellet
140
55.659
4.0
P
Bulk
Powder
3.
75.59
49.3
Results and Discussion
Heat capacity data for PVC suspension-polymerized pellet
(S), bulk-polymerized p ellet (B) and bulk-polymerized powder (P) are listed chronologically in table 2 and shown in
figure 1. The second character of the two-character code
2 In order to adequatel y describe materials and experimental procedures, it was occasionall y
necess<u1' to identify commerc ial produl'Is by manufacture r's name and labe l. In no inSlanres does
such identifi('at ion imply e ndorsement by the National Bureau of Standa rds.
denotes the history of the sample, i.e., 1, 2, A, and Q
represent first heating, second heating, annealed and
quenched, respectively. The quenching rate applied was
about 5 K min-I . Annealing was accomplished either by slow
cooling (rate annealing) or by holding the sample at a temperture (soak annealing) about 10 to 20 K below the glass
transition temperature of the quenched sample.
Interruptions in the manual heat capacity measurement
process are noted by the blanks which separate the data for
sample S in table 2. Heat capacities of both bulk-polymerized
samples Band P were measured with an automated measurement system controlled by a minicomputer. Thus continuous
measurements on samples Band P were extended over a wide
temperature range and for a prolonged period.
In determining the temperature increment due to an energy
input, the drift due to non-adiabaticity of the calorimetric
system and that due to the configurational relaxation in the
glass transition region are combined. Therefore, the heat
capacity values listed in table 2 represent that of a glass with
its configurations fixed at mid-point of the energy input
period and do not contain the long term time effects such as
the relaxation process of the glass transition phenomena .
The thermodynamic properties of PVC from 10 to 380 K
are presented in table 3 . These are calculated from the data
of BA as listed in table 2. Throughout this paper, the
superscripts for the symbols denote the stales of the material,
e.g., A and L represent annealed glass and liquid respectively. Specific temperatures are denoted by the subscripts.
Gibbs free energies, G - H o , are not listed in table 3.
Since th e residual entropi es of th ese glassy PVC samples are
not determined, the combination of the enthalpy and entropy
increments from T = 0 K could only yield
However, approximate valu es of G - Ho may be estimated by
ass uming a residual entropy of R In 2 for PVC [26]. At
298.15 K th e term TS o is more than 10 percent of either
quantity, G - Ho or experimentally determinable H - Ho -
T(S - So).
Heat capacity data of all three samples in the glassy state
are compared against the s moothed va:lues of annealed bulkpolymerized pellet, BA, in figure 2. Between 50 and 250 K,
regardless of the differences in th e origins of the samples and
their thermal histories, heat capacity values are within 0.2
percent of each other. Above 250 K the heat capacities of
quenched samples begin to dev iate from that of the annealed
samples , increasing to about l·percent hi gher at 340 K.
However, heat capaciti es of differen t samples with similar
thermal history remain within an envelope of about 0.2
percent.
It has been noted that heat capac ities in th e glassy state are
relatively insensitive to physical, chemical and configurational changes [26]. In general these changes produce less
than 1 percent change in the heat capaci ty. Typical examples
may be found by comparing the results from polystyrene
samples of different tacticities and so urces [27], and cis- l,4polyisoprene samples contai ning varied amounts of antioxidant and stabilizer [28] and from different sources [29).
Greater differences are to be expected from highly plasticized samples and crystallizable samples, especially if th ere
is a large change in th e densities of th e samples, such as
polyethylene [30] and polytrichlorofluoroethylene [30a). Near
Tg , the relative percentage change in the heat capacity is
10
100 ~------------~------------~------------~----------~
I
t
It
9
80
80
60
"7
"'0
-'::.::E
')
...,
Q..
U
40
I
)
o
80
60
60
40
p
o
o
o
o
20
o
o
o
o
o
o
o
o
0
0
0
0
S
..
0
0
0
20
0
c9
0
0
0
0
0
0
0
I
0
0
0
J
200
100
300
400
T. K
(J Fi rst h ca lin ~ (I), 0 Vu~nched (Q), •
FI GU R E ) .
Heat capacities of poly(vinyl chloride} .
Annea led (Al, B Bul k-poly me riwd, I)t! lle led; P Bulk-polymerized . ptJwde r; S sus pens ion-polymerized. pelit· led. In ~t' rwd C p s{'a le applied to P unly.
11
TABLE
2.
Heat capacity data oj poly(vinyl chloride}
(-CH,CHCI-
T,
K
Cp , J K- 1mol - 1
~
TABLE
T,
K
(-C H,CH CI-
Cp ,
J K- 1mol - 1
T,K
50.43
51.60
53. 5 7
55.5 1
57.28
59.l7
309.48
319.38
326.85
33 1. 79
336.83
342.01
347.28
352.46
357.47
362.14
366.63
371.03
60.98
64.78
64.95
66.19
67. 33
68.73
70.97
74.79
84.21
90.05
92.43
94.07
Se ri es II .
<}uench ed (S<})
60.76
67.58
73.29
78.81
84.58
90.76
97.36
18.937
20.461
21.667
22.768
23.921
25.102
26. 3 15
103.49
113.51
123.29
133.17
143.19
J53.39
163.53
171.07
27.408
29.159
30.83
32. 39
34.06
35.67
37.28
38.47
181.24
191.16
201.12
21 1.10
221.01
230.96
241.06
40.05
41.62
43.12
44.69
46.25
47.85
49. 53
24.28
25.59
27.36
29.54
32.09
34.95
38.07
Seri es I.
7.700
8.231
8.926
9.755
10.704
11. 717
12.753
41.53
45.57
49.70
54.02
58.94
13.833
15.020
16. 171
17.304
18.51]
8.00
9.12
10.18
11.24
12.54
13.95
15.44
17.11
19.03
2 1. 28
23.82
26.40
28.88
1.148
1. 500
1.874
2.268
2.778
3 .337
3 .962
4.675
5.501
6 .453
7.528
8.559
9.502
243.02
253.60
263.74
273.94
283.96
293.93
49.85
51.62
53.27
54.99
56.73
58.45
30 1.06
310.74
320.48
330.16
339.69
346.86
351.65
356.3 1
360.87
365.39
369.95
374.74
59.66
61.53
63 .32
65.32
67.60
70.21
74.50
84.26
90.60
92 .21
93 .88
95.76
Seri es " I.
Cp ,
~
J K- 1mol - 1
64.299)
T,
K
Cp , J K- 1mol - 1
B. Bulk-polym e rized , pellel (B)
As Received (S 1)
248.82
256.02
267 .83
278.81
289.22
299.68
Heat capacity data oj poly(vinyl chloride} - Continued
62.499)
A. Su spe nsion polymerized , pellel (S)
Series I.
2.
305.07
309.01
3 13.03
3'17.10
321.19
325.32
329.46
333.56
337.62
341.69
345. 71
349.69
353.59
357.55
361. 58
365.64
369.76
TABLE
60.45
62.30
64.20
66.29
341.49
348.91
352.84
355.50
358. 13
362.08
367.27
372.40
67.23
70.1 2
75.06
86.55
88.75
9 1.04
92.71
94.62
Seri es II.
60.17
61.07
62.28
64.08
64. 15
64.86
65. 77
66.72
68.02
69.3 5
71.18
76.13
84.62
89.45
91.52
93.09
94.56
271.83
275.77
278.89
282.89
286.98
291.04
295.08
299.09
303 .08
307.04
311.08
315.21
319.31
323.39
327.44
331.46
335.47
339.55
343 .71
347.81
351. 76
3 55.52
359.42
363 .58
367.69
3 71. 76
Que nc hed (B<})
54.64
55.29
55.90
56.55
57.30
58.01
58 .69
59.94
60.13
60.83
61.63
62.35
63.30
63.98
64.90
65.70
66.69
67.89
69.68
73.24
81.57
88.82
90.85
92.37
93.79
95.18
2. Heat capacity data oj po/y(vinyl chloride} - Continued
( -CH,CHCI-
T,K
Cp , J K- 1 mol - I
~
62.499)
T, K
Cp , J K- 1 mol - I
Series III. Annea led (BA)
6.21
7.12
8.20
9.37
10.57
11. 76
12.96
14.28
15.75
17.34
19.10
21.06
23.26
25.73
28.53
3 1. 76
35.52
39.92
45.18
51.56
57.24
6 1. 59
65.97
70.38
74.69
78.93
83.28
87.60
91.90
96.35
100.81
105.27
109.76
132.38
136.87
141.43
146.05
150.74
Ann ea led (SA)
306.46
316.84
327.04
337.05
Firsl Heating (BI)
12
0.643
0.896
1.203
1.597
2.013
2.466
2.946
3.471
4.100
4.778
5.549
6.368
7.295
8.298
9.388
10.596
11.920
13.355
14.916
16.690
18.117
19.145
20.121
21.082
21.972
22.857
23.679
24.541
25.349
26.1 76
27.003
27.788
28.596
32.33
33.03
33.82
34.51
35.26
155.51
162.51
169.42
176.17
185.29
194.48
201.40
206.39
2J 1..59
216.74
221.84
226.89
232.03
237.25
247.52
252.73
258.03
263.27
268.47
273.63
278.88
284.21
289.50
294.74
299.94
305.23
310.60
3 15.91
321.19
326.43
331. 74
337.13
342.47
347.73
352.63
357.75
363.09
368.48
36.02
37.12
38.21
39.29
40.71
42.19
43.21
44.04
44.86
45.70
46.54
47.28
48.14
48.85
50.56
51.42
52.33
53 .19
54.08
54.83
55.77
56.65
57.57
58.42
59.33
60.28
61.2 1
62.25
63.17
64.14
65.18
66.36
67.50
69.34
85.65
90.31
92.45
94.08
TABLE
2.
Hea.t capacity data of poly(vinyl chloride) - Continu ed
TABLE
3.
T!wrrnodynarnic fiUld.ion, of poly(vinyl ch.loride)
(-CH,CftCt- = 62.499)
T, K
Cp , J K
I
moll
(- CH,C II Ct-
Cp , J K- 1 11101 - 1
1', K
Seri es IV. Quench ed (BV)
6.89
7.93
9.08
10.25
11.43
12.64
13.97
15.41
16.99
18.69
20.58
22.72
25.12
27.86
31.00
34.65
38.92
43.36
47.34
50.88
54. 11
57.10
61.77
67. 17
71.52
75.94
80 .28
84.58
89.00
93.4 1
97.82
102.23
106.66
111.12
115.62
120.15
124.73
129.36
133.91
143.04-
0.826
1. ]]8
1.491
1.908
2.336
2.811
3.335
3.9:;4
4.622
5.353
6.164
7.064
8.060
9.127
10.311
11.619
13.048
14.401
15.543
16.507
17.337
18.095
19. ]83
20.370
21.318
22.216
23. III
23.946
24.792
25.624
26.416
27.219
28.020
28. 789
29.547
30.3 ]
31.13
31.81
32.61
34.03
Series L
First Heating (PI)
10
15
20
25
30
35
40
45
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
250
260
270
273.15
280
290
298. 15
300
3 10
320
330
340
350
360
370
380
34.78
35.60
36.26
37. 06
37.72
38.50
39.25
40.03
40.71
4 1. 48
42.21
43.00
43.70
44.43
45.22
46.0446.79
47.54
48.32
49.06
53. 13
54.03
54.85
56.50
57.43
58.26
59. 18
60.02
60.94
61.91
62.88
63.83
64.89
65.97
67.34
69.2 1
73 .26
8.3.22
89.70
91. 92
93.60
147.50
152.16
156.90
161.58
166.19
170.88
175.65
180.37
185.03
]89.78
194.61
199.40
204.14
208.96
213.87
218.73
223.54
228.32
233.17
238. 12
262.62
267.68
272.57
282.38
287.42
292.43
297.39
302.32
307.33
312.43
3 17.49
322.50
327.48
332.53
337.67
342.76
347.77
352.68
357.54
362.46
367.53
C. Bulk -polymerized , powder (P)
304.29
299.73
304.68
309.75
3 14.62
3 19.42
324.34
329.34
334.47
339.57
344. 64
60.06
59.18
60.10
60.97
62.01
63.07
64.24
65.59
66.68
67.92
69.71
Seri es II.
300.52
305.75
310.94
316.08
321.17
326.22
331.22
336.18
341.08
Cp
T
K
Seco nd Heating (P2)
59.50
60. 52
61.46
62.49
63.39
64.35
65.38
66.52
68.27
= 62.499)
N-NoA
S-SOA
J K- 1 mol- I
J 11101 - 1
J K- 1 lllol - 1
1.81
3 .78
5 .92
8.01
9.95
I J. 73
13. 38
14.89
16.28
18.77
20.99
23.05
24.98
26.82
28 . .59
30.29
3 1. 94
.33 . .56
35.1 .5
36.73
38.3 1
39.88
4 1.45
43.03
44.6 1
46.2 1
47.79
49.37
50.99
52 .63
54.29
54.8 1
55.95
57.64
59.03
59.35
61. ]]
62.94
64.88
66.96
87.59
91.08
94.56
98.05
5.54
19,4
43.6
78.5
123.4
177.7
240.5
:3 11. 3
389.2
564.7
763.7
984.0
1224.
1483.
176 1.
2055.
2366.
2694.
3037.
3397.
3772.
4163.
4.569.
4992 .
5430.
5884.
6354.
6840.
7342.
7860.
8394.
8566.
8945.
9513.
9989.
10098.
10700.
11 32 1.
11960.
126 19.
13317.
14 110.
15 138.
1610 I.
0.75
1.84
3.22
4.79
6.40
8.07
9.75
I 1. 41
13.05
16.25
19.31
22.25
25.08
27.8 1
30.4.5
33.0 1
35.50
37.92
40.29
42.6 1
44.88
47.12
49.32
5 1. 48
53.62
55.73
57.82
59.89
6 1.94
63.97
65.99
66.62
67.99
69.98
7 1.60
71.97
73.94
75.9 1
77.88
79.84
81.86
84.38
86.93
89.49
The observations on PVC indi cate that the max tma would
occur at temperatures below 6 K.
Heat capacity measureme nts on PV C have not been extend.ed to temperatures below 6 K. Excessively long equilibratIOn times were ex pe ri e nced below 6 K, due probably to
the adsorption of th e helium heat exchange gas by the powdery sampl es.
approximately that of the de nsity changes. At liquid helium
te mperatures, the heat capacity difference between glass and
crystal could be relatively very large. The heat capac ity of the
glass may be 2 to 4 times higher than that of its crystalline
counterpart [26].
The rapid ri se in Cp l T3 as temperature is lowered in the
liquid helium t.emperature reg.ion , figure 3, has commonly
been ~bserved In all amorphou s materials studied [26]. The
behaVIOr of all three PVC samples is quite similar in this
aspect. A maximum in Cp lT3 generally occurs in the temperature range between 2 and 7 K for amorphous materials [26].
13
3. 1.
Linea r C1' in the Glassy State
Below the glass transition temperature the C of PVC is
h.ighly linear o~er a wide temperature range. The followin g
SImple expreSSIOns, eqs (1) to (3), may be used 10 describe
the observed C1' behavior of annealed PVC, depending upon
the pre?i~ion and the te mperature range as stated below for
the IndIVIdual expressions:
C1' = 11.45
+ 0.158
T
0.1%,140 - 240 K
(1 )
Cp = 10.95
Cp = 10
+ 0.162 T
+ 0.166
T
0.5%, 100 - 320 K
(2)
1%,80 - 340 K
(3)
3.2.
Measurements by Lebedev et a1. [2] cover the temperature
range from 60 to 300 K. From 75 to 150 K, their reported
values agree well within 1 percent of ours . From 175 to 275 K
their values are about 2 percent lower and at 300 K about 1
percent higher than ours. Their extrapolated values for 25
and 50 K deviate significantly from our measured values.
Alford and Dole's measurement [3] covered the tempe rature range from 250 to 390 K. Between 300 and 390 K their
data were shown graphically. Two linear equations were
given to represent the Cp behavior of the ir original polymer
and the ir annealed sample. The tempe rature ranges of applicability for th e two expressions were not given explicitly.
Both the ir dCp /dT values of 0 .191 J K- 2 mol- I for the original
PVC sample and 0.221 J K - 2 mol- 1 for the annealed sample
are higher than our mean dCp/dT value of 0. 176 J K - 2 mol- 1
for the te mperature range from 2 50 to 340 K. Their Cp values
intersect ours at around 280 to 3 00 K. At e ither end of their
te mperature range of measu rement in the glassy state, their
values deviate from ours about 1 to 5 percent.
where C p is in J K - l mol- I and Tin K. The simple expression
of eq (3) represents the C p of annealed, bulk-polyme rized
PVC to about 1 perce nt of the obse rved values over the entire
temperature range from liquid nitrogen temperatures to just
below the glass transition region. Equation (1) gives C p
values within 0.1 percent, about the experimental precision,
of the observed C p values for annealed PVC for a narrower
temperature range.
I
I
I
0.6 t--
o
Comparison With Literature Values Below Tg
-
o
3.3.
o x _
0.4 t--
The e nthalpy changes in the glass transition region for
sample S are shown in figure 4. In order to obtain the
precision required, th ese changes are not computed as the
integral of the heat capaci ty values reported in table 2 , which
do not co ntain the configurational contribution. Instead the
e nthalpy increments in table 4 are calculated as the sum of
the e nergy inputs, !Q, app lied to the calorime ter between a
te mperature , T, and the final tempe rature, j, in the supercooled liquid state , L, minus the enthalpy increment of the
c;
E
~ 0.2 t..:
u
<J
Glass Transformation
<>
o
o
<>
.. ..
-0
o
0
•
o •
•
----
1000 r-----~,-----,-----,-----".__,
-0.2 ~_...L..-_-:c~l_-L-_~I:--_L----:~I=-----.J
o
100
200
300
o
T. K
Comparison of heat capacities of poly(vinyl chloride}.
FIG URE 2.
Base line (.6.C p = O) -Smoothed C p of BA. X SQ, ® SA, 0 and <D RQ, • BA, <> P2.
"0
·1000
E
...,
0.003
---,
~
~
'7
·2000
I
:.::
-=E
0.002
·3000
~
~
u
0.001
OL-____
~
______
~
_______ L_ _ _ _
o
~
300
400
T'
3.
340
360
380
T. K
K'
FI GU R E 4.
FIGURE
o SQ. 0
320
Low temperature heat capacities of poly(vinyl chloride) .
Enthalpy changes of poly(vinyl chloride} in the g lass transition
region.
() 5 1, 0 SQ • • SA.
a nd CD BQ • • BA.
14
TABLE
4.
300
Enthalpy ami eruropy cha nges of poly(vinyl chloride} in the
g lass transition region
(-CH,CHCI- = 62.499)
Samp le
Designati on
No. of
T,
T,
200
H~65 - H300
Sf6.
5 300
runs
K
K
J mol - J
J K- J mol - J
SI
SA
SQ
11
11
11
304.5
301.2
296.2
364.4
364.7
367.7
4261.
4495.
4456.
12.80
13.47
13.36
Bl
BQ
BA
BQ
16
16
13
13
303.1
301.1
302.5
299.9
363.6
365.6
365.8
360.0
4371.
4522.
4571.
4525.
13.11
13.55
13.69
13.56
PI
P2
14
14
302.1
297.9
361.5
358.3
4543.
4546.
13.61
13.61
100
-
]
"
·100
200
300
300
e mpty sample container , MT, for th e same temperature interval:
F IGU R E
H'1 - H,!,
= [lQ -
(H¥'!' - HM'!')] / n
(4)
5.
Enthalpy changes ofpoly(vinyl chloride} in the glass transition
II - HB = H () 5 1.0 SQ . • SA.
whe re n is th e number of moles of th e sample . This e nthalpy
increme nt between temperatures T and! is the n adjusted to a
common reference temperature, r, in th e supercooled liquid ,
L , thus:
(5)
H~
H - HE = H -
H~65
JO(T - 365)
_r~~~7rl
-
:3652 ) J rnol- 1
Glass Transition Temperature
The glass transition tempe rature, Tg , is defin ed he re as th e
temperature of intersection of th e entha lp y curv es (so metimes
extrapolated) for th e glass and for th e supercool ed liquid. The
Tg so defin ed is determin ed by th e hi s tory by whi ch th e glass
was formed on cooling but it is inde pe ndent of th e rate of
obseJ-vation during th e heatin g in th e glass transition region.
Thi s d efiniti on is a na logo us to th at used in di latometri c
s tudi es wh ere th e break in th e volume versus te mperature
CUI-ve is defin ed as th e glass transition tempe rature. For th e
suspension pol ymerized sampl e, Tg for SQ occurred at 355 K
and for SA at 352 K. For th e bulk-polyme rized sampl es, Tg
for BQ, BA, PI and P2 a re 351, 348, 351 , and 350 K,
res pectively . Although th e heat capac iti es of th e bulk- and
suspension-polymerized samples are almost identi cal below
th e glass transition region, th e glass tra nsiti on te mperat ure of
th e bulk-polymeri zed sample is a bout 4 K lower than th at of
the sus pension-pol ymerized sa mple with s imilar hi story. The
lowering of Tg pe rhaps refl ec ts th e fact th at th e bulk-polymerized sample has a lowe r mol ecu la r weight. Diffe re nces in th e
molec ular we ight di stribution a nd/o r th e residual monom er
content may also playa part.
- lO( l' - 365)
- 0.083('['2 - 365 2 ) J mol- I
-
The exchan ge of enthalp y between th e sampl e a nd its containe r causes th e effec t of spontaneous e nthalp y relaxation of
th e sampl e to appear more isoth ermal th a n adiabati c.
3.4.
where C~ is th e heat capac it y of th e sample in th e supercooled liquid state . A reference tempe rature of 365 K was
c hosen. The energy applie d to th e calorimeter should be
corrected for any non-ad iabatic condition that existed between th e sample containe r a nd its surroundings . In th e
present instrument , th e typ ical temperature drift du e to nonadiabaticity is less than ± 0 .1 mK min - I. Th erefore for an
experiment lastin g about 10 hours and coverin g some 60 K
ra nge, th e non-adi abati c contribution is less than 0.1 pe rcent
of the total energy appli ed and may be neglec ted.
In order to show th e enthalpy chan ges in th e glass transition region in greater detail than that in fi gure 4, a base
e nthalpy increment between the reference temperature 365 K
and th e temperature of observation as calculated from eq (3)
for annealed PVC was subtracted from th e observed enthalpy
increment. The results are shown in figure 5, where
380
320
(6)
3.5.
The heavy solid lines in figure 5 connect th e initi al te mp eratures after each electrical energy input to th e calorimeter. Th e doubl e points or the thin lines linking each of the
doubl e points denote th e extent of e nthalpy relaxation which
OCCUlTed during th e temperature drift obsel-vati on pe ri od of
about one half to one hour. The sampl e container and, its
contents were in adiabati c condition with th e surroundings .
When th e sampl e tempe rature was raised or lowered spontaneo usly du e to the enthalpy relaxation , th e sample e nthalpy
was decreased or increased du e to th e corresponding temperature chan ges of th e sa mple containe r with that of th e sample .
Enthalpy Relaxation and Adiabatic
Temperature Drift
Large spontaneous tempe rature drifts have bee n obser-ved
in th e glass transition region for e ith er qu e nched or ann eal ed
glasses, due to the configurations relaxing toward th e supercooled liquid under adiabatic conditi ons [32). An exothermic peak is generally obsel-ved to occu r just below th e
glass transition temperature for th e qu enched glass, whil e an
e ndothermic peak may be seen above the Tg for th e annealed
glass.
15
Figure 6 shows th e spontaneous drifts observed for sample
S at different times after the termination of electrical energy
input to the c alorimeter heater. Due to the axial configuration
of the heater and the thermometer in th e calorimeter, the
temperature of the central heater-thermometer assembly is a
few kelvins higher than that of th e sample at the end of th e
heating period. When th e electrical energy to th e heater is
turned off, a near exponential decay of the temperature of the
thermometer is observed for the establishment of a uniform
temperature throughout the sample container. The time constant is in the order of 50-100 s at temperatures around 300
K. In figure 6A, the influ ence of the above mentioned effect
may still be observed, at 10 min after th e termination of the
electrical energy input to th e calorimeter.
The usual exothermic and endothe rmi c behaviors for
quenched and annealed glasses are observed for SQ and SA .
For Sl, an additional amount of positive drift due to the
release of stored pelleting energy appears in the tempe rature
range 320 to 350 K. Similar behavior has also been observed
for B1. For PI, the initial drift is sligh tly negative and
reaches about -0.3 mK min- 1 at about 330 K where exoth ermi c behav ior begi ns to be noticed. Therefore the initial
slight negative drift near room temperature for both Sl and
Bl may also be associated with th e original samples, although s uch a behavior is soon maske d by the release of
pelleting energy. Above 360 K, all observed drifts remain
slightly positive. Thi s is probably associated with the onset of
th ermal degradation of th e material [33].
In general the relaxation ti me constant becomes shorter as
th e temperature is increased. This behavior is readily observable by examining the changes in the drift with time for SA at
two temperatures, 356.9 and 354.2 K (fig. 6), before and
after the heat capacity measurement at a mean temperature of
355.55 K (table 1). The drift at 356.9 K is much greater than
that at 354.2 K at 10 min after the energy input. At 20 min
after the e ne rgy input , the drift at 356.9 K is already less
than that at 354.2 K.
Longer observations have been made on PI. The relaxation
time constants 'T were estimated to be 42.5 min at 350.5 K
and 126.5 min at 335 K. It is possible to construct an
Arrhenius-type plot, which yields an apparent activation
energy for the relaxation process in the glass transition region
as 69 kJ mol - I. However, as the glass transition process
cannot b e described by a single relaxation time and th e
relaxation times are estimated for a glass with progressively
changing configurations , and with the further complication
that the observations on PVC may include the effe cts of
th ermal degradation in and above th e glass transition region ,
such an estimation of activation energy is at best a rough
approximation.
Straff and Uhlmann [34] reported recently that the enthalpy of PVC decreases by 150 to 170 J mol- I after annealing for 50 to 250 hours near Tg • In our experiment the
decrease in the enthalpy is about 20 to 40 J mol - I. They also
reported an activation energy spectrum with a peak around 78
kJ mol-I for the enthalpy relaxation process of PVc.
3.6.
Stored Energy from Pelleting Process
This subject has been dealt with in detail elsewhere [31].
When the pelleted, suspension-polymerized PVC, Sl, was
being heated from room temperature for the first time, spontaneous energy release began to be observable at 'tbout 320 K
and lasted through the entire glass transition region (fig. 6).
16
8
4
O~------~---------------'~---1
-4
t:.
-8
Ie:
E
<>
-12
'"E
~
"II:
0
(B) 20 min
4
-4
-4~~~____~~__~~____~~____~
380
T,K
FIGURE
6.
Spontaneous temperature drifts of poly(vinyl chloride) in the
glass tramition region.
Il. S I. 0 S<), 0 SA.
Similar findings were observed for Bl [3 1]. The onset of the
release of the stored energy from th e pelleting process also
caused small anomalies in th e heat capacity behavior for Sl
and Bl (fig. 1).
Around room temperature, the enthalpies of Sl and Bl are
substantially higher than that of the corresponding samples
which received subsequent quenching and annealing treatments, figures 4 and 5. By combining the results of th e two
bulk-polymerized samples, Band P , it was found that the
enthalpy increments for PI and P2 lie between those of BQ
and BA [31]. Hence the enthalpy of the original material is
approximately midway between that of the material after
quenching and annealing. The difference in enthalpy be-
tween Bl and PI is about 175 J mol- I. The diffe rences in
e nthalp y be tween SI a nd the mean value of SQ and SA is
about 215 J mol - I.
3.7.
Effect of Pressure Densification on Ty
The glass transi tion phe nomenon may occur over a wide
te mpe rature region depending upon differe nt ph ys ical prope rti es be ing observ ed a nd the criteri a chosen as the indicati on for the tra ns itio n. It is ge nerally beli eved that the breaks
in the extensive the rmodynami c properti es, such as the volume, enthalpy and entropy , would occur at the same te mpera ture for a configurationally fix ed glass form ed by a particular
hi story. This view is supported by most dila tometri c and
calorimetric observation s on glasses normall y formed under
a tmosphere pressu re, and by the observ ati ons that, upon
a nnealing, all the exte nsive th ermodynami c properti es relax
to lowe r valu es, i. e. , towa rd th at of a n equilibrium s upercooled liquid state.
It is also poss i ble to redu ce the volume of a glass by
permane nt pressure-de nsifi cation [35, 36], eith er by compress ing in th e glassy state or by vitrifi cati on from th e liqui d
sta te under preSSUTe and followed by the release of pressure
around room tempe rature. The pressure-dens ifi ed glasses,
though having lower volumes, may contain hi gher enthalpi es
a nd entropi es than that of a normall y form ed glass , and a re
actually furth er re moved from the equilibrium supercooled
liquid state at normal pressure, as observ ed in the present
and other works [37, 38]. In contras t to most observa ti ons
[3 7-41] that the enthalpi es of pressure-d ensifi ed glasses are
either essentiall y un cha nged if formed at low press ures (l ess
than 100 MPa) or hi ghe r tha n that of normal glasses if formed
at hi gh pressures, refere nces [42 , 43] reported lowering of
enthalpy for pressu re-d e nsifi ed glasses. H e nce for most syste ms obse rved , the dilatometry may s how a depression of th e
glass trans iti on temperature, wh ereas the calorimetri c observa tion may show a n elevation of the glass transition tempe rature, as compared with tha t of a normall y formed glass.
Observation s on the same pressure-densified glasses indi cate
that the volumetri c and the enthalpi c relaxations do not
necessarily occur simultaneously or relax with the same time
constant [38].
Hence, the dilatome tric and calorimetric glass transition ,
defined as a break in the volumetric or enthalpi c curves, may
not necessarily oc cur at the same time nor at the same
temperature for pressure-densifi ed glasses. Thermod yna mic
studi es lead to a conclusion that more than one ordering
parameter is required to describe the glassy state [44, 45]. It
is possible that some of these parameters appear to be more
ac tive dilatometrically and others to be more active calorimetrically. Further the rmodynamic studies are underway to
clarify this apparent distinction [46] .
3.S.
c;,
uncertainti es. After th e calorim etric measure me nt , th ese
samples appear slightly tinted with une ven purp lish color.
The intensity and th e spread of di scolora ti on in th e sa mpl e
seem to inc rea se as the durati on a nd the hi ghest tempe rature
at whi ch the sampl e have been subjec ted to are increased.
Apparently the s li ght di scoloration or decompositi on does not
affect the Cp s ignifi cantl y. Di scoloration of PV C has a lso
been observed in oth er calorim etri c measurements [3]. More
detail about th e earl y stage of therma l decompos ition of PV C
has bee n reported elsewh ere [33]. Th e heat capac ity valu es of
different liquid PVC sampl es differ s lightl y, but the slopes
dCp /dT are almost identi cal. Cp of th e liquid of sample B is
0.9 percent hi gher than t hat of sample S, and that of sampl e P
is 0.4 percent highe r than that of sampl e B. Since th e onl y
differe nce between samples P a nd B is the mechan ical treatment of the pelletin g process, the lowe rin g of Cp of sa mpl e B
ma y be att ri bu ted to some residu al effec t from the peJletin g
process or stress-induced c rysta lliza ti on. However, th e low
c rystall inity of PVC [47] indi cates tha t indu ced c rystall izati on is not likely to cause th ese effects. Simil ar argument may
be used for sample S because of it s hi gher pe ll eting pressure.
However, the diffe rence in Cp between Sa nd B or P may be
du e to the differences in th e samp le conditi on a nd pre pa rati on.
The rate of change of Cp with respect to te mpe rature for
liquid PVC, dCp/dT, is about 0.355 J K- 2 mol- I for a ll three
samples . Thi s value is hi ghe r tha n th at of th e glasses at about
0.16 J K- 2 mol - I just below Ty. Thi s phe nomenon of hi gher
dCp/dT for liquid tha n that of glass just across th e Ty is not
very common. Normall y dCp/dT of liquid is obse rv ed to be
less tha n that of th e glass. Often dCp/dT of liq uid is negati ve
just above the Ty .
Hi gher dCp /dT valu es a bove Ty have also bee n observ ed in
hi ghl y c rystalline polymers, such as polye th ylene [30] a nd
polytrifluoroeth ylene [30a]. Thi s behav ior is probably du e to
the wide di stributi on of c rysta ll ite sizes a nd c rystalline pe rfecti on, and he nce th e ex tension of preme lting ph e nome na to
temperature near Tg . It is less plau s ibl e to assume th at th e
high dCp/dT of supe rcooled PVC liquid , containing about 1
pe rcent of crys tal, may be caused by s imilar mecha nis ms as
that observed in highly c rystalline polymers. The dehydrochlorination reaction at te mperatures above Tg may cause a
hi gher dCp/dT than usual.
Alford and Dole's work [3] on PVC liquid cove rs a s imilar
te mpe rature ran ge of 360 to 390 K. Judging from their graph ,
their C p values at 360 and 380 K diffe r abo ut + 0. 6 pe rcent
and -0.6 percent from our values at corres ponding temperatures. Although the ir dC vldT value for liquid PVC of 0.29 J
K- 2 mol - I is less than ours, it is ne verthe less also highe r than
thei r dC p/dT values for the glasses at 0 .19 to 0.22 J K- 2
mol- I.
4.
Summary
of Liquid
Around 360 K and above, all the PVC samples studied
reach eq uilibrium supercooled liquid state within the normal
drift observation period of one-half to one hour. Although the
temperatu re range of measurement for the s.upercooled liquid
region is relati vely narrow , the heat capacity values of PVC
liquid are reprod uc ible to wi thin 0.1 percent for each of the
three samples .
For each Sample. no progressive cha nge in the heat capacity wi th time may be detected beyond the experimental
Heat ca paciti es of poly(vinyl c hl oride) sampl es, fro m two
diffe re nt polymeriza tion proG.,esses (suspens ion a nd bulk) and
in either the original powder form or after pe lle ting were
determined from 6 to 375 K by prec is ion adiabatic calorimetry. Below 250 K C p of PVC is essentially independe nt of
sample origi n and previous histories. Between 250 K and T g,
C p is a fun ction of thermal and pressure histories rather than
a fun c tion of sample origin. C p of glassy PVC is linear with
temperature for an exceptionally wide te mperature ra nge.
From liquid nitrogen temperatures to T g, Cp of annealed PVC
17
I
I
may be represented by (10 + 0.166 T) J K- 1 mol- I to 1
percent of the observed values.
The pelleting process contributes 175 to 215 J mol - I of
energy to be stored in the sample. The stored energies begin
to release at about 30 to 40 K below T g. Thus in calorimetric
work, especially in e nthalpic measurements in reaction calorimetry , where samples are customarily pelle ted care must be
taken to eliminate the stored energies due to pelleting in
order to achieve highest reliability.
T g of PVC is however more sensitive than C p to both the
histories and sample origin. The highe r enthalpy of the
pelleted sample raises the T g as observed calorimetrically.
However, the pressure- pelleting process may cause a permanent pressure-densification of the sample. A sample sotreated would show a lowering of T g dilatometrically. Therefore the glass transition as observed calorimetrically is not
necessarily identical to that observed dilatometrically, eve n
though it is observed on the same material.
S.
References
[11 Wunderlich, B., Adv. Polymer Sci. 7, 260 (1970).
[21 Lebedev, B. V., Rabinovich , I. B., and Budarina, V. A., Vysokomol.
Soedin A9, 488 (1967).
[3] Alford, S., and Dol e, M., J. Am. Chem. Soc. 77,4774 (1955).
[41 Badoehe, M., and Li, S.-H., Compl. Rend. Aead . Sci. (Paris) 231, 50
(1950).
[5] Badoche, M. , and Li, S.-H., Bull. Soc. Chim. France 1951, 546.
[6] Dunlap, L. H. , J. Polymer Sci. A-2, 4, 673 (1966).
[7] Gast, T., KUllststoffe 43, 15 (1953).
[8] Grewer, T., and Wilski, H., Kol!. Z. u. Z. Polym. 226,46 (1968).
[9] Hager, N. E., Jr. , Rev. Sci. Instr. 35, 618 (1964) and 43, 1116
(1972).
[10] Hellwege , K. H., Knappe , W., and Semjonow, V., Z. angew. Phys.
11, 22 (1959).
[11] Hellwege, K. H., Knappe, W., and Wetzel , W., Koli. Z. u. Z. Polym.
180, 126 (1962).
[12) Holzmuller, W., Tautz, H., and Seifert, K., Plaste u. Kautschuk 9,
264 (1962).
[13] Juijin, J. A., Gisolf, J. H., and de jong, W. A., Koli. Z. u. Z. Polym.
235, 75 (1969).
[14] Martin, H., and Muller, F. H., Koll. Z. u. Z. Polym. 191,1 (1963).
[15] Martynenko, L. Ya., Rabinovich, I. B., Ovchinnikov, Yu. V. , and
Maslova, V. A., Vysokomol. Soedin. A12, 841 (1970).
[16] Mishchenko, M. I., Samoilov, A. V., and Buchatskii, V. A., Plasticheskie Massy 1966, No.3, 59.
[17] Steere, R. c., J. Appl. Polym. Sci. 10, 1673 (1966).
[18] Tautz, H. , Gluck, M., Hartmann , G., and Leuteritz, R., Plaste u.
Kautschuk 11,648 (1963) and 11,657 (1964).
[19] Vieweg, R. , and Gottwald, F., Kunststoffe 30,138 (1940).
[20] Wilski, H., Kunststoff Rundscha u 13, 1 (1966).
[21] Wilski, H. , Kol!. Z. u. Z. Polym. 210, 37 (1966).
[22] Wilski, H., Koll . Z. u. Z. Polym. 238,426 (1970).
[23] Sterrett, K. F., Blackburn , D. H., Bestul, A. B., Chang S. S., and
Hormann, J. A., J. Res. Nat. Bur. Stand. (U.S.), 69C (Eng. and
lnstr.), No.1, 19-26 (Jan-Mar. 1965).
[24) Chang, S. S. , J. Res. Nat. Bur. Stand. (U.S.), 80A (phys. and Chern.),
No.4, 669-675 (Nov. Dec. 1976).
[25] Chang, S. S. , Proc. 7th Sym. Thermophys. Prop. (1977) .
[26] Chang, S. S. , ACS Org. Coatings and Plastics Preprint 35 No.2, 365
(1975).
[27] Chang, S. S., and Bestul, A. B. , J. Polym. Sci . A-2, 6,849 (1968).
[28] Chang, S. S., and Bestul, A. B., J. Res. Nat. Bur. Stand. (U .S.), 75A
(phys. and Chern.), No.2, 1 13-120 (Mar.-Apr. 1971).
[29] Bekkedahl, N. , and Matheson, H., J. Res. Nat . Bur. Stand. (U.S.) ,
15, 50.3-515 (Nov. 1935) RP844.
[30] Chang, S. S. , J. Res . Nat. Bur. Stand. (U.S.) , 78A (phys. and Chem.),
No.3, 387-400 (May-June (974) .
[30a] Chang, S. S., to be publish ed.
131] Chang, S. S., J. Chern. Thermodynamics 9, 189 (1977).
[32] Chang, S. S., J. Polym. Sci. C, 43 (1973) .
[33] Abbas, K. B., and Sorvik, E. M., J. App!. Polym. Sci. 17,3567 , 3577
(1973) .
[34] Straff, R. , and Uhlmann, D. R. , 1. Polym. Sci. , Polym. Phys. Ed . 14,
1087 (1976).
[35] Kimmel, R. M. , and Uhlmann , D. R., J. Appl. Phys. 41, 2917
(1970).
[36] McKinney , J. E. , arid Goldstein , M. , J. Res. Nat. Bur. Stand. (U.S.) ,
78A (Phys. and Chern.) , No.3. 331-355 (May-June 1974).
[37] Price, C., Polymer 16, 585 (1975).
[38] Weitz , A., and Wunderlich , B. , J. Polym. Sci., Polym. Phys. Ed. 12,
2473 (1974).
[39] Allen , G., Ayerst, R. c. , Cleveland, J. R., Gee, G., and Price, c. , J.
Polym. Sci. C 23, 127 (1968).
[40] Ichihara, S., Komatsu, A., Tsujita , Y. , Nose , T., and Hata, T. , Polym.
J. 2, 530 (1971).
[41] lchihara, S., Komatsu, A., and Hata, T., Polymerl. 2,644 (1971).
[42] Ichihara, S., Komatsu, A., and Hata, T. , Polymer J. 2,650 (1971).
[43] Kimmel, R. M. , and Uhlmann, D. R., J. Appl. Phys. 41,4917
(1971).
[44] Davies, R. 0. , and Jones, G. 0., Adv. Phys. 2,370 (1953).
[45] Staverman, A. J., Rheo!. Acta 5, 283 (1966) .
[46] Chang, S. S., McKinney, J. E., and Guttman, C. M., to be published.
[47] Straff, R. S., and Uhlmann, D. R., J. Polym. Sci., Polym Phys. Ed.
14, 353 (1976).
18
J