568
INDUSTRIAL AND ENGINEERING CHEMISTRY
sorption and moisture-vapor transmission, but more accurate data are needed to confirm this possibility.) $n attempt was made, without success, to determine the equilibrium moisture of the films, Although it was not successful because of the inherent error of the method selected, it mas interesting to observe that the free films are much easier to prepare into test specimens a t high than a t low relative humidities. Subsequent to the blistering test, film to substrate adhesion was examined qualitatively. One hour after removal of the steel panels from the hot water, it was found t h a t lA, l B , l C , and 2A had negligible, if any, adhesion. Enamels 2B and 2C had some residual adhesion. This decrease in adhesion permitted the films to be removed from their steel substrates. It was found t h a t under-film corrosion had stained the back enamel surface. Based upon the fraction of the area which had been stained, under-film corrosion ratings were attempted. These are shown in Table XVI. CONCLUSIONS
There is some correlation, although not precise, of the results between the three types of water test as t o their deleterious effects on nitrogen resin-containing alkyd coatings. Composition of the nitrogen resin influences the results markedly. The urea resin is much poorer for resisting water exposure than is the melamine or triazine resin. Some improvement can be obtained by altering the composition of the urea resin, but formulation changes appear to be equally as effective. No clear-cut understanding of the mechanism of blister formation is as yet available. -4hypothesis based on these and the
Vol. 44, No. 3
observations of others is that adhesion when wet exerts a dominant influence. The possibility that a film swells upon water absorption and expands because of this absorption, thereby exerting a delaminating influence, is not a remote one. Much remains to be done to study this subject further, both in the way of fundamental studies and from the practical application point of view, also. It may be noted, for example, that such an important variable as film thickness has not been mentioned and it is possible t h a t factor also plays a major role in film performance ACKKOWLEDGMEYT
Acknowledgment for much help and assistance in conducting the experiments, discussing the results, and preparing this paper for presentation are extended to Emory Slaght and to Samuel Gusman. LITERATURE CITED
Boylnn and Wray, A S T M BUZZ.,141,53-5 (March 1949). Dixon, W. J., Ann. Math. Statistics, 11, 119 (1940). Elm, A. C., Oflcial Digest Federation Paint & Varnish Production Clubs, NO.267, 197-288 (1947).
Grinsfelder, H., Ibid., 312, 42-52 (January 1951). Kittelberger, W.W., IKD.ENG.CHEM.,34, 943-8 (1942). Kittelberger and Elm, Ibid.,38, 695-9 (1946). Ibid., 39,876-81 (1947). hlayne, J. E. O.,Oil d? Colour Chemists’ Assoc., 33,312-16,538-47 (JUIY 1950). Wirsching, R., Ibid.,33, 234-7 (1941). RECEIVED for review June 21, 1951. ACCEPTED November 14, 1951. Presented as p a r t of the Symposium o n ‘Crea, Melamine, and Related Resins before the Division of P a i n t , Varnish, and Plastics Chemistry, 119th hfeeting of the ERICAN AN CHEMICALSocIErY, Bost,on, &Lass.
BY LOW TEMPERATURE HYDRATIQN OF ALPHA-PHOSPHORUS(V) OXIDE R. N. BELL
L. F. AUDRIETH AND 0. F. HILL
Victor Chemical Works, Chicago Heights, I l l .
Unicersity of Illinois, Urbana, I l l .
YDRATION of the commercially available a- form of phosphorus(V) oxide may give a variety of products, depending upon the mole ratios of phosphorus pentoxide t o mater which are allo-ived t o react. The strong phosphoric acids, containing moIe than 72.401, phosphorus pentoxide and prepared by the high temperature hydration of phosphorus(V) oxide using P20s-Hz0 molar ratios greater than 1 t o 3, have been shonn t o consist of mixtures of poly- and metaphosphoric acids ( 3 ) . [The a- form of phosphorus(V) oxide is correctly represented as PlOlo. There are, however, three other recognized polymorphic forms which are more complex in nature. The empirical formula, PZO,, is used throughout this article, except where structural considerations demand otherwise.] It was therefore surprising t o find t h a t the low temperature hydration of a-phosphorus(V) oxide produces chiefly tetrametaphosphoric acid. Neutralization of this solution with sodium hydroxide has led t o t h e isolation of the tetrasodium salt, now commercially available for t h e first time (Cyclophos, Victor Chemical Works). Sodium tetrametaphosphate was apparently first prepared by Warschauer ( 1 1 ) , who obtained the compound by the treatment
of a suspension of the copper salt n-ith squeous sodium sulfide. The copper salb had been prepared by the interaction of powdered copper oxide with slightly more (5%) than t h e required quantity of phosphoric acid, followed b y slow and gradual heat,ing up to, but not exceeding, 450°C. Bonneman (4) prepared the compound by the 6ame procedure and determined its molecular weight by a cryoscopic procedure using fused sodium sulfate 10-hydrate. Bonneman also listed the characteristic x-ray diffraction spacings for both the 4-hydrate and the anhydrous salts. Three recent publications have dealt with the tetrametaphoEphate prepared by the Warschauer method. Conductance measurements of the acid derived from the salt by an ion exchange method have been presented as evidence for the existence of a discrete P a 0 1 2 ion (6). The cyclic structure of the tetrametaphosphate has been verified by x-ray analysis (1). Careful hydrolysis in alkaline solution has been stated t o yield tetraphosphate (10). Further evidence for the existence of a tetrametaphosphate has been presented by Raistrick and his coworkers ( 8 ) of the Slbright-Wilson Laboratories, who have prepared the sodium salt by a procedure similar to the method used by the authors.
March 1952
INDUSTRIAL AND ENGINEERING CHEMISTRY
569
lowering method using sodium sulfate 10-hydrate as the cryoscopic The experimental results presented herewith definitely prove solvent. Values of 395 and 415 were obtained in two experithat identical materials are prepared by the older Warschauer ments, corresponding t o a compound with the over-all composimethod and by the newer and simpler procedure involving low tion Na4P4012. The freezing point constant for the sodium sultemperature hydration of a-phosphorus(V) oxide. Hydrolysis fate 10-hydrate was found t o be 3.32, using sodium chloride for was found t o give triphosphate and orthophosphate. Electropurposes of standardization. This value was used in calculating metric titration of the acid prepared from the sodium salt by ion the molecular weights reported below: exchange confirms the formation of a meta acid. Distinguishing properties of the tetrametaphosphate ion are also presented. Molecular NazSOclOHzO, NarPdOln, Grams A Tf Weight Grams The most striking indirect evidence of the composition and 0.5 0 105 395 40 structure of the tetrametaphosphate is offered by the authors in 0 06 415 40 0.3 the presentation of a mechanism for the hydration of a-phosX-RAYDIFFRACTION DATA. Further proof of the identity of phorus(V) oxide. the product obtained by the low temperature hydration of phosTETRAMETAPHOSPHORIC ACID BY HYDRATION OF PHOSPHORUS phorus(V) oxide is offered by Tables I1 and 111, in which are pre(V) OXIDE. Fifty grams of or-phosphorus(V) oxide are added slowly t o 300 ml. of water below 15' C. The solution is stirred sented the x-ray diffraction data for t h e anhydrous salt and the vigorously with a mechanical agitator, so that the oxide is drawn two tetrahydrates. Data previously published by Bonneman into the water quickly in order t o avoid local high temperatures. (4) and by Thilo and Rate (IO) for the anhydrous salt are in The rate of addition is adjusted so that the temperature of the excellent agreement with values obtained by the authors. Recogsolution does not exceed 15' C. Stirring is continued until the oxide has been completely dissolved. Titration of such a solution nition of the existence of two polymorphic forms of the 4hydrate with sodium hydroxide indicates t h a t the hosphorus (V) oxide accounts for the discrepancies which are apparent from a comhas been converted largely into a metaphospioric acid. Using 50 parison of data of the earlier workers. It is apparent that grams of phos horus(V) oxide (0.352 mole as PzOJa total of 98 Bonneman had in hand a fairly pure sample of the high temperagrams of 3 0 g s o d i u r n hydroxide (0.735 mole) was required t o effect neutralization t o a pH of 7.0. This corresponds t o a molar ture form, whereas Thilo and Riitz obtained a product consisting ratio of sodium hydroxide t o phosphorus pentoxide of 2.09, largely of the low temperature form. Both workers used the whereas complete conversion t o a metaphosphate would have reolder Warschauer (11) method for the preparation of their quired a molar ratio of 2.00. materials, SODIUMSALTOF TETRAMETAPHOSPHORIC ACID. T o a solution of sodium tetrametaphosphate, prepared as indicated above, 30 grams of sodium chloride were added. The resulting solution was allowed t o stand overnight. Sodium tetrametaphosphate 10TABLE11. X-RAYDIFFRACTIONDATAFOR NarPIOll hydrate crystallizes from solution if the temperaure is kept below This Work Thilo and Riitz ( 1 0 ) Bonneman (4) 25" C. (A 4-hydrate may be obtained a t temperatures above d , A. Intensity d, A. Intensity d , A. Intensity 40' C.) T h e crystals were removed by filtration, washed with water at 5" t o IO" C., and air-dried. Yields of 60 t o 65% of vvs 5.60 5.67 vs 5.67 vs 5.11 VW5 theoretical were obtained consistently. 4.71 vw= d 4.36 4.34 M 4.36 S HYDRATES OF TETRASODIUM TETRAMETAPHOSPHATE. Observd 3.83 W 3.73 M 3.81 3.61 ing the IO-hydrate cyrstals under a polarizing microscope while vw M 3.35 S 3.46 9 3.41 heating on a micro hot stage indicates that crystal changes occur 3.11 M0 V8 2.92 vvs 2.94 vs 2.92 a t approximately 40°,54', and 100" C. Heating curves show an M 2.82 exothermic break betFeen 400" and 450" C. and an endothermic S 2.76 M 2.74 M 2.73 2.51 M 2.51 S 2.52 M break a t 620' to 630' C. Dehydration experiments demonvw 2.48 W 2.43 W 2 . 2 8 W 2 . 3 2 W 2 . 2 9 strate that a 4hydrate is obtained a t 40' C. and that this par2.21 VW W 2.21 ticular hydrate exists in two polymorphic forms: a low temperaW W 2.16 2.17 W 2.10 W 2.12 VW d 2.11 ture form which is stable below approximately 54' C., and the M 1.99 VVW d 2.04 W 2.01 1.95 vw high temperature modification which forms above this temperavw 1.87 ture, but does not revert on cooling. The 4-hydrate loses water vw W 1.83 1.84 W 1.76 W 1.77 around 100' C. to give the anhydrous compound. The break 1.72 1.67 vw 1.67 near 400" C. is due t o conversion t o trimetaphosphate. The vs 1.64 M 1.64 x-ray pattern of material heated t o 450' C. is identical with vvw 1.54 vw 1.53 vw vvw 1.49 sodium trimetaphosphate prepared by heating monosodium dihyw 1.42 1.38 d 1 . 3 5 drogen orthophosphate a t the same temperature. yvw 1.32 REFRACTIVE INDEXES OF VARIOUSMETAPHOSPHATES. The 1.27 1.17 ? various hydrates are readily identifiable by their refractive 1.16 1.15 VW 1.14 vvw indexes, which also serve t o differentiate between tetra- and tri1.12 vw 1.10 1 metaphosphates. Refractive indexes accurate to within 0.002 vw 1.07 1 1.07 are presented in Table I. 5 Correspond to strong lines in high temperature phase of 4-hydrate.
TABLE I. REFRACTIVE INDEXES OF VARIOUS METAPHOSPHATES NarPaO~z.lOHnO 1.420 1.424 1.436
NaaP401z.4Hz0 High temp. form 1.438 1.459 1.468
Low temp. form 1.441 1.459 1.472
NaaPnOs.6HzO NarPaOo.Hz0 1.433 1.442 1.446
1.499 1.500 1.504
Intensity symbols. VVS = very very strong VS = very strong 8 = strong M = moderate
W = weak VW = very weak
VVW = very very weak d = diffuse
~
COMPARATIVE TITRATION CURVES FOR METAPHOSPHORIC ACIDS. Further verification of the discrete character of the product obtained by low temperature hydration of a-phosphorus(V) oxide and of the cyclic structure of both tetra- and trimetaphosphates MOLECULAR WEIGHTDETERMINATIONS OF ANHYDROUS SODIUM was obtained by the eleclrometric titration of the pure acids prepared from the sodium salts by ion exchange. Both acids can be TETRAMETAPHOSPHATE. The identity of the sodium tetrametatitrated as strong acids with the empirical formula HPOa. The phosphate with the product previously prepared by Bonneman titration curves for tetra- and trimetaphosphoric acids (Figure 1) ( 4 ) and by Thilo and Riitz (IO) was first checked by carrying out show all hydrogen atoms to be strongly ionized and of equal determinations of apparent molecular weight by the freezing point
570
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 44, No. 3
80
TABLE 111 X-RAYDIFFRACTIOX DATAFOR iTa*P40124H20 Low Temperature Phase This Work Tkilo and Ratz ( 1 0 ) d . A. Intensity d , A. Intensity 7.73 6.22 4.79 4.33 4.32 3.84 3.83 3.24 3.27 3.17 3.08 3.09 2.81 2.81 2.74 2.67 2.62 2.62 2.54 2.47 2.42 2.38 2.38 2.21 2.26 2.19 2.10 2.09 2.01 1.96 1.92 1.91 1.86 1.85 1.83 1.79 1.76 1.76 1.68 1.68 1.65 1.58 1.61 1.55 1.53 1.53 1.49 1.50 1.48 1.44 1.44 1.42 1.39 1.39 1.37 1.37 1.34 1.31 1.32 1.27 1.25 1.25 I . 24 1.24 1.22 1.22 1.19 1.19 1.18 1 16 1.17 1.15 1.15 1.14 1.13 1.11 1.07 1.05 1.02
High Temperature Phase This Work Bonneman (4) d , A. Intensity d , A. Intensity 9.36 9.47 VS 6.16 6.39 v 5.65 w 3.10 5.02 S 4 . 85 4.70 4.40 4.40 M 4.24 3.78 hI 3.86 3.48 3.49 vw 3.31 3.20 3.20 W 3.09 3.11 vs 3.05) 2.90' 2.78 2.82 hl 2.71 2.63 2.62 S 2.59) 2.50 2.47 2 48 &I 2.38 2.40 2.32 VW 2.27 2.21 2.23 w 2.18 2.11 17%2.10 2.06 VW 2.05 2.01 2.01 vw 1.97 1.98 VW 1.92 1.94 UT w 1.87 1.89 1 85 W 1.84 W 1.81 1.82 1.77 1.76 VW 1.71 1.72 W 1.68 1.69 U' 1,67 1.62 1.60 1.58 1.57 w 1.54 1.53 1.52 w 1.52 W 1.49 1.50 w 1,46 1.46 1.44 w 1.44 1.42 1.39 1.37 1.34 1.32 1.31 1.29 1.27 1.25
6
801
w
5.);
w
T I M E I N DAYS
Figure 2. Hydrolysis of Sodium Tetrametaphosphate 170solution in water a t 100' C.
weakly acidic hydrogen atoms t o be present. The curve for the acid prepared from the metaphosphate glass reveals the presence of relatively few weakly acidic hydrogen atoms; this indicates that solutions of the metaphosphate glam consist of long-chain units or of a combination of ring and chain structures.
ORTHOPHOSPHATE
T R IFriOSPHATE
LA h ETRAMETAPHOSPHATE
TIME IN HOURS
1 34
Coirespond t o strong lines in high ternperature phase. Intensity symbols same as for Table 11. 0
1.23 1 20
JV
1I !g 15
VTT ,v_\T n
1 12
\v
1 14
1 11 1 10 1 09
'
Q
1% solution in 1'3% sodium hydroxide a t 100" C.
TI7
M' IT
rv
strength. Titration curves for pyro-, tri-, and the so-called "hexametaphosphate" are also presented in this figure for comparison. The curves for the two poly acids show both strong and
HYDROLYSIS O F TETR~METAPHOSPHATE. The hydrolysis of the tetrametaphosphate was studied in both alkaline and acid solutions. As hydrolysis of the tetrametaphosphate ion results in a lowering of the pH, it would seem from the data preBented that such hydrolysis is catalyzed by the hydrogen ion, since the rate of conversion increases with respect to time. When the pH of the solution is maintained between 5.8 and 6.4, hydrolysis is much slower (Figure 2). At a temperature of 100' C. the addition of 1yo sodium hydroxide to the solution of tetrametaphosphate causes complete disappearance of tetrametaphosphate in about 4 hours (Figure 3). The analytical data indicate that the principal hydrolysis reaction can be represented by the following equation:
+
o ML OF 01 r\ NaOH
Figure 1. Titration of Phosphoric AcidsrPrepared by Ion Exchange 0.20 gram of acid
k
Figure 3. Hydrolysis of Sodium Tetrametaphosphate
+
+
P*On---40H- +PaOio----PO*--2Hz0 That hydrolysis proceeds in accordance with the above equation is further indisputable evidence of the existence of a discrete tetrametaphosphate ion. Rupture of the eight-membered PaOl9 ring should give t h e tetra,phosphate as the intermediate product. No direct evldence of its formation under the given conditions was:found in'the present work, although it is conceivable that tetraphosphate would respond t o the same analytical test used in the determination of tri hos hate by the Bell ( 2 ) method. The triphosphate and orthopgospEate do, however, represent the most probable further hydration products of such a linear tetraphosphate, although formation of pyiophosphate could also be postulated. P4Oi2----
-1- 20H-+
[P,Oi3------ +HzOI---t POa---
+
p 0 ----_ 3
io
March 1952
TABLE IV. REACTIONS OF TETRAMETAPHOSPHATE WITH SOLUTIONS OF METALLIC SALTS PH 7.0 4.7 6.6
Salt Used
Preoipitate
Forms slowly Slight None
None
6.3 4.6 3.3
None None
Y
A
P-0’ bond distance indicates that polar character in the bond is predominant. This shorter bond distance is reflected in the chemical roperties of phosphorus(V) oxide, since such bonds are pi-esumabyy very stable and the thermal stability and resistance t o reduction of P 4 0 1 0 are well h o w n .
STEPWISE HYDRATION OF PaOla. The initial attack of water on the P4O10molecule may be assumed to cause rupture of a P--0-P bond with the resultant formation of only one possible hypothetical intermediate, which is represented graphically by I in Figure 5. Addition of the second molecule of water may be assumed to bring about bond rupture of I to give two possible products, IIa and IIb. Each of these is characterized by the empirical composition (HPO&, but only one, the tetrametaphosphoric acid, IIa, is actually known. Tetrametaphosphoric acid may be represented structurally as an eight-membered ring with alternate phosphorus and oxygen atoms and as a cyclic condensation product of four PO1 tetrahedra, each of which shares two oxygen atoms with its immediate neighbors.
C
B
Figure 4.
Structure of PaOlo Molecule
COMPOSITION FORMULA
4. Planar representation B . Ball and rod model C. Tetrahedral representation p4010
Sodium tetrametaphosphate, like the trimetaphosphate, precipitates very few heavy metals. Observations with respect to formation of precipitates and the pH, after the addition of 10 ml. of a 1% solution of sodium tetrametaphosphate to 50 ml. of a 5 t o 10% solution of the metallic salt, are recorded in Table IV. The tetrametaphosphate also resembles the trimetaphosphate in its inability to complex calcium ions; it resembles the so-called “hexametaphosphate” in its ability t o precipitate protein. This property suggests a possible use for the tetrametaphosphate in the tanning of leather. MECHANISM OF HYDRATION OF ALPHA-PHOSPHORUS(V) OXIDE
STRUCTUREOF ALPHA-PHOSPHORUS(V) OXIDE. The alpha form of phosphorus(V) oxide has been shown t o exist as discrete PaOlo molecules both in the vapor state and as the solid. It may be depicted structurally as consisting of four PO4 tetrahedra connected through six P--0-P linkages. The molecule may also be regarded as a tetracyclic condensed system of four six-membered rings, each containing alternate phosphorus and oxygen atoms. The bond distances and valence angles which characterize this form of phosphorus(V) oxide have been determined by Hampson and Stosick (6) and are reproduced in Table V.
TABLE V. BONDDISTANCES AND VALENCE ANGLESOF PHOSPHORUS(V) OXIDE(6) P-P P-0 P-0‘
=
571
INDUSTRIAL AND ENGINEERING CHEMISTRY
2 . 8 4 zt 0.03 A.
= 1.62 f 0 . 0 2 A . = 1 . 3 9 f 0 . 0 2 A.
OPO = 1 0 1 . 5 f 10 f 1; OPO = 116.5 =t1 POP, = 123
Such a structure may be depicted on a plane surface as indicated in Figure 4, A. However, such a presentation is apt t o be misleading, as it does not show directly the probable equivalency of the P-0-P bonds. A model in perspective, as depicted in Figure 4, B, is more revealing. To simplify the approach to this problem, it is still more convenient to depict the PaOlomolecule as a plain tetrahedron, Figure 4,C, each edge of which represents a P-0-P bond. All P-0-P bonds in the P*Olo molecule are equivalent and of equal strength. There are two different types of oxygen atoms in the molecule: those which form a bond between two phos horus atoms and those which form a bond to only one phospEorus atom. The latter are represented in Table V by the symbol 0’. The shorter
SIMPLEST PERCENT ACID p40lO COMPOSITION
100
94. I
00.0
84.0
79.7
75.9
72.4
Figure 5. Graphical Representation of Mechanism for Stepwise Hydration of a-Phosphorus(V) Oxide
When the third molecule of water is allowed t o react with the P ~ O ~a Ovariety , of products is obtainable, depending upon which of the structures, IIa or IIb, is involved, and which of the P-0-P bonds in each of these intermediates is broken. Only one product, the straight-chain tetraphosphoric acid, IIIa, could be obtained by hydration and rupture of the ring structure assigned t o tetrametaphosphoric acid. The hypothetical intermediate I I b could react with water in one of three ways, (1) to give an equimolar mixture of ortho- and trimetaphosphoric acids, represented graphically by IIIb, (2) t o form the straight-chain tetraphosphoric acid IIIa, or (3) t o form a n “isotetraphosphoric” acid, depicted graphically by IIIc. The reaction of 1mole of P4010with 3 moles of water could yield a mixture containing orthophosphoric acid, trimetaphosphoric .acid, tetraphosphoric acid, and an unknown iso-(branched-chain) tetraphosphoric acid, even though the empirical composition is represented by the formula H6P4018. A complete analysis of the hydration process leading eventually t o the formation of orthophosphoric acid is presented in graphical form in Figure 5. Although this mechanism represents an ideal treatment, it emphasizes the fact that no reaction involving treatment of phosphorus(V) oxide by less than a 3 to 1 molar ratio of
572
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
H,O-P,OI will lead to products corresponding to, or derivatives of, the formula for the simplest acid. This had already been demonstrated experimentally by a study of the composition of the strong phosphoric acids (S), and also accounts for the complex nature of solvolytic products which are obtained when a-phosphorus(V) oxide is subjected t o treatment with ammonia, amines, alcohols, and ethers. ACKNOWLEDGMENT
The authors desire to express their appreciation for the advice and help offered to them by Howard Adler, Willard H. Woodstock, and Benjamin Toubes of the Victor Chemical Works Research Laboratory, and t o thank hI. J. Klein (Y),H. H. Rogers (9), and Ralph Gher, Jr., for a confirmatory study of tetrametaphosphoric acid and its salts.
Vol. 44, No. 3
LITERATURE CITED
(1) Andress, K. R., Gehring, W., and Fischer, K., 2. anorg. Chem., 260, 331 (1949). (2) Bell, R. N., Anal. Chem., 19, 97-100 (1947). (3) Bell, R. N., IND. ENG.CHEM.,40, 1464 (1948). (4) B o n n e m a n , P., Compt. rend., 204, 865 (1937). (5) Davies, C. W., and Monk, C. B., J . Chem. SOC.,1949, 413. (6) H a m p s o n , G. C., a n d Stosick, A. J., J . Am. Chem. SOC.,60, 1814 (1938). (7) Klein, M. J., thesis, University of Illinois, 1948. (8) Raistrick, E., personal c o m m u n i c a t i o n ; Discussions Faraday Soc., No. 5, 234 (1949). (9) Rogers, H. H., thesis, University of Illinois, 1949. (10) Thilo, E., and Rats, R., 2. anorg. Chern., 260, 255 (1949). (11) Warschauer, F., Ibid., 36, 137 (1903).
RECEIVED for review September 5, 1950. ACCBPTEDOctober 11, 1951 Abstracted in part from a thesis submitted to the Graduate School of the Cniversity of Illinois by Orville F. Hill, in partial fulfillment of the r e quirements f o r the doctorate, in January 1948.
Cogelled Chromia-Alumina Catalyst for Naphtha eforming EVALUATION OF HYDROCARBON TYPES AND NAPHTHAS E. C. HUGHES, H. Rl. STINE, AND H. A. STRECIdER Standard Oil Co. (Ohio), Cleoeland, Ohio
S. C. EASTWOOD, C. L. GUTZEIT, W. A. STOVER, AND S. J. WANTUCK Socony-Vacuum Oil Co., Znc., Paulsboro,
N.J .
ATALYTIC reforming of petroleum naphthas is increasing in 10). The reactions by rrhich aromatization occurs are dehydroimportance as a process t o provide gasoline of high octane cyclization, in which paraffins undergo dehydrogenation and number, aromatic concentrates for aviation gasoline, and pure cyclization, and the dehydrogenation reaction, in which cycloaromatics. The first paper of this series (8) showed that a paraffins are converted t o analogous aromatic compounds. Hercoprecipitated catalyst composed of chromia and alumina was ington and Rideal (6) have s h o r n that not only cyclohexanes but of unusual interest for this purpose. One outstanding charactercyclopentanes with suitable side chains undergo this reaction, istic of the catalyst was that it enabled the coke deposition on the the latter after first undergoing an isomerization step. catalyst to be controlled to a low level by a lower partial pressure OPERATING VARIABLES of hydrogen than is used with any other catalyst. This chroniia-alumina catalyst has been developed into a The primary operating variables are temperature, space cogelled bead form, mechanically strong and highly resistant to velocity, pressure, ratio of recycle gas to naphtha and ratio of wear, a t the Paulsboro laboratory. A separate paper fioin this catalyst to naphtha, which will be discussed in detail in a forthlaboratory will describe the development of a naphtha reforming coming paper. process utilizing this catalyst. I n the course of the work on chroniiaalumina reforming catalyst, straight-run TABLE I. PROPERTIES OF REFORUER CHARGE STOCKS and coker naphthas from many crudee California Straight-Run Illinois Straight-Runb have been tested over a range of operatWest La + coker East + coker Tibuing conditions. Some substantial differTexas Gloria gasolinea Texas gasoline Heavy Petrolea ences have been noted between naph52.4 50.7 53.8 53.7 52.1 45.9 46.7 54.4 Gravity, PPI thas from various sources. As the first A.S.T.M. distillation, F. 254 205 217 I.B.P. 166 209 226 206 280 paper described the results of but one 278 270 267 241 10 vol. 7 264 252 223 284 306 316 323 264 288' 50 vol 9 241 293 338 naphtha, it is believed of interest to show 348 387 376 314 339 380 90 "01: % 312 279 425 380 E.P. 417 362 400 340 the results on the others. Many of the 352 401 99.0 98,O 98.5 98.5 99.0 97.5 Recovery, VOI. % 98.5 99.0 naphthas tested give yields much higher 1.3 0.8 1.2 0.6 0.8 1.6 Residue, vol. 7 1.0 0.7 33 33 64 52 48 55 60 44 Octane NO., F-1 cfear than the one reported in the first paper. 0 .03 0 . 5 6 . . . 0 . 0 5 . , , 0.02 0.15 Sulfur, wt. % ' 0.17 Relationships are drawn between naphtha Composition, vol. % 12 9 12 17 16 17 9 Aromatics 19 composition and the results obtained, in 8 0 2 13 1 1 1 0 Olefins 47 42 36 40 20 29 63 Naphthenes 33 order to explain the differences observed. 44 51 66 23 48 54 20 Paraffins 47 The reactions which hydrocarbons unBlend cons/sts of 60 vol. % ' straight-run naphtha and 40 vol. % coker gasoline. dergo in the presence of a chromium oxide b Blend consists of 80 vol. % straight-run and 20 5.01. % ooker gasoline. catalyst have been reported (1-4,6, 7 , 9 , 0
