Adsorption - Industrial & Engineering Chemistry (ACS Publications)

---

i

'

1 UNIT OPERATIONS REVIEW

C H E M I C A L ENGINEERING REVIEWS

I I

II Adsorption I

THE

field of adsorption continued to grow both in breadth and depth over the past year. Some 800 contributions were reported, well above the capabilities of reporting in any such revieiv as this. T h e author has Xvinnowed these to some 300 so as to hit most of the highlights, with, it is hoped, as feiv significant omissions as possible. The greatest increase, and consequently the field most sharply cut, \vas the application of adsorptive, and particularly liquid-phase chromatographic procedures, to the analysis of trace components, and to analytical resolution of laboratory samples in lieu of standard analytical procedures. A great many Fvorkers are engaged in building u p chromatographic data for resolution of specific mixtures. These techniques belong more properly to a revie\v on analysis in some cases, and to the area of ion exchange in others. .Although i r will be a stupendous task, this field seems ripe for compilation of such analytical procedures into a handbook, and this reviewer expects to see some such effort vithin the nest several years. Thd most explosively developing area is that of gas-liquid partition chroinatography. This area is ne\v and fills a great need for techniques of analysis and detection applicable with high degree of confidence by personnel Icith little background. Relatively little effort has been expended in this field in the area of adsorption vis-a-ois partition, although both subjects are being developed concurrently. O n e additional spur to the development of the techniques is that this subject holds forth promise of giving on-the-spot answers to plant operators by automatic and semiautomatic use, and perhaps within the foreseeablr fu-

460

ture, it may have applications to automation as well. A field which seems to be neglected, whereas it should be exploited, is the application of adsorptive techniques IU large-scale rectification. While a fair amount of background exists, and some processrs havr been developed 10 r h r point of patenting and a small degrer of exploitation, a potential exists here which is virtually untapped. I n particular, marriage of specific adsorbents to the stripping operation, for separation of difficult mixtures, racemates, and for general use for more volatile components has been coinpletely ignored. 11 is the opinion of this reviewer that, while tlirrc. is a great deal of scientific interest in adsorption \vorld \\+de, very lirtle engineering application in gmeral has been made, and the time for such application is ripe. Industrial Applications Details of the fluid char adsorption process verbally reported on last yrar have nmv been published by Etherington and associates (2.4). The procrss has been carried through pilot plant phase, and consists of a bubble cap column

through which flows a fluidized charcoal countercurrcnr to the rising gas strram. 'The opcration is in many resprcts analogous 10 a rrhoiled oil absorber. Commercial dt:signs and applications were discussed. Xccording to Etherington, no serious deactivation problrrns were obtained when xvorking with a typical rrlincry residur gas. This \vas thought to be due to the small particle sizr of the char. Because of the early stage of rrsults, howevrr, they do riot recominend iinmrdialc application to other than "clean" gas ( 1.4). (:orrelations w ~ also c given (3.4) for cquilihrium adsorption of pure hydrocarbons and their mixtures. From thew data predictions Xvere made stated to be good up to 250' I . and 2.50 pounds pt*r squarr inch gage pressure. 'The correlation of adsorption cif unsatui,ated gases O I I charcoal rrported by Regak and Smil-nov last !.c.ar h a s b r r n translated into English by Consultants Bureau l/jLl). '1'hc)- rcportrd for flow systrms coiirlatioii of the f'oirn h'u = 0.21 X 10 - - 5 RrO.R l+ -1F, -0.04 [,'-a. 3 10,:). \vhcrc the dinirnsionless paraincters have thcir usual Inraninps. Scha!, and Szrkely (8.4) havr evaluated adsorption in ;i fixed bed, by a inethod bascd on priii-

8. L.' HARRIS was born in Savannah, Go., in 1917. He received his B.E. (1938) and Ph.D. (1941) in Chemical Engineering at the Johns Hopkins University. After wartime service with the Chemical Corps, he tought at Johns Hopkins and consulted until 1953, when he returned to the Chemical Corps, now being employed b y the Chemical Warfare Laboratories at Army Chemical Center, Md. Harris is a registered professional engineer and a member o f AAAS, ACS, AIChE, Sigma Xi, and Phi Lambda Upsilon.

INDUSTRIAL AND ENGINEERING CHEMISTRY

ciples of frontal chromatographic analysis. Haul has discussed the use of gas adsorption for separation (5.4). The dynamics of adsorption in a fluidized adsorbent were compared with those in a fixed bed by Romankov and Lepilin (74. Breakthrough time was found to be a linear function of packing height instead of the usual correction factor ordinarily assumed as a “critical depth” which does not contribute to the life. The volume of vapor adsorbed per unit volume of adsorbent was independent of packing depth, gas rates of three to four times those for static beds were achieved, and only a slight temperature rise was observed during adsorption. Vagin and ’ZhukhovitskiI (9.4) investigated the theory of thermal separation of gaseous mixtures adsorbed in a bed and desorbed a t constant pressure by a moving temperature gradient. For complete separation they reported that only the isobars of the pure components need be known. Vyakhirev and others (70A, 7 7.4) have developed an analysis system for gas mixtures by gas-phase chromatography based on selective desorption with a moving temperature zone. Gas chromatography, discussed more fully below, has been applied to process control by Dudenbostel and Priestley ( 7 4 . A 20-foot alumina-packed column was adapted for intermittent continuous-duty determination of propane in propylene in 30 minutes or less. The sample is trapped between solenoid valves and sent to the column, eluted with air, and sent to an analyzer which is a modification of the MSA water analyzer. The “wet” stream and “dry” stream are sent to adsorbers in parallel, the streams being alternated frequently. The wet stream gives up adsorbate and is thereby heated; the dry stream takes up adsorbate and is cooled. The temperature difference is recorded by thermopile and the results are presented to the process operator on a bar graph, with the height of the bar equal to concentration.

Surface Area and Porosity Studies

As would be expected, the volume of work specifically oriented toward surface area determination was small. Tempkin has prepared tables for calculating the volume of the unimolecular film from the adsorption of nitrogen a t various pressures (238). Royen, Orth, and Ruths (2OB)found that the pressure a t which the monolayer of nitrogen was completed on various adsorbents was constant for the same material even though the area was altered by progressive sintering of the surface. Furt h e r m o r e , Brunauer-Emmett-Teller (B.E.T.) areas agreed with areas de-

termined from average particle size measured by x-ray diffraction. The previously reported phenomenon of a negative intercept for the B.E.T. plot on very pure sodium chloride has been explained by MacIver and Emmett as being due to departure from linearity a t low relative pressures on homogeneous surfaces. Below 0.1 relative pressure, linear B.E.T. plots of reasonable surface area were obtained

(77B). Kel’tsev and Khalif proposed technical grade propane at 20’ C.for area determination (72B). When compared to nitrogen areas on various silica gels, values differing only 10% were obtained, using a molecular area for propane of 44.2 sq.A. Maron and others suggested the use of soap titration for determination of the surface areas of polymers. Soaps of various acids were used and agreement with nitrogen area for very coarse carbon blacks was obtained. For finer blacks the soap area was much the smaller, but they suggested that the soap area was the preferable value for studies on rubber reinforcement (78B). Data on the effect of adsorbates and solvents on the area determined by stearic acid adsorption were also presented by Suito and coworkers ( 2 7 B ) . Danes (7B) has modified adsorption apparatus for rapid and precise measurement of adsorption a t low temperatures. He recommended use of neon for dead space determination. DeMarcus, Hopper, and Allen (8B) refined their previously proposed surface area method by the use of a more realistic interaction potential function near the surface, and by inclusion of quantum effects. Parts ( 7 9 B ) compared the Langmuir, B.E.T., and Huttig equations with that of Gregg and stated that the latter is not an independent method of area determination. Craig, Van Voorhis, and Bartell found the free energy of immersion of graphite in nonaqueous liquids to be independent of area and pressure of compressing the graphite into porous plugs; it was dependent only upon the nature of the surface (4B). Graham measured surface areas of Graphon using dye molecules and found that nominal pores of less than 13 A. in diameter were inaccessible to the dye. Acidic substituents on the surface also hindered adsorption of anionic dye in proportion to the concentration of the substituent in the surface. Culver and Heath (5B) prepared gas adsorbent carbons from Saran ranging in surface area up to 3100 sq. meters per gram having pores from 7 to 8 A. to 15 to 20 A. in radius. Activation increased pore size somewhat, as well as the number of pores. Dacey and Thomas (6B)found it possible to modify saran charcoals by pyrolysis of vinylidene chloride adsorbed

thereon. Avgul and others found hysteresis on nonporous and large-pore carbon samples (7B). Barrer and Strachan (2B) found no such hysteresis on nonporous carbon, but a pronounced hysteresis appeared on compressed plugs, which disappeared as the compression was increased. The plugs were constrained mechanically to constant total volume during adsorption. r, Kiselev and coworkers preparea a series of carbons ranging from nonporous, through large-pore to small-pore samples and tested theni for adsorption of phenol from water. The adsorption below relative concentration of 0.1 was similar for all but the fine-pore sample; above this value the coarse-pore sample was more adsorptive than the mediumpore or nonporous sample. The finepore sample was more adsorptive a t low relative concentration, and adsorbed little more thereafter ( 7 5 B ) . Similar results were found by the same workers on various other adsorbents (73B, 74B). Tsitsishvili and Barnabishvili (24B) found evidence of bottle-shaped pores in bentonite which were opened up as activation progressed. Voigt and Tomlinson (25B) have derived a general expression for relative adsorbate volume as a function of the poresize distribution, with specific functions for the cylindrical pore and the “inkwell” type of pore. Using these, in conjunction with experimental isotherms showing hysteresis, calculation of pore size, distribution, and surface area is possible. ’ Teichmann (22B) has patented a method of varying pore size in permeable barriers using an adsorbable liquid phase which can be frozen in position after predetermined partial saturation. I t has been applied to modification of sandstone or sintered barriers using Wood’s metal. Cosgrove (3B) has measured the porosity of anodic oxide coatings on aluminum and found disagreement between krypton values at -195.8’ C. and butane values a t 0’ C. for large pores. He recommended use of solid krypton vapor pressure data to better the agreement. Folman and Shereshefsky ( 7 0 B ) measured vaporpressure lowering for toluene and isopropyl alcohol in 0.8 to 3.0 micron capillaries and found disagreement with the Kelvin equation which they attributed to long-range forces. Everett and Whitton ( 9 B ) stated that the sharp drop in the desorption isotherm and the closing of the hysteresis loop may not give reliable information about the most common pore size.

liquid-Phase Chromatography Some hundreds of articles appeared during the past year on the subject of

VOL. 49, NO. 8 , PART II

MARCH 1957

461

While there is much world-wide scientific interest in adsorption, very little engineering application has been made, and the time is ripe.

chromatography. The bulk of these de7:t ivith the use of chroniatography as an analytical tool in laborator>-analysis. in forensic medicine, and in control of processes, and \cere concerned largely Lvith paper chromatography. S o attempt is made here to review these articles exhaustively; instead some of the other and more general aspects of the subject are discussed. Revieivs specifically limited to chromatography have appeared in numerous ,journals. Development and theory \vere revielved in Russian ( , 5 K ) and French (33C). methods in German (32C). nc\v progress in Italian (JC). and large niolecules in German (37C'). Paper chromatography has been rather conipletely revielved i n German ( 9 C ) ; shorter reviews appeared in English (17C, 5 7 C ) , Spanish (6C): and German (39C), and a review of electrical techniques appeared in Japanese (35C). Potterat ( I Z C ) and Sulser (53c')have extended the previously reported ivork of Matthais on neiv radial paper techniques, permitting the application of large fiiter paper sheets and making this method suitable for serial it-ork. Precise limired bands ivere obtained: allowing resolution of some difficuli material. M'arren and Fink (582)described the use of a fed solvent vapor atmosphere to give rapid separations of 2 to 20 minutes for quantitative, and 2 to 4 minures for qualitative results. .Alcock and Canne11 ( 7 C ) studied the temperature effects of sugar solutions and related rhein to solubility. Sominer studied the effect of salts, acids, alcohols, and paper on inorganic cations (5OC) ; Gorbach and Deinme1 (27C) developed a quantitative method for "test-tube" work on amino acids. Ganguli (77C: 7SC) studied the effect of types of cut on R, values and proposed an optimal system. Prijs and Erlenmeyer (13C) developed retention paper chromatograms of bases {vith hydrochloric acid and thus evaluated very small quantities. Giri described a simple system for separating large quantities of material. up to 200 mg. on a single sheet of paper; large quantities were handled on multiple sheets held i n an appropriate apparatus (79C). Ganguli (76C) investigated a number of compounds on both unidimensional and circular papers. He found that the square of the circular R, values was equal to the linear R , values. Jayme and Knolle (23C) found that glass-fiber papers did not behave

462

as crllulose papers unless impregnated. Proper impregnation gave advantages i n some cases, although some sugars could not be rrsolved. Schauer and Btilirsch (J7C) developed a modification of' the hlarrin eqiiation by ivhich analysis of structural details of a substance can be made if sufficient values of chromarographic constants arc measured. They have prepared tables of constants for numbers of atonis. functional groups, and structural modifications. Pollard and others ( I 7 C ) have investigated the multiplt~-spot.phenoinenon and stated causes, whether competitive rflircts. complex formation. or impurities in the paper. Pickerins and Jacobs ( K C ) discussed more fully the role of complex formation. Kertes ( 2 i C ) found the Ri values of se with the water content of Lvater-miscible organic solvent so as to be linear Lvith the dipole moment of the mixed solvrnt for univalent and bivalent cations, as \veli as for some organic materials. Franc and Jokl (7iC) developed a general equation for Rj values of aromatic isomers if certain basic data are known. Precipitation of inorganic ions in gels \vas investigated by lliione and corvorkers ( S I C ' ) and proposed to explain Liesegang phenomena. Krebs and others (2ciC) analyzed their previous data and set up six rules governing the resolution of racematrs on starch columns. Uchida and Ogino modified the Amundson general equation and applied the result to separation of propanol from xvater in a charcoal column (38C, , 5 K , 57C), finding agreement tvith a proposed rr.eclianisin of solid diffusion through a film around the absorbed particles. Lervis (ZSC) pointed out the analogy of laboratory chromatography to stepwise countercurrent liquid-solid extraction, \vhich latter he proposed for pilot plant evaluation. If a continuous process is desired, he proposed an analog of a liquid-liquid extraction column, wherein the eluent passes u p a column very slowly and the adsorbent passes down, the feed stock entering midwray between the points of entry of the two. .A number of contributions were made to the theory of chromatography. Klinkenberg (25C) pointed out the resemblance of the results of Giddings and Eyring to a theory based on transfer coefficients. Bastian and Lapidus (3C) have developed equations for longitudi-

INDUSTRIAL AND ENGINEERING CHEMISTRY

nal cliflusion olwariiig i n coliirniis of finite Irngth undri. rquili tiriuni conditiOl1S. n1.aice has dr gradienr elution an grnpliicai nierhod oi' drtrr.iiiin;ttion of the soi.ptioii of' ~ldsorb;iir for ccirvrd isotherins thcwx)f', Practic,ai advantays of gradient r.lution analysis i w r t ' pi~intecl o i i t for both linrar a n d c:sIioiiciitial infi ut'n t gradient. 'The kinetics of soivc.nr flo\v in paprr chroinatograpliy \vas tlisc.ussc.ci b y Ilrclos and Vavruch (/JC)> \vho utiliztd a simple model of dn rfii,crivc c ~ i ~ ~ i l l a r > ~ . Glueckauf (3//c')tcxeci \-arious cinliiric,al dilfision cquations as app1ic.d t o pi~aciical chroinatoqaphic brrak-iiiruuqli ciirvcs. and found a simple rquation [ I ] si\(, s u ifici e n t rigor. :'icId i r io 11a1 ma thrma Iital trt'atnient n~ p r t J s r l i r c d IJV l \ l ; i r h l I t , 137C) and predictions ol' posiilions 01 \.arious adsorbates and conditions undcr ivhich separation facton c.ouid br v'iritd, \vas made thereby. Breslrr and Sainsonov have made a suniinar!- rrviciv of thcorrtical chromarogi-aph! ihrough 1954 (5C). A ntiinbcr of n a v tcchiiiqucs and appararus ivrre dcscribed in tlie liieraturr. Thompson and blarion 7X') s u q q r s i tlir use of borosilic,a[r glass instead of steel as a trough for descending lzipc'r chromatography. C:ontrol of elution \vas discussed by s capillar>. dcvice for rluting from cut-out spots of a paper chroniaro~rani( / / C ) , improved autornatic fred for gradicnt se~niautoinatic( /iC;j and ciution (-/X'), automatic ( / I C ) apparatus for incrrasing solvent polarity, use of salr solution instead of \vatrr IO elute from ion tsxchanqc resins (&C), and the usual automatic fraction collectors i7C, 30Ct > K ) . Llunz (36C) has traced thr. c a u m o f uncvrn running of papar chroinatograms to local heating and evaporation of solvent. Colas described preparation ~f ii silica useful for difficult isomeric separation (SC). Continuous chromarographic apparatus !vas developed by Solins (49C) and by Svensson and coivorkers (54C). Rankin (3SE) proposed that sorption tubes be fillrd half full of solvent and that the packing material be added slurried in the solvent. S o pressing of the packing is then required. hIethods of evaluation of chrornatograms lvere the primary concern of several authors. Eger (72C) drscribed an apparatus for measurement of adsorption spectra on paper chromato-

ADSORPTION grams, involving a n attachment for a standard spectrometer. Fellegi and Slama described a simple photometer (74C) and Langer (28C) described the use of direct polarographic scanning. Houston (22C) developed a proportional divider easily improved for rapid determination of the R, values of paper chromatograms. Barrollier suggested coating such chromatograms with a polyurethane resin which dries to a transparent film, for photometry (2C).

Gas-Phase Chromatography Gas-phase chromatography divides logically into two divisions, adsorptive and partition chromatography. Very little activity was reported for the former; the growth of the latter over the past few years has been explosive. Some developments apply equally well to both applications, especially auxiliary apparatus. Chromathermography is one area of interest which is preponderantly adsorp'tive. Turkel'taub and his coworkers have been active in this area. They have found that longitudinal diffusion causes diffuse band formation (60C) and recommended use of thin tubes and fine adsorbent as corrective measures. In the absence of longitudinal diffusion, an equation for nonequilibrium was derived (330). Apparatus and techniques were also discussed

(320). The greatest activity in gas-liquid partition chromatography has centered about resolution of gaseous hydrocarbon mixtures, particularly in the range C I to CS, although apparatus offered for sale includes ranges of resolution for materials boiling up to 250' C. Resolution of C1 to Cg mixtures was reported by Fredericks (700) and Taramasso (370),of isomeric hexanes and hexenes by Sullivan and others (300), of Cg and CO mixtures by Eggertsen and coworkers (90),of cg to (28 mixtures by Lichtenfels and others ( 2 7 0 , 220), and a general study was reported by Brooks and Collins (30). A known mixture of 35 Cg to c8 aliphatic hydrocarbons was resolved on various substrates by James and Martin (770). In the more general area, Greene and others resolved air, carbon dioxide, carbon monoxide, hydrogen, methane, ethylene, and acetylene (720). They increased the temperature of the column continuously and increased the upstream pressure to maint'ain constant flow to the detector. Podbielniak and Preston presented a tabulation of approximate retention times for alcohols, ketones, esters, fatty acids, liquid hydrocarbons, and nonhydrocarbon and light hydrocarbon gases on various substrates (250). Kamer and coworkers studied the series

of volatile fatty acids from formic to dodecanoic acid (780) and James separated and estimated quantitatively the volatile aromatic amines, using a paraffin hydrocarbon stationary phase (760). General discussions were presented by several authors (20, 60), as well as a review by Green ( 7 7 0 ) . Porter, Deal, and Stross (270) have determined partition coefficients from gas-liquid partition chromatography, and have developed a n expression for retention volume which gives consistent results a t low solute chargings. The partition Coefficients obtained therefrom agreed with those measured directly. Pierotti and others have determined solvent effects and developed a general relation (240). Hoare and Purnell found the peak retention volume to be logarithmically related to vapor pressure a t the elution temperature (750). For related compounds a single curve was obtained, and temperature dependence of retention volume was used to determine solubility which agreed with published data. The equations found were also derived experimentally (74'0). Elution time was found by Wiebe (340) to vary with column length, inversely with carrier gas flow rate, and logarithmically with the reciprocal of absolute temperature. Resolving power improved with decrease in column length, passed through a maximum with flow rate, and varied nonlinearly with column length, approaching a maximum asymptotically with time and dilution with carrier gas. A number of the studies reported above concerned varying substrates. Neimark investigated silica gels of varying preparation (230) and Keulemans, Kwantes, and Zaal (790) suggested various substrates for resolution of hydrocarbons, aliphatic from aliphatic, aliphatic from aromatic, and aromatic from aromatic, as well as oxygenated compounds. The applications of gas-phase partition chromatography are varied. It has not yet been applied to automatic process control, although it is understood that a great deal of effort is being expended in thjk direction. Applications so far reported included preparation and collection of pure materials (70, 260), mass spectrometer analysis (80), and study of kinetics (40)and of catalytic reactions (200). Rock (280) has discussed the relation between gas partition chromatography and extractive distillation, and has used the former to predict the liquid to be used as entrainer in the latter. Numerous developments in apparatus and technique continue to be reported. Helium has been stated to give greater sensitivity as a carrier gas than nitrogen, but a loss of resolving power (350).

The apparatus requirements have been discussed generally by Dimbat, Porter, and Stross (70). Scott (290)and Henderson and Knox (730) discussed the microflame detector. Recently a gasdensity balance has been reported which operates entirely on a reference gas and so does not crack the sample; quantitative recovery of the resolved components is achieved. The reading is proportional to the molecular weight of the unknown and the apparatus has resolved peaks, thought to be the same by observation of other apparatus (50). Apparatus currently on the market include the Burrell Kromo-Tog and Fracton, the Perkin-Elmer Vapor Fractometer, the Podbielniak Chromacon, Research Equipment Corp.3 Distillograph, the Hallikainen-Shell Chromagraph, the Fisher-Gulf Partitioner, the Beckman Gas Chromagraph, and the Consolidated Electrodynamics Gas Chromatograph.

liquid-Phase Adsorption This subject may be further subdivided into two sections, adsorption on solids and on liquid substrates. The former accounted for the bulk of the published work. The mechanism of the sorption by active carbons was elucidated by Okamoto (35E),who found that the primary concern was the pore size and distribution rather than surface properties. Korenman (22E) showed that in coprecipitation the micro component could either be adsorbed or be isomorphous with the macro component. In the former case a typical isotherm results; in the latter a linear plot obtains. Koral reported studies of poly(viny1 acetate) on solids and found that chemical variation and molecular weight were more important than other properties of the polymer (27E). hloilliet reviewed sorption of surface-active agents

(30E). Graham and Hansen (72E)found that carbon sorbs aliphatic alcohols and carboxylic acids flat on the surface a t low coverage; at higher coverage only part of the carbon chain is bound directly to the surface. Morrison and Miller used similar adsorption studies to determine pore size of active carbons (3723). They found that the pore diameters must be corrected by 2 A. to bring them into agreement with values determined by water. Other studies of interest included the sorption of acids by clays (47E), of fatty acid on mechanically activated metals (43E): the use of fly ash as sorbent for waste phenol liquors (39E),and sorption as a means of elucidating the aging of sulfides (40E),and the interaction of oxides (33E). Use of radioactive tracers

VOL. 49, NO. 3, PART II

8

MARCH 1957

463

UNIT OPERATIONS REVIEW was reported in two studies of organic acids (79E, 50E). The specificity of silica gels prepared in the presence of organic dyes previously reported by Dickey has now been investigated by him more rigorously and for a large variety of organic compounds (7E). H e found that the effect \vas not observed for compounds not adsorbed readily by usual silica gels, and that low pH, low electrclyte composition, and low temperature favor the development of specificity and its resistance to aging. Haldeman and Emmett confirmed Dickey's early svork and reported the slow loss of specificity upon aging a t room temperature ( 7 2 9 , Bartell has determined that the extinction coefficient for a d>-ein a film of solution of molecular thickness is not significantly different from its value in bulk solutions (7E). hIattuck questioned the use of the Hartman method of such determinations of experimental and theoretical grounds and concluded that the method is too insensitive to bc of practical use (ZGE). He further developed an experimental system utilizing two barium stearate steps on chromium. which was theoretically evaluated (27E) and applied to studies of plasma albumin, in collaboration \rith other workers (ZE). Adsorption on liquid substrates was reported briefly. Grand (73E) studied the adsorption a t the surface of polarized mercury and Kryukova ( W E ) showed that this effect caused errors in polarographic curves over those in the absence of adsorption. Pilpel measured adsorption a t the benzene-ivater interface and proposed a system of identification of unknown molecules based upon the observed, as compared M.ith the calculated, molecular areas (37E). Meguro (28E) showed that coagulation of dyes takes place when a small concentration of surfactant is present and dispersion takes place when the concentration increases. Zutrauen and hfinassian-Saraga (57E) showed that an integrated form of the Gibbs equation holds for some long-chain compounds, but not for others. Cutting and Jones (5E, 76%) measured the adsorption of insoluble vapors on water surfaces. Sobotka (45E) measured the penetrability of molecular layers and postulated a mechanism of continuous opening and closing of the molecular array in the course of thermal motion. Lahfer and Blank (24E) showed a method of separation of mixed molecular layers by flow through a surface channel from a region of higher to lower surface pressure. Frisch and Simha proposed a model for monolayers composed of linear macromolecules a t a n oil-water interface (70E). Results indicated that rough indications of the shape and mean

464

dimensions of the macromolecules could be obtained from film studies a t both air-water and oil-water interfaces. The Guastallas have extended the DuClauTraube rule to adsorption from concentrated solutions ( 7 3 E ) . Schafer explained surface properties on the basis of the difference betrveen thermodynamic properties of the surface molecules and nonsurface molecules (-17E). Tunitskii and others have developed a d>-namic theor!- of adsorption and chromatograph!-, now extended to ion exchange adsorption in interdiffusional kinetics (48E). Calvet found the thermokinetics of adsorption of acetone on various organic materials to be indicative of hydrogen bonding (4E). Heats of immersion and the calculation of heats of adsorption therefrom were reported by several authors (8E, IO'E, .ME). Kiselev has proposed an acid-base mechanism for adsorption on silica gel, in contrast Jvith physical theories (20E). Mill (-39E) has taken issue \vith Elton on the latter's revision of his \vork on the adsorption of binary liquid mixtures on solid adsorbents. More practical studies of the liquid phase adsorptive effects were varied. Deryagin described a friction gage for measurement of adsorption of surface active molecules ( 6 E ) . \Vith Izmailova ( 7 7 E ) he studied the effect of adsorption on friction. Levine and Zisman (25E) studied the effect on lubrication of adsorbed close-packed monolayers of various organic and fluorinated organic materials, \vhich compared favorably \\.ith hydrocarbon lubricants. Fraioli and others have found that friction decreases irith increasing. adsorption of organic boundary films up to a monolayer ( 2 I G ) , Suzuya ('-16E)has revieived the errors due to adsorption on analytical xvare in microanalysis. Patrick and colvorkers studied the effect of oriented monomolecular Ia!.ers on adhesion (36E). Shepard and Ryan (32E) used radiotracer technique in the study of adhesion, Ohrn [3JE) studied the effect of adsorption on capillary dinensions of viscometers, and ascribed thereto the effect of higher viscosity in smaller capillaries when measuring polymer solutions a t low concentration. Nathan (3ZE) found that strongly adsorbed amines and organic acids inhibited attack of iron powder by acetic acid more than did weakly adsorbed substances. For hydrochloric acid attack no inhibition was observed. Smith and Leva have discussed the basic concepts of drying organic liquids with solid adsorbents. They also discussed regenerative dryer designs (44E). Bergstrom (3E) has developed a system of drying of adsorbent solids after liquid treatment and before regeneration.

INDUSTRIAL AND ENGINEERING CHEMISTRY

Chemisorption The bulk of work on this subject is reported as catalysis and all of chemisorption applies to ihat area. A few studies lvere reported strictly as chemisorption, particularly on tungsten and on nickel. Ehrlich (O'I;. 71;) investigated the adsorption of nitrogen on tungsten a t IOU pressure and various temperatures. He identified three different adsorption states a t lo\\ temperature and desorption i n t u i i disrinct temperature r'angcs. Goinrr a n d \Vortman studied the Ji-.obility of 11) drogen on bPI(J\Y 20' F. tungsten a t ti~ii~pcratiir~s by field emission techniqiic-. arid found it to be mobilr. i)i~>vidingthat a source of activation encr,qy (sticli as incident light) \cas aLailable (XI;), In other studies of surfacc dify~isionoi' h!-drogen and oxygen on tungsten ( / O F ) ; they found t\co types of surface Iit.ti=rogeneity to be irnportant---cr.~sialloirrat,hic anisotrop!- and esistciicr 111' iiihrrenrly different adsorption siws riri a given face. Xfoore and ,Allison studicd the behavior of strontium and barium on e tungstcn by the same t ~ ~ h n i q i iand concluded that despite tht, Iiiqh activation energies, tlit: adsorption \vas essentiall!, physical ( 73F). h r h a t y i and Shuppe (77E) fwncl that sodium, potassium. and rriagnesium migrated to crystal boundaries on the t u n g s t m surface and desorbed difrertmtly from the various c r p a l faces. Kwwi (70F) and Kaal [ / ? I ; ) studied chemisorption on iron catal>-srs. Raal stated that his results contiadicted previous reporis of t h e chen:isorption of carbon monoxide as occurring on the free iron surface. Sastri and \*isu.anathan ( I C i F ) found that presorlition of carbon monoxidr on cobalt FischrrTropsch catalyst causes cnhaiicerncnt of h;.drogen; the reverse is trur for high values of h!,drogen. not so for lo\\. hydrogen adsorption. 'This is inrcrpreted as being a function of tht, arrivr sites; for more active sites carbon monoxide increased hydrogen; thr r true. For less active sites, mutual enhancement occurs with complex formation. Smith and coworkers reported that the carbon-oxygen surface complexes 011 carbon surfaces are cleaned off' by hydrogen; no hJ-drogen complexes are formed (78F). Ll-ubovskaya and Ravdel found that as getters titanium is preferable to zirconium for hydrogen; the reverse is true for oxygen ( 7 7 F ) . Huttig related chemical resistance and microhardness to heat of sorption (SF, -17G). Davtyan and Ovchinnikova (5F) studied the chemisorption and oxidation of sulfur dioxide on solid catalysts and proposed a hypothesis based on the Balandin multiplet hypothesis and the Eyring complex-formation theory.

ADSORPTION Balandin and others derived equations for determining the equilibrium constants of adsorptive displacement of one set of substances by another from a catalytic surface, expressed in terms of dimensionless variables (4F). Balandin has proposed a new “second law of dehydrogenation catalysis,” that the catalytic excess stress (excess of free energy) is proportional to the adsorption excess stress with a reverse sign. His previously published first law stated that the overstress depends only on the unsaturation of active centers, not on the nature of the adsorbate ( 3 F ) . Molinari pointed out the relation between the constants of the Arrhenius equation for catalysis and its significance for desorption, explaining thereby the accelerating effect of a n adsorbing gas on the desorption of a n already adsorbed species ( 7 2 F ) . Fal’kovskii and L’vov derived an equation to explain the effect of isothermal conditions in processes of liquid chemisorption ( 9 E ) .

Miscellaneous General Theories. Neimark (6OG) extended the classification of adsorbents by the Kiselev method (nonrigid adsorbents and four types of rigid adsorbents: nonporous, coarsely porous, finely porous, and mixed). Singh and Band (78G) found that small atoms (helium) could submerge in the lattice to form two-dimensional alloys, resulting in apparently very high density. Nitrogen did not show the effect. Shereshefsky and Weir described two-dimensional phase formation for oxygen on glass spheres (77G). Temkin and Kul’kova postulated that in addition to the physical and activated types of adsorption, a third type exists, “deep” adsorption, as characterized by oxygen adsorption within the silver lattice (87G). Danon (78G) stated that the distribution function of the energy of adsorption on a heterogeneous surface can be deduced statistically from the isotherm, if known a t low concentrations. DeMarcus (79G) derived the transition probabilities between states of a n adsorbed molecule and showed them to agree with more rigorous treatment of Lennard-Jones. Danon reviewed and discussed the general area of statistical theories of adsorption (77G). Hepler discussed the limits for integration of an isotherm equation for nonuniform surfaces and derived the distribution of energy sites from heat of sorption (34G). Antropoff (2G) discussed adsorption and the Boltzmann-Euken “e-principle’’ for a modification of the Langmuir equation. For homogeneous surfaces he calculated the adsorption volume from heat data. Hill (36G) showed that a principle of

corresponding states could be used to furnish a reference behavior with which the location of steps in stepwise adsorption can be compared. Honig and Rosenbloom have developed a mathematical technique for the problem of surface heterogeneity which permits determination of extreme bounds defining the maximum possible variation in the isotherm data for a given system consistent with the simplifying restrictions (38G-40G). Kwan, Freeman, and Halsey (55G) developed expressions for the second and third virial coefficients between a solid and a binary gas mixture. Enderby discussed a domain model of hysteresis and suggested possible verification of some thermodynamic relations (23G). Zimm and Lundberg (93G) proposed a n exact “clustering” theory for various systems, bafed on statistical mechanics. Koutetskif (53G) simplified and generalized the Vol’kenshtein chemical adsorption theory. Seidl (76G) found that some metallic vapors impinging upon a surface a t high velocity do not adsorb unless they first strike another object. He concluded that adsorption takes place only when the energy of a molecule is below a critical value and that Frenkel’s theory is only approximately correct, and in the lower temperature field. Higuchi and coworkers (35G) concluded that the decrease of adsorption heat, activation energy, and activation free energy with surface coverage is due to the adsorption in a n ionic state and consequent double layer formation, for adsorption of alkali atdms on tungsten. Timofeev concluded, from assumption of exponential adsorption, that the adsorption mechanism was one of concentric layer formation, whereas the desorption was due to general lowering of a concentration which was uniform throughout carbon grains (82G). Gordieyeff (30G) found that no enhancement of toxicity of toxic vapors was obtained in the presence of biologically inert aerosols. Schay (74G)attempted to narrow the gap between the Langmuir and Polanyi theories by introduction of the concept of a n area of denial of adsorption to other molecules by a n adsorbed molecule, replacing the assumption of definite adsorption sites. He derived the Langmuir isotherm by statistical mechanics. Pavlovic (64G) found it possible to linearize the isotherm by a combined B.E.T.-Huttig equation. H e concluded from thermodynamic data that the second and third layers of several gases on nickel tungstate are not formed in a n ordinary manner. Todes and Lezin (83G) discussed the possibility of two types of sorption, depending upon the relative size of the velocity of heat propagation and velocity of solution

propagation. Aston, Tykodi, and Steele (4G) conceived a model of a heterogeneous surface with five types of adsorption sites and no lateral interaction. Predicted heat of sorption curves for this model agree with the types most often observed experimentally. Tovbin and Baram (843) studied liquid-phase desorption of iodine from charcoal and derived correlations for time and agitation. Thermochemistry of Adsorption. In addition to the thermodynamic and thermal data discussed above, a number of other studies in the area were published. Pierce (65G) and Millard and others (59G) compared heat of wetting and heat of sorption of carbon blacks. The former concluded that capillary condensation occurred. Amberg, Spencer, and Beebe ( 7 G ) showed that graphitized carbon black has a highly homogeneous surface, based upon heat data and the isotherm. Ross and Winkler showed that these data were consistent with the model of the adsorbed gas as an ideal two-dimensional gas a t lowcoverage, a n ideal van der Waals gas a t intermediate coverage, and agreed with the Langmuir equation at higher coverage (7%). &For the stepwise adsorption observed in such materials Clark (75G) showed that the vertical discontinuities represent equilibrium conditions. He found the isosteric heats calculated therefrom to agree with those measured calorimetrically by Amberg and others (7G). Huttig and Hart1 (42G) postulated that the energy of reaction of ammonia with metal halides was due to the expansion of the lattice and then fitting the ammonia molecule therein. The heats of adsorption observed agreed with the calculated second portion of the composite energy. Umeda and coworkers (85G) discussed the adsorption isotherm for a heterogeneous surface, with heat of distribution uniform, or varying linearly, exponentially, or hyperbolically. They concluded that the heterogeneity of the surface is not given precisely by the distribution function. Halasz, Schay, and Szonyi (37G) derived three different relations by which the differential heat of sorption could be calculated. Zettlemoyer and coworkers (92G) proposed a model for localized adsorption on a heterogeneous surface without lateral interactions. Bagg and Tompkins (5G) measured the heat of sorption of various gases on iron films and deduced heterogeneity of the surface, attributed to different modes of rehybridization of surface orbitals. Honig (37G) modified his previous comments on determination of adsorption energy heterogeneity of surfaces. Garden and others ( 2 7 G 2 9 G ) de-

VOL. 49, NO. 3, PART ll

*

MARCH 1957

465

UNIT OPERATIONS REVIEW terinined the thermodynamics of intercrysralline adsorption and measured thermodynamic properties of gases adsorbed on chabazite. Balandin derived equations for free energy of sorption of dehydrogenation catalysts (6G). Thermodynamic data \\-ere published for sorption of hydrocarbons on rutile (75G), krypton on titania (63G), and carbon oxides and disulfide on charcoal (77G). Ross and Good investigated the mobility of adsorption of burane on carbon black (69G)and concluded that a localized oscillator model is more reasonable than that of a tivo-dinirnsional gas. Utsugi (86G)derived the thermodynamic properties of a mobile adsorbed molecule in a heterogeneous external field, and also for the case of partly restricted motion. \Vhite (SOG) derived equations expressing the conditions of equilibrium for a strained solid adsorbent in contact \vith liquid adsorbates. Preparation of Adsorbents. This subject has been covered in part above. Blackburn and Kipling described the effect of ash and methods for its removal from charcoal (7OG). Sarin and Puri (77G) suggested the use of sulfonated coal as a desiccant and coinpared it to silica gel; the). state that it is a promising industrial material. Hashirnoto patented a calcined mixture of prrlite and bentonite for adsorption of oil and water mixtures from floors (JUG). LIaxted and Josephs have studied the regeneration of poisoned catalysts by gasphase desorprion (57G). The bulk of the preparations reported concerned silica gel and similar materials. Iler (43G) discussed preparation of “estersils,” modified gels io make them compatible fillers for compounding. by preparation Lvith ester groups. :\rnctt and others gave added organophylic properties to silica gel by grinding in the presence of polymerizable eth!-lriiic monomers ( 3 G ) . A Spanish patent (73G) !vas issued for silica gel useful as a desiccant \vhich ivould adsorb u p to 80y0 of its ireight of Lvater. \‘ariation of pore size of silica gel \vas shocvn to be possible using alkali or gaseous hydrofluoric acid. Berestneva and Kargin described i\vo ne\c silica-alumina gels, but gave no data (7G). Ku\rada and Sugawara prepared globular gel by several methods. The effect of drying temperature and wash liquor on the properries of silicaalumina, carbon-silica, and carbonalumina was also derived (5-G). Higher porosity was obtained \rhen drying was carried out a t a slightly elevated temperature. Wankat ( 8 9 G ) produced firm hydrogel spheres by passing a sol through a suspending medium of light gas oil or kerosine. Methods. Bering and Serpinskif de-

466

scribed a new apparatus for adsorption of gas mixtures, consisting of a combination of volumetric and gravimetric procedures. The adsorption is measured by the quartz spring technique and the residual gas volumctric mcthods (3G). The same workers developed a method of independent {veighing, using duplicate quartz spring apparatus a t t\vo differenL temperatures (SG). A microelectrode apparatus for adsorption studies of halogens was reported to give sensitiviry of 7 X 10-5s (JD‘G). Garbarski and Folman (36G) used the change in electrical capacity to study adsorption of water on glass, lvhile Yoshino used infrared adsorption to study loiv t c m perature adsorprion of nitrogen Jvithin a thin film of silica gel (97G). Zirninerman studied \cater on silica gel \virh nuclear resonance techniques ( 9 l C ) . while Deryagin and Zorin used optical micropolarization for various vapors on plane glass surfaces (2UG. ? E ) . Xieyerson (5%)discussed thr effect of adsorption on mass spectrometric studies. and conxwted a liability into an asset to detect minor components. Khodot usrd high pressure autoclaves to establish adsorptive equilibrium and determined the adsorption by rcIno\d of successive increments of sample and then desorbing and thus derectinq 1he adsorbate (XiC). James and 5IcKinries used adsorption a t moderate pressures rn influence the course of reactions Ivhich normally require high pressurc. The reaction took place in the condensrd phase upon the adsorbent (-1-lG). Schafrr and others determincd adsorption rates using a retardation of diffusion process 1 7 X ) . Rosenbers (liSG) discussed the use of thermistors to measurc adsorption of vwy sinal1 areas of solids by use of krypton. Khodot and Yanovskaya discussed a niultiple quartz spring apparatus for simultaneous determination of sorption of coal (77G). Clauss (76G‘) studied tht: iorsional elasticity of a gold \rire as afyecied by adsorption and found that the efyect !vas greater undcr static than under dynamic conditions. Keppler and coworkers (-/E)developed a multiple gas-liquid chromaiographic apparariis for LIS? “1’ lo 2 X C c:.

Adsorbents of Major Interest Charcoal ( / . I - >.I. 5A IR, i R . OB. OB 7dB--7AR, 5hC, 371;. 3iE, 5F,2G, 70G. 77G. 7JG. 27G. 38G. IIG, i i C . 66G.82G, 81C.

___

X