INDUSTRIAL AND ENGINEERING CHEMISTRY
October, 194!3
in the over-all economic picture. Consequently, considerable attention is paid to the design of such a column in the selection of the optimum reflux ratio. By utilizing the two-component equilibrium curves, much of the tedious algebraic calculations have been eliminated. The following molal material balance for an isobutane tower, taken from the design of a commercial unit, is used to illustrate the method of application: Feed
Molw/hr. C:Ha ieo-C~Hle n-C&e Alkylate
98.9 1272.8 467.4 251.0
2090.1
Overhead
M~IWIHL. 98.0 1203.5 161.1
Bottom
Mole&.
....
9.3 816.8 251.0
.... -
-
1513.5
576.6
To illustrate this case, the following simplifying aasumptions are made: (a) Feed enters as liquid a t the boiling point; (a) constant molal reflux is assumed in each section of the column; (c) constaut relative volatility, a,is used as follows: a 2.5 1.28 1.o
0.05
For the solution of this problem, a two-component diagram is drawn first, ulling equilibrium data for iso- and n-butme. This is curve b of Figum3. This curve is then wed as a guide in the selection of a preliminary reflux ratio since the true multicomponent equilibrium curve will be outside this curve but close to it. A reflux ratio of 4 to 1 is selected, and several equilibrium points
959
are calculated for the top and bottom sections of the column by direct equilibrium calculations. Equilibrium points corresponding to the feed tray and the bottom tray of the fractionating section are now calculated in the manner explained in the previous paper, making use of the fact that alkylate and propane will be essentially at their theoretical minimum concentrations on the feed tray and tray above the feed, respectively. The calculated points for the light key component and lighter (propane plus g diagram as in Figure 3. isobutane) are now plotted on an 5 If deemed advisable, additional equilibrium points in the vicinity of the feed tray may also be calculated directly in order to define accurately the position of the equilibrium curve in this section. With the equilibrium curve completely defined and drawn in (curve a of Figure 3), the theoretical number of trays is stepped off in the conventional manner. The solution of the reference problem showed thirty-one theoretical trays plus a re boiler and total condenser which checked exactly with a stepwise algebraic solution. Other solutions within the accuracy of the equilibrium data can be determined for moderate changes in reflux ratio, using the previously determined equilibrium curve. The examples wed for iIlustration in the present paper me based on problems in which a relatively sharp separation is desired between adjacent components. The method has also been applied Ruccessfully to problems involving a split key component.
-
LITERATURE CITED (1) Brown, G.G.,Singer, 8.C., and Wilson, R. R., IND.Em.CEIDY,, 28,824 (1926). (2) Jenny, E'.J., Tram. Am. Inst. C h .E n ~ r e .35, , 636 (1939). (3) Obryadchakoff,€3. N., IND.EHQ.CHBY..24,1166 (1932). (4) Shiah, C. D., R 4 ~ i m Natural Gasoline Mfr., 21, 182 (1942).
1
0
Thermoplastic Laminates C. W. EURENIUS, R. H. HECHT, WILLIAM KOCH, AND H. C. MALPASS Hercules Powder Company, Wilmington, Del. Thermoplartic laminntee prepared by bonding fillers such am cloth and paper with cellulose aretnte and ethyl cellulose plastics w o r e evaluated in direct comparison to commercial phsitolii4onded laminutea. They were found to POSIWDR a r t urtttwal degree of toughnear, and were readily drawn i n l o contplex shapes with inexpensive equipment when softenrd b y heat. They have the additional advantagesof quic*kand easy fabrication arid an unlimited range of color Imnsihiliiien. Pmpertim nucsh as low-temperature Bexihiliiy, itnpwt smenqth, modulitr, of elantirity, water a h r p t ion, uttcl electrical irtwulrrt ion behavior can be controlled b y proper aelcction uf t&e h n d i i y plastic and the filler mutsriul.
T
HE art of combining plastic materials with fibrous fillers is old, for some threa decades ago fabric-filled phenolics were developctd and placed on the market for electrical insulating and mechanical ww. Other thermosetting binders, such as the ureas and melamine, have expanded into laminar constructions, and during the war the so-called contact resins have come into prominence. All of these materials have their individual sets of properties and charactaktics which tend to direct them into specific
end uses where they logically belong. However, even though they differ in many respecta, they have one thing in common; that is, they are thermosetting and must be polymerized by heat or a combination of heat and pressure. Furthermore, being threedimensional polymers, they have relatively low impact strength and must depend to a large extent on the filler for impact strength or shock resistance. During the past year this company has investigated combinations of celluloRe derivative plastics, which are noted for their inherent toughness, and various types of fillers, such as fabrio, paper, and asbestos paper. The purpose of this article is to present some of the results of this investigation. Two different types of laminar constructions were examined; the differences were primerily in the methods of preparation rather than in the materials employed. Low-pressure combinations of thermoplastics and cloth, which were studied in some greater detail, were p r e pared by prwsing a uandwich lay-up of plastic sheeting and filler. A simultaneous application of heat softens the plastic sufficiently to allow it to flow into and adhere to the cloth under the exerted pressure. These are called low-pressure cosstructions to die tinguish them from the other type which is based upon the successive lay-up of several plies of wet-plastic-impregnated cloth over a form, and which depends upon solvent r e l m e for setting
Vol. 37, No. 10
INDUSTRIAL AND ENGINEERING CHEMISTRY
960
TARLE I.
P H Y S I C A L PROPERTIES OF
THERMOP1,ASTIcLAMINATES
(Ratio of plastic t o cloth, 70:30; type of cloth, %ounce drilI) Cellulose Acetate Ethylcellulose Formula 1, aoft Formula 3, hard Formula 4, soft Formula ti, hard ~
P 4
Imrd iinpact strength, ft.-lb./in. nobrh, rdgenihr Flexural strength. !b./ljq. in. Modulus of elasticity in flexure, lb./sq. in. X 108 Tensile strength, lb./sq. in. Elongation, % Modulus of elasticity in tension, Ib./sq. in. X 102 Water absorp,tion. % 100% relative, humidity, 168 hr., 77' F. 24-hr. immersion" Thermal conductivity, cal./sec./sq. cm./" C./cm. X 10 -6 Thermal expansion, in./in./" C. X 10-5 Sp. gr. a t 77' F. Dimensional stability, % change at 80% R.H., 168 hr.
680 14 500 75 16
P
L
P
L
7.3 1600 41 4300
2.9 9600 230 7200 26 260
5.5 10400 260 8040 8 390
400 5.7 460 37 9
4-
4.36
6.80 3.30
10
79
....
.. ..
1:26
74 6 1.31
..
-_- - .
lAn0 F
... .
Lengthwise Edgewiee Flat win? '1
L
..
.. ..
1:32
.. .. ..
54 3 1.33 -0.34 -0.49 +3.0
___..-
Grade C
P
L
5.8 1900 62 2940 7 59
4.0 5000 108 3100 15 140
5.4 5900 150 5370 7 337
3.9 13000 440 8000 1.3 850
3.66
..
7.99
4.30
5.78 2.00
4.93
,.
71 5 1.16
54 7 1.13
43 2.9 1.35
-0.36 -0.49 +0.60
+ O . 18 +0.20 +0.67
1:06
.. .. ..
..
..
.. 1:09
.. ....
..
Determined by A.S.T.M. D570-42: edges were sealed.
were cross-laid in order to obtain balanced physical properties i i i both directions. As far as possible, A.S.T.M. tests were employed. In tests requiring specimens with a greater thickness than f / g inch, composite specimens were used in accordance with A.S.T.M. practice.
into a rigid unit. The latter type will be disrussed later in thk report. LOW-PRESSURE LAMINATES
Three primary factors are recognized as having a n influenee on the properties of thermoplastic laminates: plastic composition, type of filler, and ratio of plastic t o filler. The broad range of properties obtainable by variation in plastic composition is founded on the ability of thermoplastics to be modified with high or low percentages of plasticizer giving soft and flexible or hard and rigid plastics.
EFFECT O F PLASTIC COMPOSITION ON PROPERTIES
Table I presents some properties of thcrmoplastie laminates and compares the unfilled plastics, designated P , and the laminates based on these plastics, designated L. The data, obtained on a commercial Grade C phenolic laminate are also included. "Grade C" is the National Electric Manufacturer's Association designation for the high-impact cloth-phenolic combination which is generally recommended for mechanical uses. One method of interpreting the data in Table I is to observe the effect of cloth when combined with the different plastic binders. Whether the plastic formulation is hard or soft, or whether the bindcr is based on cellulose acetate or ethylccllulosc, the general effcct of tho fabric filler is one of increasing the impact, flexural, and tcnsilc strengths, the modulus of elasticity in flexure and in tension, and the thermal conductivity. On the other hand, thermal expansion and percentage elongation decrease as cloth is combined with plastic. Water absorption, as would be expected, is greater in the laminates than in the unfilled plastic. Data on all compositions were obtained by subjecting the test specimens to 1 0 0 ~relativc o humidity for 168hours a t 77" F. This procedure was used because a broad range of plastics was being tested; some contained large percentages of plasticizer while others had very little. Immersion of these materials in water, as required by A.S.T.M. Method D570-42, would result in relatively large amounts of plasticizer loss for some of the soft specimens, with the possibility
I n the examination of the effect of formulation on the properties of laminates several plastic compositions, based upon cellulose acetate and ethylcellulose of varying degrees of plasticization, were combined with an 8-ounce drill fabric under pressure and heat. 'This drill cloth weighed 7.7 ounces per square yard, with a 72/60 thread count, 61.570 of the weight being in the warp threads. The cellulose acetate, type PH-I, had a combined acetic mid content of 52-53.570 and a viscosity of 85-120 seconds by the conventional falling-ball method in a 20% solution with 9O:lO acetone:ethanol. The ethylcellulose (type N-100) had an ethoxyl content of 46.8-48.570 and a viscosity of 100 centioises in a 5Y0 solution, the solvent being 80:20 to1uene:ethanoi. hese two cellulose derivatives are standard materials used in commercial plastic work. The ratio of plastic to cloth was held constant at 70 parts plastic to 30 parts cloth by weight, which R'BS found to represent an optimum combination of properties. The materials to be laminated, consisting of alternate layers of plastic sheeting and fabric, were placed between 3/,@-inch,polished, chrome-plated steel platens in a heated press .which was closed, but no pressure was applied during the initial heating period of 6 minutes. The temperature ranged from 275' F. for the very soft formulas to 350" for the hard formulas. Pressure was then applied and maintained while the press was cooled. This required from 10 to 12 minutes. Pressures ranged from 100 pounds per square inch for the soft formulas to as high as 400 pounds for the hard formulas. The alternate layers of cloth
5
~~~
OF FABRIC TYPEON PROPERTIES OF THERMOPLASTIC LAMINATES TABLE 11. EFFECT
Type of cloth Count, Breaking strength A.S.T.M. raveled strip method A.S.T.M. grab niethod Weight oa./yd. No, of iayers of cloth Isod impact strength. ft.-lb./in. notch, edgewise Flexural strength, Ib./sq, in. Tensile strength, lb./sq. in. Elongation % Elastio mohulus in tension lb /sq in X 108 Water absorption (100% d.H:, 168 hr. 77' F J , %
(Ratio of cellulose acetate PH-1 t o Santicizer M-17, 75 t o 25; Ratio of plastic t o cloth, 70 t o 30) Balloon Basket Chafer Chafer Harvester Drill Airplane Weave Duck Duck Hose Duck Duck Cloth 80 X 80 104 X 98 35 X 25 23 X 23 26 X 22 32 X l 6 l / ~ 80
x
80
...
4.0 8 2.7 16,500 11,200 6.9 900 5.7
120
x
120
...
5.6 6
3.4 17,200 15,100 3.5 1000 6.4
92 2 . 6 1 5.6 6 2.5 14,000 8,700 4.5 610 7.0
201 X.191 14.0 2 3.6 12,400 7,300 7.6 430 7.2
119 X'122 9 3 3.5 12,100 8,000 6.6 440 7.2
546 X'432 31.5 1 5.2 12,700 7,400 11.9 350 6.5
... ... ... 8
4 5.5 10,400 8,040 8 390 6.8
Odober. 1945
INDUSTRIAL AND ENGINEERING CHEMISTRY
961
Figure 1. P a m of Drawing Men,drel ( l e f t ) end Assembled (right)
thet wster absorption data would not be comet even though the procedure described in D5'70-42 is suppwed to compensate for such Io=. However, limited data were obtained on formulas 3 and 6 according to D5'70-42 in order to obtain comparative values. The edges of the specimens subjccted to the high humidity teat were not scalod 80 that the specimens would come to equilibrium with the humid stmosphcre within the test period of 188 hours. haling the edges by costing them with a solvent solution of the plastic with whieh the laminate was made would greatly lower rater absorption hy reducing the wicking action of the fabric. Another interpretation of Table I is the sndysis of the effect of plastic formulations on physical properties. This is done by comparing the two L columns under formulations 1 and 3 for cellulose acetate or the L columns under formulas 4 and 6 for ethylcellulose. The general effect of decreasing the plasticizer content, or increasing the cellulose derivative in the bonding plastic, is to lower impact strength and thermal conductivity; on the other hand, flexural snd tonsilestrengths and elastic moduli in flexure and in tension increase sub8tsntially. The thermoplastic laminate which shows the closest degree of similarity with the Grade C phenolic lminste is baaed upon the hsrd celtulose eeetate formulation. The phenolic atmctum bas higher Rexurd strength, higher modulus, lower water absorption, and slightly lower thormsl conductivity while the thermoplastio oonstrootion has greater impact strength and higher elongation
The two are oompaiahle from the standpoints of tensile struai(th and thermal expansion. In ooneluding this phase of investigat.ionwhere gdastie formulation and its effecton physical properties have been covered, it is pertinent to point out that the range of plastic cornpitions, from soft to hard, ia not complete. The limit on the soft end of the rango has been approached and ia determined by plasticizer retention. Since the maximum temperature availahlo in our laboratory equipment is approximately 350"F.,and since this is slso the maimurn with l a w commercial Iminsting presses, the hsrd formufationa contain, as Littlc plasticizer as pmihle, consistent with proper impregnation st 350" F. and up to 400 pounds per square inch pressure. Attempts to obtain impregnation of harder plastics with greeter pressure rcre unsuccessful due to lateral Row of the plastic and tearing of the fabric. When higher temperatures an:made svailahle in commercial equipment, tho use of harder formul&ms will result in greater tensile and flexural strengths and greater elastic moduli, all of which &re important properties in msny end we8. However. research with fabrim other than Sounce drill cloth has resulted in considerable improvement in theae properties (Table 11). EFFECT OF FABRIC TYPES ON PROPERTIES
With plastic composition and ratio of plastic to cloth held constat, resulta in.Tshle I1 indicate that the type of febric has a definite bearing on the physical properties of tho laminated strocture. For instsnec, the use of balloon cloth resulted in laminates having greatly increased tensile and flexiirsl strengths TULE 111. ELECTRICAL PILOPERTIES OP T H E ~ ~ O P L M T ~ C and an elastic modulus in tension of more than twice that obLAMINATEB tained with %ounce drill. However, these properties &PB inCsllulass Aoetate Ethylaeliuloar phsnoiia, c r e d st the expense of impact strength which is considerably Formula Formula Formuln Formula C i d s lower then that obtained with the drill cloth. The four duck Laoft 3, hard 4.soft 8.ha.d LE weave8 fall in between the two extremes when strength qualitia Didsotrio strength (26' C.. BO 07are examined, and all of the fabrics give water absorption msults rrlsa). "olfalrnll i a l w bystep) within B fairly narrow range. Dry
wets
Powerl*clor i26'C.i Bo oy0l.a
320 27
0.299
1,Mo.MoOYEies
.Dry wet Dielectric aonatant (2S'C.)
no cycles
DW Wet I.wo.000 Dry Wet
>0.70 0.0649
0.132
360
370
0.w 0.104
0.088
210
0.0434
0.133
82
480
68
380
78
0.136
0.071 0.046
0.31 0.68
0.0443
0.0289 0.107
0,0570
0.100
0,122
ELECTRICAL PROPERTIES
The electrical properties of thermoplastic laminates (Table 111) were determined on soft and hard combinations hy standard
A.S.T.M. methob. For comparison, tests were conducted at the samo time on a Grade LE phenolic lsminate which is commonly
4.3
12.8
>30.0
8.2
6.8 18.4
13.1
13.8
>30.0
8.0
1.1
4.2 6.6
3.8
6.2
4.6
cycle.
13.1
8.6
8.1
6.8
used for ehotriosl appliostione requiring toughness. Both the thermoplastic and thermhsetting eompitions breek down rather bsdty after aosking in water. In the dry state, however, the thermoplastic laminates are good insulating materials. For &k$ric a m @ b A h odlulave scedske skmetures are zwgfily 00mp~ab1eto the Grade LE phenolic laminate while the ethyl-
962
INDUSTRIAL AND ENGINEERING CHEMISTRY
cellulose is far superior. The power factor of ethylcellulose laminates is vastly superior to the thermosetting compound; cellulose acetate is somewhat better. Both cellulose acetate and ethylcellulose laminates show a much better dielectric constant a t 60 cycles, with ethylcellulose again being the best. At l,OOO,OOO cycles the superiority of the cellulosic compositions is still notice able but by a closer margin. EFFECT O F FILLERS
Physical properties obtained on combinations of thermoplastics with fillers other than cloth are shown in Table IV. Asbestos paper produced a laminate with remarkable properties in all respects; the significant value is the elastic modulus of nearly 2,000,000. Glass fabric also produced an outstanding laminate; extraordinarily high impact strength was combined with high tensile strength and high modulus, and a relatively low water absorption value.
Vol. 37, No. 10
or the surface layer of cloth may be colored or printed with a design and given the appearance of depth by the use of a clear plastic binder. This advantage should be important in civilian applications. The proper selection of plastic composition is important for low-pressure laminates. For instance, where extreme low-temperature flexibility or good electrical properties are desired, formulations based on ethylcellulose are indicated. On the other hand, if delicate shades of color or high grease resistance is wanted, it would be advisable to choose a plastic based on cellulose acetate. I n conclusion, the addition of cloth to cellulosic plastics increases all of the strength properties of the thermoplastic compositions themselves and contributes other factors which may direct them into applications where they have not previously been used. NO-PRESSURE (SOLVENT-RELEASE) LAMINATES
A different type of laminated construction using cellulosic EFFECTOF MISCELLANEOUS FILLERS ON PROPERTIES thermoplastics as binders is based upon solvent release for obOF LAMINATES taining rigid constructions, and has been used in a limited number Asbestos Glass Kraft of applications for many years. The basic advantages of solventType of Filler Paper Fabric Paper process laminating lie in the ease of fabricating large or com30 3 25 Number of layers [rod impact strength. ft.-lb./in. plicated laminates without the need for expensive equipment or 3.7 >14.9 1.7 notch.-edrewise skilled labor. They differ from other no-pressure methods of 19400 8500 17900 Flexural strength. lb./sq., in. Tensile strenath. 1 b . h . in, 23100 25800 14500 laminating where thermosetting resins are used, in that no heat Elongation % . 1.86 2.8 2.5 Modulus df elnsticity in tendon, is required since the cellulosic plastic binders need not be cured. lb./sq. in. X 101 1890 1160 1000 The equipment and materials necessary for no-pressure conWater absorgtion (100% R.H., 168 hr., 77 F J , % 4.4 3.34 5.6 struction are simply a wood or plaster mold or a skdeton frame, a dope solution containing the plastic binder, an open-weave fabric which lends itself to easy impregnation, and brushes for applicaThe laminate based on kraft paper also has exceptional strength tion of the dope solution to the fabric. The method of applicaproperties with the exception of impact. However, this value is tion consists in cutting the fabric to the shape required to cover quite good when compared with similar. values for paper-base the mold uniformly, laying the cloth over the mold, and impregphenolics. These determinations are edgewise and would, therenating it with the plastic solution. This procedure is repeated fore, be considerably greater if flatwise specimens were tested. with successive plies of treated cloth until the desired thickness is The best results to date have been achieved with an absorbent attained. An interlayer of cellophane between the mold and the type of kraft furnished by the Forest Products Laboratory. first layer, or a coat of wax on the mold, ensures quick release The values in Table I V are based on a laminate prepared with when the laminated unit is ready to be removed. The time interthis paper. val for solvent release may range from several minutes to several OTHER CHARACTERISTICS hours, depending on the thickness of the construction. One of the most important qualities of thermoplastic laminates with cloth fillers is their adaptability to easy and economical TABLE V. PHYSICAL PBOPERTIES OF NO-PRESSURE THERMOfabrication. They can be formed and drawn into complex shapes PLASTIC LAMINATES when heated, and component parts may be sealed together quickly A.S.T.M. by solvents or heat (Figure 1). I n examining the drawability of Value Test No. various fabric fillers, a drawing mandrel was used. The apparatus Izod impact strength. ft.-lb./in. notch, edge6-10 D63842T D25641T consists of a base block to which is attached a ram 2 inches in wise 1300-4400 Tensile strength Ib./sq. in. diameter. This ram is guided into an opening 26/~einches in 1500@--170000 D638-42T Modulus of elasjicity in tension, lb./sQ. in. 8-15 D638-42T Elongation, % diameter in the upper block. The clamping ring, supported by Water 24-hr. absorption, immersions 9% 6-12 0570-42 springs, serves to prevent wrinkling around the edge of the draw; 10-12 ..... 100% R.H.,168 hr., 77O F. pressure is adjusted to the extent that the hot laminate can a Edges were sealed. slide over the ring and into the mold to increase the ultimate depth of draw. The procedure followed in evaluating the draw characteristics of different fabrics waq to heat 6 X 6 inch laminates, In examining no-pressure thermoplastic laminates, the pribased on these fillers, to a predetermined temperature and to mary consideration has been the study of physical properties transfer them immediately from the heating unit to the drawing obtained when various binders such as ethylcellulose, cellulose device. The pressure for drawing was supplied by a small hydrauacetate, and cellulose nitrate were employed. The over-all relia press. Motion was stopped when the first sign of rupture sults are described in Table v. Specimens for these tests caroccurred, as seen from the opening in the upper block. The ried varying percentages of cloth ranging from 30 to 64%. The value for drawability was taken M that depth to which the unit amounts of plasticizer in the Cellulose bonding agents also were was drawn to cause rupture. adjusted from 15 to 30% of the plastic composition. Osnaburg The deepest draws were obtained with a high-twist open-weave cloth with a weight of 8 ounces per yard was used throughout. fabric such as bootleg duck or with very high-tensile-strength I n general, as the ratio of cloth to plastic is increased, the fabrics such as balloon cloth. The drawing characteristics of impact strength, per cent elongation, and water absorption values laminates bonded with either a cellulose acetate or an ethylcellushow corresponding increases while the values for tensile strength lose plastic were comparable. and elastic moduli go down. The range found most workable There is practically no limit to color versatility with thermofor the wet method of lamination was 4545% fabric content: plastic laminates. The plastic binder may be dyed or pigmented, T ~ L IV. E
~
October, 1945
INDUSTRIAL A N D ENGINEERING CHEMISTRY
the optimum solids content of the dope solutions from the standpoint of application by brush was 1045%. Cellulose acetate binders were dissolved in a 90:10 mixture of acetone and ethanol; the solvent for ethylcellulose WM ethyl acetate. The effect of increasing the percentage of plasticizer in the bonding plastic was t o lower water absorption and elastic moduli while the percentage elongation increases. Impact and tensile strength values showed no consistent trend and were quite comparable. Water absorption values were determined both by direct immersion in water for 24 hours in accordance with AB.T.M.recommended procedure and by exposure to 100% relative humidity a t 77" F. for 168 hours. It was considered satisfactory to immerse the no-pressure solvent-release specimens, since the plastic binders had relatively low amounts of plasticizer and therefore plasticizer loss would not be significant. I n comparing the resulting products of two methods of lamination with cellulosic binders-that is, the low-pressure type which
963
was first discussed and this solvent-release build-up type-the following facts are apparent: The pressure typeA with comperable plastic formulations have greater tensile and flexural strengths and much higher moduli; they also have lower water absorption. On the other hand, solvent-release no-pressure constructions are superior in impact strength and are lighter in weight, having densities around 0.8 to 1.0 aa against 1.15 for ethylcellulose lowpressure structures and 1.3 for cellulose acetate low-pressure types. The method of producing low-pressure laminates, involving mechanical cloth treating and pressing equipment, indicates that they lend themselves to volume production. On the other hand, no-pressure, solvent-release laminations do not lend themselves so c a d y to mechanization but do have the advantages of requiring simpler and lower-cost equipment for fabrication, and can be built up into larger units. PEIwaNTaD before the Division of Paint, Varniah, and Plsstias Chamiatry at the 108th Meeting of the AYIORICAN CHEMICAL SOCIETY in New York, N. Y
Aconitic Acid from Citric Acid by Catalytic Dehydration J
. I
J
ROBERT R. WMBDENSTOCK
PAUL F. BRUINS
Charles P3ser & Company, Znc., Brooklyn 6, N . Y .
Polytechnic Inatitute of Brooklyn, N . Y.
A
,
CONITIC acid is the common name of propene-l,2,3-tricarboxylic acid. This unsaturated acid occurs in nature in Aconitum napellus, Equiwtum flzrviatik, sugar cane, beet root ( l a ) , and sorghum (11). Aconitic acid has been suggested as a possible ingredient of modified alkyd resins. Esters of aconitic and tricarballylic (dihydroaconitic) acids have been employed aa plasticiaers for plastics and Buna-type synthetic rubbers, and in the manufacture of wetting agents (IO). Aconitic acid can be readily decarboxylated to itaconic acid (7)) whose esters can be polymerized to produce plastics (2). Previously reported methods for aconitic acid include recovery from sugar cane sirup residues (8),dehydration of citric acid by sulfuric acid (3, 18), hydrochloric and hydrobromic acids (6')) and H.PO4 (S),and direct synthesis from sodium malonic and acetylene dicarboxyIic esters (9). Because citric acid is relatively cheap and, in normal times, is abundant, most aconitic acid produced in the past has been derived from citric acid by the mineral acid dehydration process. Rather low yields (4144% with sulfuric acid, 3) and extremely corrosive conditions in the mineral acid processes were responsible for the present investigation. Three general methods were employed: (a) homogeneous aystems in which the citric acid and catalyst were dissolved in a suitable organic soIvent; (b) heterogeneous systems in which the citric acid was suspended and the catalyst suspended or dissolved in an organic liquid; and (c) a vacuum fusion process in which the citric acid and catalyst in intimate mixture were heated under 23-24 mm. mercury absolute pressure; in thie process fusion occurred at the temperatures employed which were always below the normal melting point of citric acid (153' C,). Because of the relative instability of aconitic acid above 150" C., the work was limited to the range between 150" and 120' C. (the minimum t e m p e r a t d a t which citric acid could be dehydrated even under the influence of a catalyst). Previous investigation (Y) showed that the pyrolysis of citric acid can proceed in two directions; both liberate water but only one produces aconitic acid:
H-CH-COOH HO--b--COoH H-Ck-COOH
CH2-COOH
AH2--COOH
kZCOOH
Aconitic,Acid
Acetone Dicarboxylic Acid
-1 &-COOS
+
&€€-COOH Citraconic and Mesaconic Acid (cis-trans Isomers)
-2c0,
6-COOH
CH:
AHAOOH Itaconic Acid
t"* Acetone
The decomposition products of aconitic acid are the isomeric acids itaconic (methylene succinic), and citraconic (methyl maleic) and mesaconic (methyl fumaric). Acetone and carbon dioxide are the decomposition products of acetone dicarboxylic acid. DEHYDRATXON PROCEDURE
APPABATUS.The reaction apparatus consisted of a threeneck flask with the center neck containing a motor-driven agitator and one side neck containing a Dean-Stark tube and total condenser. In the case of runs made a t atmospheric pressure the water liberated during the course of the reaction was measured by collection in the Dean-Stark tube. I n vacuum fusion ex-