THE JOURNAL OF
PHYSICAL CHEMISTRY (Registered in U. 6. Patent Office)
1
(Copyright, 1952, by the American Cliomioal Society)
Fo1L?ldCd by J,Vildc~D.Ha~Crojt NOVEMBER 15, 1952
VOLUME56
NUMLtJGl
8
THE STKUCTUKE OF GRAPHITE 0XIL)E BY B. J. BECKETT AND R. C. CROFT Uwision oJ lnduslrial C'lieinislry, CottiirrotLweullh, ScienliJic uiLd Induslrial .Research Oryuniaulion, Alclbocirnc, Australia Received J u l y SI, 1961
Difficulties relating to tlic dcterniinatjion of t l ~ struct>urc e of graphite oxide have been reviewed. Fresh cvidriice indicating the structural nature of t8hiscoin ound is presented. This was obtained from electron niicroscopc studies of small particles of South Australia11 graphite andPt8hcositle prcpared from it. The latter exhibits considerable folding but the fomier does not,. This diffcrencc is atttril)uted to the redist,ribution of valency linkages between carbon atom8 in the layer planes of graphite occurring wlicii the latter is converted to graphite oxide.
Introduction Most of the lamellar compounds of graphite have been shown by means of chemical and X-ray methods to consist of the original aromatic carbon layer planes of the graphite crystal separated by intercalated ions, atoms or molecules, which are regularly arranged both relative to each other and to the carbon atoms in the layer planes above and below them. The many attempts, however, which have been made to obtain similar data for graphite oxide have met with little success. The applicahas failed to establish tion of chemical rt definite empirical formula for this substance. Examination of the oxide by the X-ray diffraction method5 only shows that it is composed of parallel lamellae spaced at greater distances than the caimbon layer planes in pure graphite. Failure of both methods to elucidate the structure of graphite oxide appears to be due t o the ease with which this substance absorbs water and other liquids. Siiicc the complete removal of these absorbed liquids is virtually impossible,6 chemical analysis of oxides prepared in different ways yields variable results. Furthermore, because liquid absorption increases the interlamellar spacing in graphite oxide until the lamellae are ultimately dispersed as two(1)
U. Hofmann, Kolloid Z.,104,
( 2 ) Zusammengefasst
112 (1943). bei U. Hofmann, Erg. esnkt. Naturwiss., 18,
229 (1939). (3) G. Ruess, KoEloid Z.,110, 17 (1945). (4) H. Thiele, 2. anorg. allgem. Chem.. 190, 145 (1930); Kolloid Z., 66, 129 (1931). ( 5 ) U. Hofniann and R. Hoist, Ber., 71, 754 (1939). (6) U. Hofmann, A. Frenrel a u d E. Csrtlaii, Ann., 610, 1 (1934).
dimensional sheets, it evidently weakens those interlamellar forces which might preserve a regular structure, as in graphite, and therefore permits relative displacement of the oxide lamellae. The extinction of all lines in the X-ray diffraction powder pattern of graphite oxide other than those of (0001) and (hkiO) planes is further evidence of this effect. Of the few structures s u g g e ~ t e d for ~ , ~graphite oxide, that proposed by Hofmann appears to be best supported by experimental data. He obtained evidence from X-ray measvrements that the oxygen i n graphite oxide is bonded t o the carbon atoms of the hexagon layer planes by an epoxy linkage (see Fig. 1). He was unable, however, either to measure changes in the C-C distances or to confirm buckling of the carbon layer planes, both of. which would be caused by a redistribution of carbon valencies. There is some chemical evidence that the graphite oxide structure can also contain hydroxyl ~ o u s . ~Ruess3 ,~ has confirmed this and has in addition obtained results which show that graphite oxide is capable of being hydrated to various extents. Hofmann and Holst2 showed also that it is possible t o prepare graphite oxide in which the atomic ratio of carbon t o oxygen approaches 2. The carbon atoms in the layer planes of which graphite is composed are linked by strongly directional bonds of the order 1.5. If, however, oxygen is linked in a covalent manner t o these atoms, this bond system is broken down. Taking epoxy group (7) J. Woiss, Naluve, 114, 744 (1940).
R. J. BECKETT AND R. C. CROFT
030
VOl. 56
TABLE I DETAILSOF PREPARATION O F COLLOIDAL GRAPHITE A N D GRAPHITE OXIDESUSPENSIONS Sainple
Obtained f r o m
1 (Graphite oxide) 2 (Graphite oxide) 8 (Graphite)
4 (Graphite oxide)
.
3 (Graphite oxide)
Oxidation t r e a t m e n t
-325 mesh purified South Australian graphite
Staudenmaier
Same Pipetted from water suspension of homogenized -325 mesh graphite, at 28 em. after 32 hr. settling Graphite pipetted from water suspension of homogenized -325 mesh graphite, a t 2 cm. after 2 weeks settling Same graphite as used to prepare sample 3
Brodie
Formation as ail example of how oxygen may become attached to the carbon layer planes when graphite i s converted t o graphite oxide, it is apparent from Fig. 1 that the epoxy group, if itself rigid, only confers structural rigidity on every second C-C linkage. Also, formation of this group results in the bond .order of the other C-C linkage being reduced from 1.5 in graphite t o 1 in graphite oxide. Whether saturation of carbon valencies in the graphite layer planes is achieved by formation of epoxy groups or other types of covalent groups, viz., peroxide, the redistribution of C-C valency bonds must be the same as described above. This view is supported by results of chemical analysis performed on graphite oxide by Hofmann and Holst.2 Their results indicate that the oxygen in this compound is attached only t o adjacent carbon atoms. r---0
A
0-OXYGEN @-CARBON
Fig. 1.-Layer-plane
of graphite oxide according to Hofrnann.
.U.
If it is preaumed that the rigidity of layer planes decreases as the bond order of linkages between carbon atoms composing them is reduced, then, according to the arguments set out above, thin lamellae of graphite oxide should show a greater tendeiicy to bend than graphite particles of the same dimensions. Confirmation of this theory \vas sought by examining dispersions of each of these materials under the electron microscope. Experimental Coarsely crystalline graphite from Uley, South Australia, was purified by altcrnatc treatments with hydrochloric and hydrofluoric acids until impurities were reduced to less than
....
Method of preparing susiiension
Oxide shaken with dilute (1: 3) ammonia for 24 hr., then agitated by ultrasonics for 12 hr. Same Pipet sample diluted with 1:3 ammonia subjected to ultrasonic agitation for 8 hr.
Three Urodic oxidations
Shaken with dilute (1:3) ammonia for G hr.
Same
Shine D
0.5% and was finally ground to minus 325 mesh (Tyler). The purified graphite was shown by X-ray powder photographs* to contain 95y0of the Bernal structure, the remaining graphite being in the Lipson and Stokes form. Suspensions of colloidal graphite were obtained by homogenizing a mixture of the graphite with tannic acid and water in the ratio 100:5:120 by weight. After homogenizing the paste was agitated with water and allowed to settle. Pipet samples were removed from this suspension a t depths and after settling periods shown in Table I. Some of the pipet samples were subjected to agitation by ultrasonic waves (500 kc./s.) with the object of further reducing the size of particles suspended in them. Graphite oxide samples were prepared by Brodic’sg method both from the sized fractiuns and from the original purified graphite. This was done by warming graphite with a mixture of fuming nitric acid and potassium chlorate. Some samples, however, were prepared by Staudenmnier’slo method. In this case oxidation is effected by agitation with a cold mixture of concentrated nitric and sulfuric acids ( 2 : 1 by volume) containing potassium chlorate. After washing, the partially oxidized graphite is heated with acid permanganate solution. Various methods were employed to disperse the oxides to the extent necessary for inspection under the electron microscope. Some samples were shaken with either water or dilute ammonia. I n the case ot the oxide prepared from the original purified graphite, large particles were separated by decantation. Some of ihe ovide suspensions were subjected t o ultrasonic treatment. The methods of preparation of colloidal graphite and graphite oxide suspensions are summarized in Table I. Suitably diluted suspensions were mounted on specimcn screens in the usual manner for electron microscopy and the specimens shadow-cast with 10 A. thickness of uranium before examination in the electron microscope, the shadowing ratio being 4:1. The InicroscoDe uscd was an R.C.A. model E.M.U. instrument fitted kith a self-biased gun and having the objective lens asymmet8ry corrected. An adjust,able objective aperture was einploycd11 and t,he magnification was calibrated by the method of Farrant and Hodge.12
Interpretation of Micrographs The main feature of graphite oxide evident from electron-micrographs in Figs. 2, 3, 5 and 6 is the folding which occurs in particles of this substance. Fully oxidized samples of the oxide (containing 30y0 oxygen) whether prepared by Staudenmaier’s method (Fig. 2) or by Brodie’s method (Fig. 3) (8) These were taken by 4 . McL. Mathieson of this Division. (9) E. C . Brodie, Phil. Trans., 149, 249 (1859). (10) L. Staudenmaier, Ber.., 31, 1481 (1898); 32, 1394 (1899); 33, 2824 (1899). (11) J. L. F a r r a n t aiid 4 . J. Hodge, J . Sci. Znslrumenls, 27, 77 (1950). (12) J. L. Warrant arid A. J. Hudge, J ,A p p l i e d Pliys., 19, 840 (1948).
032
Nnv., 1952
THESTRUCTIJRE OF GRAFHITK OXIDE
933
934
I'ig. li. (;r:qhiilc 5); Ihirkwss of pariiclvs 511-2110 .\.
show extensive folding which is general and not confined to mv part,icular area. Other samples containing smaller amounts of oxygen exhibit ,less folding. In all cases, however, folds extend rlght through even the thickest oxide particles and appear also t,o be predominantly about straight lines. The latter have no obvious preferred direction. Although Figs. 5(a) and 6 show that it is possible for sheet,s of oxide to be folded back on themselves, t,his vas seldom observed. Further evidence of the readiness of t,he oxide lamellae to bend is shown in Fig. 5(b), in which a long narrow sheet of oxide is seen to conform to the contour of an underlying part,icle like a carpet on st,airs. It mas mentioned earlier that the powder diffract,ion patatern of graphite oxide contains only (hki0) and (0001) reflect,ions. This implies that, although parallel, the oxide lamellae are not orientated in a particular manner relative to each other. Figure 5(a) suggests that slipping has caused rclative displacement, of the lamellae in a composite particle of graphit,c oxide. It shows a number of almost parallel lines at one edge of a relatively thick pIat,e of oxide. Sineeall of these lines are the same shape and sho\y t,he same irregularities, they presumably originat,e from t,he same graphit,e flake. Micrographs of graphite show that, although some of the particles phot,ographed are t,hinner than those of graphite oxide, they do not exhibit, the folding which may possibly be expected of thin flexible sheets. Comparison of Figs. 4(a) and 4(b) with Figs. 5 and 6 illustratcs this.
Discussion The conclusion that the above results support the theory that, valence linkages between carbon atoms in graphite arc redistributed when the latter is converted to graphite oxide, is based on the assumption that the aromatic structure of graphite layer planes originally proposed by Debye and Scherrer'a is valid. This assumption, however, is just,ified by data obtained by more recent investigators, among whom Bernal," Laidler and Taylor,15 Lipson and Stokes,IG are notable. These data all confirm that carbon at,oms in the layer planes of graphite are disposed hexagonally and linked, each with it,s three neighbors, by 1.5 order valcncy bonds as in aromatic compounds. It is suggested, therefore, that oxygen intercalated in graphite forms cotTalent linkages &h carbon atoms in the Iaycr planes by saturating the free valencies which, in graphite, these at,omsshare by resonance. The difference in the rigidit.ies of layer planes of graphite and its oxide is demonstrated by folding of lamellae which occurs only in the case of the oxide and is believed to be due, as explained earlier, to modification of strongly directional 1:5 order C-C linkages in graphite layer planes by covalent linking 6f oxygen to adjacent pairs of carbon atoms in these planes with consequent r e d i d o n of the order of bonds between remaining carbon atoms. This interpretat,ion is in agrcemcnt with Hofmann'ss (131 P.Debye and P.Seherrer, Phimile. Z., 18, 281 (1817). (14) ,I. D. DernJ, P ~ cROV. . Soe. (Landon), A106, 748 (1824). (16) D. S. laidler and A. Taylor. N o l w r , 146, 180 (19401. ( 1 6 ) FI. Limon and A. R. Stokes, < h i d . . 149. 32R (1942).
Nov., 1952
POTER'TIALB OF
ADSORPTI~R'-D.ES~RPTIOR' CAPACITY
proposed graphite oxide structhre which, because of the relative disorder of the oxide lamellae, would be difficult to confirm by X-ray measurements. Although sheets of graphite oxide appear folded in electron micrographs, it is probable that, when suspended in a suitable liquid, they are extended
PRAXS
935
and flat. The possibility that the folding just referred to was due entirely to the method of mounting specimens and not to structural weakness in the oxide lamellae is discounted because the same method of mounting did not produce similar effects in graphite lamellae.
APPLICAATZONOF THE CATHODE-RAY OSCILLOSCOPE T O POLAROGR4PHIC PHENOMENA. 111. POTENTIALS O F ADSORPTIONDESORPTION CAPACITY PEAKS AND SURFACE CHARGE DENSITY RELATIONSHIPS EXHIBITED BY ALCOHOLS AT AQUEOUS SALINE SOLUTION-MERCUR'I' INTERFACES BY J. WESTLOVE LAND'^ A N D PHILIPJ. ELVIR'G'~ The Pennsylvania Slate College, State College, Pennsylvania Received August 28, 1961
Differential capacity oscillograms for the saturated solutions of n-octyl and n-heptyl alcohols show four capacity peaks on each of t8hecat,hodic and anodic sweeps ratjher than the espected two peaks, indicating double film format,ion. The patterns for a sat*uratedn-hesyl alcohol solution are characterized by two and sonietimea three capacity peaks on the charging and discharging curves. A saturated solution of n-amyl alcohol gives only two adsorptioii-desorl)tion capacity peaks in each trace, indicating the formation of a mono film layer. The lower the molecular weight of the alcohol, the greater is the potential span between capacity peaks for the saturated solutions. For undegassed supersaturated solutions of the four normal alcohols, only two peaks per branch were observed. This picture reverts to that of the saturated solution upon degassing. The potentials of the reversible capacit'y peaks are given for both saturated and supersaturated solutions. In general, the desorption processes were found to proceed a t the same or a t a faster rate than the adsorpt,ion process aa indicated by the sharpness of the corresponding capacity peak heights. Calculations of surface charge density according t80the capacity peak potentials for the saturated solutions indicate that adsorption processes depend directly on the surface charge existing at the mercury surface.
In previous papers2 experimental arrangements were described which were capable of producing differential capacity and surface charge density patterns on the face of an oscilloscope. The application of these techniques to the observation of film formation a t the dropping mercury electrode, D.M.E., was indicated. The present paper is concerned with the results obtained from a systematic study based on oscilloscopic observation of the adsorption-desorption phenomena exhibited by various alcohols at the D.M.E., ie., a t the interface between the mercury and aqueous salt solutions. Heretofore, the study of the effect of non-electrolytes on the surface properties of a mercury electrode have been carried out for the most part by the determination of the surface tension which exists between a mercury drop surface and an electrolyte solution to which has been added some organic material. The usual procedure is to determine the surface tension over a selected applied voltage range and to plot the values of surface tension as a function of the applied voltage to give what is commonly called an electrocapillary curve. Gouy3f4published the most extensive surface tension data available for electrolytic solutions, both with and without the addition of non-electrolytes, The shape of the electrocapillary curve for a pure
electrolyte is parabolic in form with a maximum coming at the potential of zero charge on the surface of the mercury. In the presence of slightly soluble organic compounds the maximum is either flattened or shifted to a potential different from that of the original electrocapillary maximum. F r ~ r n l t i n likewise, ,~ found that the presence of adsorbable organic substances in solution modified the shape of the electrocapillary curve from that of the pure solution. Frumltin, et aL16determined the influence of capronic (caproic?) acid and phenol on the surface tension of a mercury-sodium sulfate solution boundary by a capillary-electrometer method and found that in the neighborhood of the saturated solutions the two organic compounds are adsorbed as multilayers. Alternating current measurements by Prosburnin and Frumkin7 of the capacity a t a mercury surface in contact with a solution of sodium sulfate and octyl alcohol, showed abrupt increases in the capacity of the electrical double layer at potentials at which the adsorption and desorption processes of the alcohol took place, and a decreased capacity at intermediate potentials. Grahames studied the capacity and resistance a t a mercury electrode in contact with solutions of potassium nitrate, sodium chloride and hydrochloric acid, all saturated with octyl alcohol, as a function
(1) (a) Sun Oil Co.; Norwood, Pennsylvania. (b) University of Michigan, Ann Arbor, Michigan. (2) (a) J. W, Loveland a n d P. J. Elving, THISJOURNAL, 56, 250 (1952); (b) J. W. Loveland a n d P. J. Elving, ibid., 56, 255 (I9ij2). (3) G. Gouy, Ann. c h i m . phya., [SI 8 , 291 (1906). (4) G. Gouy, ibid., [SI 9,75 (1906).
( 5 ) A. Frurnkin, 2. P h y s i k , 35, 792 (1926). (6) A. Frurnkin, A. Gorodetskaya a n d P. Chugunov, Acta Physi-
cochim. U.R.S.S., 1 , 12 (1934); C. A . , 29, 2040 (1935). (7) A. Proskurnin and A. Frunikin, Trans. F a r a d a y Soc., 31, 110 (1935). (8) D.C. Graharne, J . Am. Chem. Soc., 6 8 , 301 (104F).
