I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
September 1949
2067
ACKNOWLEDGMENT
TABLE VI. PHENOLIC SUBSTANCES FROM HYDROGENATION OF LIGNIN
Substance Phenols Phenol o-Cresol w-Cresol Guaiacol 2 4-Dimethylphenol 2~Methoxy-4-methylphenol 2-Methoxy-4-ethylphenol 2-Methoxy-4-propyl phenol Higher-boiling phenol8 a n d methoxyphenols
Hydrogenation, % At 325O C. A t 400' C. 1.2 0.2 3.2 4.8 4.9 3.3 12.0 12.0
1010 6.0 9.0
-
-
2.8 1.4 2.2 2.3 5.3
5.1 1.8 3.0
40
81.6
Catechols Catechol 4-Methylcatechol 4-Ethylcatecbol 4-Propylcatechol Higher-boiling products
-14.0
4.0 2.7
The authors wish to express their appreciation to the U. S. Bureau of Mines Pittsburgh, for the use of hydrogenation equipment, t o H. H. &orch, L. L. Hirst, and co-workers for carr ing out the hydrogenation, and to A. Eisner for a preliminary anaf sis of the products. They also wish t o thank John Traquair, $he Mead Corporation, for supplying the lignin for this work.
5.5
15.0 9.1
28.6
84.9
2.5
4.1 18.6
The nonphenolic oxygen compounds, consisting principally of ketones, appear t o be accounted for as products that resulted from the pyrolysis of 2-methoxy-4propylphenol derivatives. The hydrocarbons represent the products of greater degradation. It is possible that a knowledge of the hydrocarbons produced at the lower temperature will be of value in the determination of structure but not to the extent that the phenolic compounds are. The high-boiling heavy tar obtained in this work is similar in amount and in properties to that obtained in the hydrogenation of other lignin preparations. Further hydrogenolysis at long periods of time converts this product to cyclohexane and cyclohexanol derivatives, which are in keeping with the proposed building unit.
LITERATURE CITED
(1) Adkins, H., "Reactions of Hydrogen," p. 12,Madison, University of Wisconain Press, 1937. (2) Adkins, H., Frank, R. L., and Bloom, E. S., J. Am. Chem. SOC., 63,549-55 (1941). (3) Corson, B.B., and Ipatieff, V. N., J . Phye. Chem., 45,431(1941). (4) Coulthard, C. E.,Marshall, J., and Pyman, F. L., J . Chem. Soc., 34,280(1930). (5) Fenske, M. R., Tongberg, C. O., and Quiggle, D., IND.ENG. CHEM., 26,1169 (1934). (6) Freudenberg, K., and Adam, K., Ber., 74B, 387-98 (1941). (7) Harris, E. E., D'Ianni, J., and Adkins, H., J . Am. Chem. SOC., 60,1467(1938). (8) Harris, E. E., Saeman, J. F., and Sherrard, E. C., IND. ENG. CHEM.,32,440(1940). (9) Ipatieff, V. N., Corson, B. B., and Kurbatov, J. D., J . Phgs. Chem., 44,670(1940). (10) Kester, E. B., IND. ENO. CHEM.,24, 1121 (1932); 25, 1148 (1933). (11) Lipscomb, W. N.,and Baker, R. H., J . Am. Chem. SOC.,64, 179 (1942). (12) Osokin, A. S., J . Gen. Chem. (U.S.S.R.), 8, 583-7 (1938). (13)Ibid., 9,1315-25(1939). ENG.CHEM., 32, 1399 (1940). (14) Plunguian, Mark, IND. (15) Saeman, J. F., and Harris, E. E., J . Am. Chem. SOC.,68,2507 (1946). (16) Storoh H. H., Hirst, L. L., Fisher, C. H., and Sprunk, G. C., U.8. Bur. Mines, Tech. Papar 622 (1941). RECEIVWD September 10, 1948. Presented before the Divieion of Cellulose Chemistry at the 114th Meeting of the AMERICAN CHEMICAL SOCIETY, Portland, Ore. A portion of t h e work on t h e isolation and fractionation of t h e oxygen-containing aompounds in t h e neutral oil fraction was conducted by C. B. Bergstrom, a n d the report is presented a8 fulfilling a part of the requirements for a master's thesis a t the University of Wisconsin.
Viscosity of Normal Paraffins near -the Freezing Point J
E. B. GILLER AND H. G . DRICKAMER University of Illinois, Urbana, I l l . m
3
The viscosities of a series of normal paraffin hydrocarbons have been measured over a temperature range down to the freezing point and into the supercooled region. The results show that the free energy of activation per unit volume is substantially independent of temperature except at the freezing point and below, where an increase is noted, probably due to increased molecular orientation. A relation i s shown between freezingpoints of compoundsand freezingpoint viscosities.
T
HE purpose of this work was t o investigate the viscosity-
temperature relations of a homologous series of paraffin hydrocarbons down t o the freezing point, as a possible contribution to the establishment of a kinetic theory of liquids. The cryostat used was a slight modification of the one described by Egerton and Ubbelohde (6),capable of holding + 0 . 0 5 O in the range 0" to -50' C. and *0.lo in the range -50" to -140'.C. The bath liquid was low-boiling petroleum ether, cooled with li uid air (6). %he Cannon-Fenske modified Ostwald viscometer was used.
Flow time generally was well above 200 seconds as recommended which gave an err& of *0.2% for relative viscosities and *0.5$ for absolute viscosities; the exceptions were pentane and hexane at 20" C . Flow time was measured with a stop watch, checked against the international time signal for an error of less than 0.2%. Pure grade n-pentane and n-heptane were supplied by Phillips Petroleum Company; n-hexane n-octane, n-decane, n-dodecane, and n-tetradecane were obtained from the Connecticut Hard Rubber Company. The normal paraffins were tested with potassium permanganate for a color reaction. n-Pentane and n-heptane showed no reaction, but the other paraffins gave a positive test. n-Hexane, n-octane, n-decane, n-dodecane, and n-tetradecane were shaken with concentrated sulfuric acid until no color formed in the acid layer. They were then washed five times with distilled water and dried over calcium chloride for several days. No chemical treatment was given to n-pentane and n-heptane. All compounds were distilled in a 2-foot column filled with small glass helices. The middle cut boiling over a 0.1 O C. range was retained for use in the investigation. All samples were dried over sodium wire for several weeks before being tested in the viscometer. The measured properties are compared with the best available data from the literature in Table I.
22
3 v a w
v)
-- --
-.-.--.-.-.__._._.
--
w -
- . .. - .
a 1823 2 *L
/
- - - - - _ - - - - - - - - - =>= .o. . - . . - . - , ,. - ..w *
a
0
ent of temperalure for each compound and depends only slightly on niolecular cveight, a t least for the compounds up to decane. There is a small but consistent increase in &/V for each compound near the freezing point, which would represent the increased activation energy necessary because of a certain degree of order developing in the liquid. As far as the experimental
.....- .-
=5
-I-,-,--I-l-I-
-...._...-... - ...0 .,.I - I- I - I*
---...I
I
6' c7
I
'8
I-
,
-I-I.I-I-
CIO l2 C 14
o
F. P.
through the freezing point and in the supercooled region. Whrn
DETERMINATION OF VISCOSITY
A11 viscosity determinations were started a t 20" C. and repeated a t lower temperatures until solidification occurred. The compounds were filtered through a fine sintered-glass filter and immediately loaded into the viscometer according to the technique described by the American Society for Testing RIaterials ( 2 ) . The bath was brought to constant temperature, and the viscometer placed in it,. The viscometer limbs were connected to a calcium chloride drying system to prevent entry of water vapor from the surrounding atmosphere. Ten minutes x a s allowed for the viscometer to come t o bath temperature, and readings were taken until constant within 0.2%. The bath temperature was then lowered and the procedure repeated. In the neighborhood of the freezing point,, runs were made at, 0.5" C. intervals or less. I n all cases it was possible to supercool at, least slightly and obtain viscosities below the melting point,, sometimes as much as 5" to G o C. The density data tabulated by Egloff ( 7 ) were used to calculate viscosities. In the temperature range where comparison was possible, the results agreed with those of Geist and Cannon (8) within 2y0 in all cases and within 1 in most cases. The viscometer was calibrated twice with conductivity water tvith a temperat,ure control within +0.02" C. The viscometer constants were determined at 40' and 20" C. and then extrapolated as a straight line over the temperature range of this investigation, as outlined by Cannon and Fenske (6). The calibration dat,a follow: Temp.,
c.
20.0 40.0
Time, Min. 4.155 2.725
Viscosity of HzO Poise Stoke 0,01007 0.01005 (3) 0.006360 (4) 0.006611
Constant, Stokc/Min 0,002424 0.002426
TABLE I.
n-Pentane n-Hexane %Heptane ri-Octane n-Decane n-Dodecane n-Tetradecano
TABLE 11. Temp., " C. / -
20.00 -19.9
-anI _ . n_
-100 0 -118.6 -120.7 -124.0 -126.0 -128.0 -128.7 -129.5 -130 0 -130.5 -131.0 -132.0 -133.0 -134 0 -136 0 20.00 -20.0 -40.1 -60.1 -80.2
-92.3 -94.3
EFFECT OF TEMPERATURE AUD MOLECULAR WEIGHT
By considering the flow of liquids as a rate process, Eyring ( 9 )derived the following equation for viscosity: haV AFT exp. __ V RT
17 = -
When solved for the free energy of activation per molc, this gives:
As might be expected, AFT depends on the temperature and the molecular weight of the component. However, the quantity AF'/V (activation free energy per unit volume) is proportional to the free energy necessary to make a hole of unit size. In Figure 1 A F S / Y is plotted against temperature for the various compounds. It appears that A F r / V is substantially independ-
1"
Con1i.iound
-90.3
Table I1 gives the results.
c.
PEOPERTIES O F PARAFFISS
--Literature
--s5,3 -95.7 -96.3
-96.8 -97.5 -98.5
>I.p.,
nD
1.3575 ( I ) 1,3749 ( 1 ) 1.3876 ( 1 ) 1 ,a975 ( 1 \ 1.4121 ( 7 ) 1.4218 ( 7 ) 1.4360 ( 7 )
O
-129.7 -95.3 -90.6 -56.8 -29.7
(1)
(1) (1) (I) (7) -2,s (7) 0.J0( 7 )
--Experivental-nY F.P., 1.3574 1.3744 1.3873 1.3875 1.4118 1 421-4 1.4286
0
c.
-130.5 -96.7 -90.7 -37.3 -30.5 -10.5
4.70
VISCOSITY O F PARAFFIXS .4T V A R I O C S TEJIPERATURES
Viscosity. Stoke
n-Pentane0.00371 0.00516 0.00783 0.0170 0.0297 0.0325 0.0364 0 0395 0.0431 0,0444 0.0462 0.0472 0.0482 0,0495 0.0528 0,0554 0 0581 0 C649
Viscosity, Poise 7
Temp., C. 7 -
0,00342 0,00232 0.00348 0.0125 0.0223 0.0245
0.0275 0,0299 0,0327 0.0337 0.0351 0.0359 0.0367 0.0377 0,0402 0.0423 0 0444 0 0499
-n-Hexane- -7 0.00470 0.00310 0.00691 0.00481 0,00892 0.00635 0.0123 0.00894 0,0187 0.0139 0.0184 0.0244 0.0195 0.0258 0,0209 0.0276 0,0216 0,0285 0.0219 0.0288 0.0223 0.0294 0,0226 0.0298 0.0231 0.0304 0,0239 0.0313
,---n-Heptane----------. 20.00 0.00599 0.00739 0,OO -20.0 0.00982 -40.0 0.0130 -60.1 0.0194 -80.3 0.0341 -82.3 0.0367 -84.3 0.0393
o.oo+io 0.00518 0.00683 0,00955 0.0146 0.0261 0.0281 0.0302
-88.4 -89.3 -90.3 -91.0 -92.0 -93.0
0,0358 0.0373 0.0390 0,0399 0.0417 0.0438
20.00
0.00
-20.0
-40.0
-;o.o
--a2 0
-54.0
-56.0 -57.0 -37.6 -58 0 --59.0 -60.0 -61.0 -62.0
Viscosity, Stoke
Viscosity, Poise
n-Octane--------. 0.00539 0.00767 0.00977 0.00703 0.0132 0.00968 0,0193 0.014,5 0.0185 0.0243 0.0193 0.0257 0.0270 0,0206 0,0286 0.0218 0,0295 0.0223 0.0300 0.0230 0.0235 0.0303 0.0241 0.0314 0.0248 0.0324 0,0257 0.0835 0 0343 0 . 026,;
---Decane---------. 20 on 0.00
n
0124
n
nnqn7
-10.0 -20.00 ,-24.1
.-26.1
-28.1 -29.1 -29.6 -30.1
-30.6
-31.1 -32.1 -33.1 7 -
20.00 10.00 0.00 -5.00
-7.0 -8.0
-9.0
-9.5 -10.0 -10.5 -11.0
-
n-Dodecane0,0199 0.0239 0 0295
0,0332 0.0348 0.0357 0,0366 0.0371 0.0375 0,0380
0.0385
n-Tetradecane-0.0300 20.00 15.00 0.0336 10.00 0.0379 0.0394 8.00 0.0404 7.00 0.0415 6.00
0.1048 0.0181 0.0226 0.0255 0,0268 0.0275 0,0282 0,0286 0 .o m 0,0294 0.0297
.__
,5,.50
5.00 4.80 4.50
0,0420
0,0425 0.0427 0.0431
0.0229 0.0258 0.0292 0.0306 0.0313 0.0321
0.0325 0.0329 0.0331
0 CJ331
INDUSTRIAL AND ENGINEERING CHEMISTRY
September 1949
-
I
1
I
n
0.06
-
LITERATURE C I T E D
-
-
’ 0.080-VISCOSITY
AT FREEZING POINT
OF Cg AND C , A T “NORMAL“ F R E E Z I N G P O I N T
__
X-VISCOSITY
-
W
: I
1I
t-
1
2 0.04 -U < I v)
2069
I
(1) Am. Petroleum Inst., Research Project 44 on Collec%ion,Analysis and Calculation of Data on Properties of Hydrocarbons,
Washington, D.C., Natl. Bur. Standards, 1945.
(2) Am. Soc. Testing Materials, “Standards on Petroleum Products
and Lubricants,” 1936. (3) Ibid., 1944.
(4) Bingham, E., and Jackson, R., Bull. Bur. Standards, 14, 75 (1918). ( 5 ) Cannon, M., and Fenske, M., IND. ENG.CHEM.,ANAL.ED., 10, 297 (1938). (6) Egerton, A,, and Ubbelohde, A,, Trans. Faradall SOC.,26, 236 (1930). (7) Egloff, G., “Physical Constants of Hydrocarbons,” A.C.S. Monograph 78, New York, Reinhold Publishing Corp., 1940. (8) Geist, J., and Cannon, M., IND. ENG.CHEM.,ANAL.ED.,18, 16 (1946). (9) Glasstone, L., Laidler, K., and Eyring, H., “Theory of Rate Processes,” New York, McGraw-Hill Book Co., 1941. (10) Kauzmann, W., and Eyring, H., J. Am. Chem. Soc., 62, 3113 (1940). RECEIVED September 15, 1948.
Figure 2.
Effect of Molecular Weight on Freezing Point Viscosity
crystallization occurred, it was almost instantaneous in time and in temperature. Furthermore, for tetradecane the liquid was frozen and then reheated to 0.1’ C. above the freezing point, where viscosity measurements were made which were repeated at higher temperatures. The viscosities obtained in the heating cycle exactly reproduce those obtained in the cooling cycle. This shows that viscositymeasurements do not indicate any greater orientation in the remelted liqujd than in the liquid before freezing. Figure 2 shows freezing point viscosities plotted against molecular weight. The viscosities of the compounds with even numbers of carbon atoms give a smooth curve; the viscosities of the odd numbered compounds are displaced from this curve. Kauzmann and Eyring (IO)suggested that the freezing point viscosity for normal paraffins is relatively constant for compounds 8crith five t o fourteen carbon atoms. The difference in viscosity between odd and even numbered carbon compounds seems to depend on the well known anomaly between freezing points of odd and even numbered chains of normal paraffins; odd numbered compounds have relatively lower freezing points than the corresponding even numbered compounds. This anomaly has been explained by statistical considerations, as there are fewer ways of fitting the odd numbered chains together in a crystal lattice than the even numbered chains. If the freezing points of the paraffins with even numbers of carbon atoms are plotted against molecular weight, it is possible to establish a “normal” freezing point for the odd numbered compounds. The viscosities corresponding to this normal freezing point for pentane and heptane were plotted as shown in Figure 2. The points fall essentially on the freezing point viscosity curve established for the chains with an even number of carbon atoms, an indication that, as far as resistance to flow is concerned, the compound is, in a sense, supercooled below its normal freezing point. NOMENCLATURE
F* = free energy of activation, calories per mole h k
N
= Planck constant
= Boltzmann constant
V
= Avogadro number = N k , calories/” K. = absolute temperature, = molar volume, cc.
q
= absolute viscosity, poises
R T
O
K.
CORRESPONDENCE Critical Constants, Density, a n d Viscosi ty-Correspondence SIR: Attention is called to an error in the paper by Arnold Boas [IND. ENG.CHEM.,40, 2202 (1948)l. Equation 4 is taken correctly from Souders. However, as the logarithms are to the base 10, the reduced form should appear in Equation 5 a5 log ‘tr = ( 1 0 ” ~ ~ ~ ) ( 1 0 - 2 . @ - )log ‘to Further, some confusion may result from the author’s statement that, “log qo is substantially constant for the nonpolar liquids studied.” It is constant only within about 10% as shown by a tabulation of qc in the paper of Uyehara and Watson [Natl. Petroleum News (Oct. 4, 194411.
E. Harvey Barnett 5873 Julian St. Louis 12, Mo.
,.....
SIR: The comment of Mr. Barnett is quite correct. Equation 5, page 2203, should have read as he has indicated. However, this equation is not used in any further calculation but is merely used to illustrate a method of reducing a n equation and to establish the universal constant, md,. The error was noted upon publication and was corrected in every reprint that has been sent out. rlC
Benzene Ethyl ether Ethyl propionate Toluene
312 268 284 a06
log
9c
2.49 2.43 2.45 2.49
Due to the units of q in Souders’ equation-that is, millipoisesthe constancy of log q c varies much greater than if we had used a relation with the viscosity in micropoises. Using a few values of qc as noted in Table I1 of Uyehara and Watson with ?a in micropoises it is seen that log qo is practically constant. However, it should be noted that, even with the constancy of log qc within 10% as pointed out by Mr. Barnett, we are still able t o establish the universal constant md, and predict critical density values from this constant. Hydrocarbon Research, Inc. 115 Broadway New York 6, N. Y.
Arnold Boas