THE DIRECT EXAMINATION OF SOLS BY X-RAY DIFFRACTION

---

T H E DIRECT EXAMINATION O F SOLS BY X-RAY DIFFRACTION METHODS’ W 0 MILLIGAN

AND

HARRY B WEISER

Department of Chemtstry, The Rzee Instztute, Houston, Texas Reeeaved June 11, 1936

The application of x-ray diffraction methods to the determination of the constitution of sols has usually been indirect. X-ray diffraction studies have been made on the dry powder or moist gel obtained from the sol by precipitation with electrolytes, ultrafiltration, or centrifuging. In previous papers (10, 12) it was shown that, in general, the moist gels obtained by ultrafiltration of sols give the same x-radiograms as the dry powder. Thus with alumina, stannic oxide, and indium hydroxide sols, the moist gels obtained by ultrafiltration give the pattern of y-A1203.HzO, SnOs, and InsOa.3H20 or In(OH)3, respectively. Although it is often assumed that the sol particles have the same constitution in the sol state as in the moist precipitate, the direct examination of sols by x-ray diffraction methods should prove whether or not this is the case. Very little work has been published on the direct examination of sols by x-ray methods. Kraemer (5) in discussing the application of x-ray methods to colloids states, “this technic [x-ray method] yields no results for the surface of particles, nor is it effective (as yet) in dealing with colloidal solutions.’’ Bjornsthhl (3) examined gold and silver sols by allowing them to flow through a tube of gold-beaters’ skin centered in the camera and obtained patterns which agreed with those for metallic gold and silver, respectively. The important investigations of Bohm and Niclassen (1) on the gels of various hydrous and hydrated oxides and hydroxides, has been supposed by some people to include investigations of sols in the sol state. Professor Bohm in a private communication states, “Die Aufnahem fur die Arbeit in der Z. f . anorganische Chemie, 132, 1 (1924) wurden, wie Sie richtig vermuten, an den meist feuchten Ruckstanden von Solen nach dem Koagulieren oder Eindampfen (so beim Crum’schen Sol) gemacht.” Bohm and Ganter (2) examined laquid sols by flowing aged ferric oxide and vanadium pentoxide sols through a Mark tube and observed some indication of orientation of the needle-like particles. Our first work (10, 12) on the direct examination of a sol was 1 Presented a t the Thirteenth Colloid Symposium, held a t St. Louis, Missouri, June 11-13, 1936 1095

1096

W. 0. WILLIGAN AND HARRY B. WEISER

carried out on a thixotropic stannic oxide sol, which gave the pattern of stannic oxide. More recently Heller, Kratky, and Nowotny (4) examined various ferric oxide sols in thin glass capillary tubes. These investigators obtained the patterns of FeOCl, 0-FeOOH (ll),and 7-FeOOH, only; no indication of the formation of complexes was observed. In this paper will be given the results of an x-ray study of a number of representative sols with the object of obtaining direct evidence concerning the constitution of the colloidal particles. EXPERIMENTAL

The chief difficulties in the direct examination of sols are (a) the relatively low concentration of the dispersed phase, (b) the scattering of x-rays

1

SOL

FIG.1. X-ray diffraction camera for the direct examination of sols by the water in the samples, and (c) the possibility of coagulation of the sol by the x-rays. The first and second difficulties may be obviated by using fairly concentrated sols; and the third by flowing the sols through the camera. Some objections may be raised to flowing the sol through a tube of any kind. For this reason, in most cases the sol was allowed to flow in an uninclosed column, in the same way that Debye originally examined liquids. THE CAMERA

A Debye-Scherrer type of camera was employed, a diagram of which is shown in figure 1. The sol (or liquid) was allowed to flow into the capillary

EXAMINATION OF SOLS BY X-RAY DIFFRACTION METHODS

1097

tube. By maintaining a suitable head the rate of flow was adjusted so that a smooth, uninclosed column was obtained. I n some cases a thin Mark tube was inserted in the capillary tube, giving a column of sol inclosed in glass. The scattering from the Mark tube is so small that it causes little or no difficulty. A possible objection to the use of the Mark tube is that solid particles might deposit on the walls and give the x-radiogram. However, there was no indication of this. A blank x-radiogram, obtained for a used but unwashed Mark tube as a sample, gave the same result as a new Mark tube. Furthermore, the same patterns were obtained using the inclosed and the uninclosed column of sol. PREPARATION OF SOLS

Ferric oxide sol. A hydrous ferric oxide sol was prepared by the addition of a slight excess of ammonium hydroxide to a solution of ferric chloride. The resulting gel was washed rapidly by decantation until almost free from chloride, and a few drops of hydrochloric acid were added. Peptization and aging were brought about by warming for several hours a t 60-65 C., while stirring vigorously with a mechanical stirrer. It has been shown (9) that the freshly precipitated hydrous ferric oxide gives no x-ray diffraction lines or bands, but that the material aged as described above gives the a-FezOs pattern. This sol contained about 90 g. of ferric oxide per liter of sol. AZumina sols. The two alumina sols examined were made by methods already described in detail (10). Sol I was prepared according to the method of Thomas (7) by peptization with hydrochloric acid. Sol I1 was prepared by the peptization of precipitated alumina with hydrochloric acid. The gels from these sols gave the y-A120~.H20pattern (10, 12). Alumina sol I, which contained 11.0 g. of alumina per liter originally, was concentrated by evaporation to 55.0 g. of alumina per liter. Alumina sol I1 was concentrated by evaporation to 41.2 g. of alumina per liter. Stannic oxide sol. The method of preparation of this Zsigmondy stannic oxide sol has already been described (10, 12). The concentration of the sol was 33.4 g. of stannic oxide per liter. I n d i u m hydroxide sol. A hydrous indium hydroxide gel was prepared by the interaction of a slight excess of indium chloride solution and a solution of ammonium hydroxide. The gel was washed by centrifuging until peptization began, after which a few drops of hydrochloric acid were added. The sol contains 22.1 g. of indium oxide (InzO3)per liter. Beta ferric oxide monohydrate sol. The gel formed by the interaction of solutions of ferric chloride and ammonium carbonate was repeptized with an excess of ferric chloride. The sol was purified by dialysis in the cold for three months, and was concentrated by boiling on a hot plate. It would be expected (9) that p-FeOOH would form under thePe conditions,

1098

W. 0. iUILLIGAIi A S D HBRRY B. WEISER

since this material results from the slow hydrolysis of ferric chloride solution. But since the sol is dark red, and 6-FeOOH is yellow, the dispersed particles must, not be all P-FcOOH. The sol contains 89.6 g. of ferric oxide per liter. Titanium diozidc sol. This sol was prepared by the slow hydrolysis of titanium tetrachloride solution. d slightly acid solution of the salt was hydrolyzed by heating to boiling; the resulting gel of hydrous titanium dioxide was centrifuged to remove excess acid, and repeptized by suspending in water t o which a few drops of hydrochloric acid were added. The sol contained 30.8 g. of titanium dioxide ppr litcr. From the method of preparation, one would expect (8) the particales to consist of the rutile modification of the dioxide.2

FIG.2. X-ray diffraction pattern for a hydrous ferric oxide sol (a-Fe203)

FIG.3 . S - l a y diffraction patterns

Silcer sol. The silver sol used was a commercial protwted silver sol (“hrgyrol”) contaiiiiiig approximately 40 per cent silver. Silver iodide sol. A d v c r iodide sol was preparcd by mixing solutions of silver nitrate and hydriodic acid of such concentration that thc re.;ulting sol containcd 80 millimoli ilvcr iodidc and 8 millimoles hydriodic acid prr litrr. Thc pol concentration was t,licrrforc 18.7 g. of d v c r iodidr prr litcr. X-RAY BSAhfIKATIOU

A hydrous fcrric oxidc $01 n as prepared and examined first in February 1934 The sol was not examined in the camera described in this paprr, but 2 Investigations uliich vi11 he irpoited lntei a l e in plogless on the pieparation of gels of thc rutile modification of titanium dioxide

A

I

I

AG METAL

N 0

P "4:

II

Ad SOL

I,

It

A d POWDER

I

I nl

d Q 0 Q

-

Fro. 4. X-ray diffraction patterns 1099

1100

W. 0. MILLIGAN AiVD HARRY B. WEISER

was flowed through a “nonex” glass tube in the General Electric diffraction apparatus, using Rlo K, radiation. The pattern obtained after ninetysix hours exposure is reproduced in figure 2. The results compared with

A

B

E G

I K M 0 FIG.5 X-ra) diffraction patterns. A, HzO; €3, -y-AILOIHLO; E, SnOz; G, In(OH)3;K, @-FeOOH;K, TiOl(ruti1e); M, A g ; 0, AgI.

a-Fe20apowder are shown diagrammatically in figure 3. The pattern obtained with the sol is identical with that from crystals of a-Fe203. The other sols dcscribed above were examined in the special camera shown in figure 1, using a Philips cross-focus tube with Cu K, radiation

EXAMINATION OF SOLS BY X-RAY DIFFRACTION METHODS

1101

(nickel foil filter). The exposure time was 30 minutes in all cases reported. While workkg with the ferric oxide sols a filter of aluminum foil was placed between the sample and the film to prevent fogging of the film by fluorescent x-radiation from the iron. For purposes of comparison, the pattern of distilled water was also obtained. To identify the patterns from the sols, x-radiograms were made from the corresponding powders, using the same camera. This was done by placing the solid in a Mark tube attached to the capillary tube in the camera. The results obtained from pure water, the various sols, and the various solid materials are given in chart form in figure 4. Reproductions of some of the negatives are given in figures 5. Alumina sol I containing 11.0 g. of alumina per liter (not included in figures 4 and 5 ) gave the pattern of y-Al&.H20. DISCUSSION OF RESULTS

From inspection of figures 4 and 5 it is apparent that in every case the sol pattern is a composite of the patterns from pure water and the material making up the sol particles. Thus in the case of the weak indium hydroxide sol the water bands are relatively intense, whereas in the case of the strong silver sol the water bands are hardly visible. Only a slight indication of fibering in the pattern of the a-FezO3 sol was observed. This sol wm probably not aged as long as the one examined by Bohm and Ganter (6), who reported definite indications of orientation. As pointed out under the method of preparation, the 8-FeOOH sol is red in color, whereas pure 8-FeOOH is definitely yellow. The results indicate that the sol consists of two portions: (a) yellow p-FeOOH particles which give the lines in the x-radiogram, and (b) dark red particles of a-FezO3 which are too fine to give a definite x-ray pattern (9). This conclusion is supported by an earlier observation (9) of the slow settling out of some P-FeOOH particles during the aging of dark red sols. The results confirm the previous conclusions of the authors (10, 12) based on the examination of moist gels, namely, that typical hydrosols of oxides, elements, or inorganic salts consist essentially of aggregates of minute crystals of the respective hydrous oxides or simple hydrates, simple elements, or simple salts. There is no indication of the presence in the sols of complexes such as postulated by Pauli (6), Thomas (7), and others. SUMMARY

The following is a brief summary of the results and conclusions reported in this paper: 1. For the first time, a number of representative hydrosols have been examined directly by x-ray diffraction methods, taking precautions so that the resulting x-radiograms are for the actual sol particles.

1102

W. 0. MILLIGAN AND HARRY B. WEISER

2. A Debye-Scherrer type of camera has been designed especially for the direct examination of a completely uninclosed, flowing column of sol. 3. Alumina sols prepared (a) by the action of amalgamated alumina on water, and (b) by the pept,izationof precipitated alumina gave the pattern of ~-Al?Os.H20. 4. Ferric oxide sols were found to consist of particles of a-E’ezQs or P-FeOOH, depending on the method of preparation. 5 . X stannic oxide sol (Zsigniondy), a protected silver sol (“Argyrol”), a n indium hydroxide sol, a negat’ivcsilver iodide sol, and a titanium dioxide sol gave respectively the patterns of Siloa (cassiterite), metallic silver, In(OH)3, XgI, and Ti02 (rutile). 6. It is concluded, in agreement with previous x-ray studies on moist gels, that the common iiiorqanic hydrosols consist in general of particles of simple oxides or simple hydrates, elements, or salts, and not of coinplexes as postulated by some invest,igators. REFERENCES (1) BOHMAND XICLASSEN: Z. anorg. allgem. Chem. 132, 1 (1924). (2) R o H h f A N D GAXTER: Unpublished results reported in a private communication to the authors. (3) BJORXSTKHL:Dissertation, Uppsala, 1924. (4) HELLER,KRATKY, A K D KOWOTXY: Compt. rend. 202, 1171 (1936). (5) KREAMER: in Taylor’s Treatise on Physical Chemistry, Vol. 11, p. 1611. D. Van Nostrand Co., New York (1931). (6) PAULIA X D VALKO : Elektrochemie der Kolloide. Julius Springer, Vienna (1929). (7) THOMAS A N D COWORKERS: J. Phys. Chem. 36,27 (1931); J. Am. Chem. SOC.64, 841 (1932); 66, 794 (1934); 67, 44 (1935). (8) WEISER A N D MILLIGAS:J. Phys. Chem. 38, 513 (1934). J. Phys. Chem. 39, 25 (1935). (9) WEISER A N D MILLIGAN: (IO) WEISERA N D MILLIGAN:J. Phys. Chem. 40, 1 (1936). (11) WEISER ASD MILLIQAN: J. 4 m . Chem. SOC. 67, 238 (1935). Trans. Faraday Sot. 32, 358 (1936). (12) W E I S E R .4ND ?fILLIG