positions of the Si and 0 atoms in crystals of di- bonded dimer, while the broad band appearing a t alkylsilanediols. If the linear (sp) structure higher concentration and still lower frequency is were correct, one would expect greatly decreased attributed to higher polymers.21g22 The silanols, basicity, since in hydrogen bond forniat,ion T - like alcohols. develop two lower-frequency OH electrons are much less basic than unshared pairs. bands as the concentration is increased. The Pi-bonding involving both electron pairs of oxygen position of the first low-frequency band for the tnay be important in the nenrljp lincar silosai~es'~conipounds studied is given in Table I in the and perhaps also in silanolate ions. "dimer" column, in terms of the frequency difIntermolecular Association of Silanols and Car- ference (Av)between this band ancl the non-bonded binols.-Data on the degree of association also are OH absorption. There is as yet no firm evidence summarized in Table I. The fraction associated that the nature of the species responsible for the was determined by the deviation from Beer's law first low frequency band is the same for alcohols with increasing concentration shown by the first and silanols. However, it is interesting to note overtone of the non-bonded OH absorption."' that the hydrogen bonds in this first associated Only the alkyl substituted silanols and carlinols species appear to be substantially stronger for were sufficiently soluble in CC14to permit ineasure- silanols than for carbinols. This confirms Ratuev's ment of the degree of association. ;2rnong these observation that the band shift on polymer forcompounds, the silanols are somewhat more as- mation in trimethylsilanol is exceptionally large." sociated than the carbinols a t any given concen- Particularly strong hydrogen bonds in the associtration, but the differences are not large (Table I). ated species would be expected for the silanols if The differences between the triethyl and trimethyl they are much more acidic but only slightly less compounds in either series are as large :is the dif- basic than alcohols. The frequency of the free 0-I-I absorption band ference between, e.g. trimethylsilanol and trimethvlcarbinol. This suggests that steric factors of the silanols studied is significantly higher than are very important in determining the degree the free 0 - H band for alcohols. The silanols of association and that decreased hindrance alone absorb in the region 3675-8688 cm.-' while the might explain the enhanced association of the carbinols absorh at 3iOO-3614 cm. -I. Carbinols generally absorb in the region 3605-3645 cm-1.23 silanols. As the concentration of an alcohol in an inert Batuev and his co-workers previously have noted solvent is increased, the first absorption band t o the high frequency of the free 0-H vibration in appear a t lower frequency is by common consent trimethylsilanol ancl attributed this shift to enattributed t o the 0-H vibrations in the hydrogen- hancement of the 0-H bond force constant by the electropositive silicon atom." Pi-bonding from (17) M. K a k u d o a n d T. Watase, J . C h e m . P h y s . . 21, l f i 7 (1!153); oxygen to silicon, and the probably larger Si--0-H Bzlll. Chem. SOL.J a p a ? ~27, , 6 0 5 (1954): I f .K a k u d o , S . Kasai a n d T. Watase, J . Chenz. P h y s . , 21, 1891 (1953). bond angle. may also influence the position of the (18) M. Tamres, THISJoURN.4L, 7 4 , 3375 (1952) ; R.TVe?t, ihid.. 0-H absorption and a t present the evidence seeins 81, 1614 (1959). to us insufficient to explain the phenonienori. (19) R. C. L o r d , D. W. Robinson a n d W. C. Schumb, t b l d . , 78, 1,527 ~
(1956); 6. R. F, Curl, Jr.. a n d K. S. Pitzer, ibid.,80, 2371 (19A81. Acknowledgments.-- The authors are indebted (20) If the nature of the association reactions mere knorrn, o u r specto Lliss Karen J. Lake for her patient assistance tral measurements would permit calculations of t h e cquililirium cons t a n t s for these reactions.21 I n the absence of definite knowledge with the preparation of diphenylsilanol. a b o u t t h e nature of the association reaction in silanols a n d carbinols, we have preferred t o leave our d a t a in the form of per cent. association (22) F, 9. Smith a n d G , C. Creitz, J . Rescorch T a l l . Bur. .Slandnvds. a t several concentrations. These fifiures should be directly com~iarahle 46, 151 (19.51). f o r silanols a n d carbinols. (2:j) L. J. Bellamy, "The Infra-red Spectra of Complrx lIolPciiIes,'' (21) R. Mecke, Disc. F a v a d a y SOC.,9 , l G l (1050); N. D. C ~ ~ g a e s h n l l 2 n d Ed., R,Iethiien and Co., T.ondon. 1958, 1). q7. a n d E. I,. Saier, THISJ O U R N A L , 73, 6414 (1951); Cy. T-iddel and IS n. MADISON. WISCOXSIS Becker, S j h ~ c l ~ o ~ hdic~l n ,. 10, 70 (19,57).
[CONTRIBUTIOS FROM THE CHEMISTRY DEPARr5IENT O F
ST. LOUISUSIVERSITV
Reactions between Dry Inorganic Salts. X. The Effect of Rubidium Chloride on the Transition Temperature of Cesium Chloride BY LYXAN J. ~ ' O O D ,CHAS.STVEENEY, S.J.,' AXD SR.39.THERESE DERBES' R E C F I V r R .\PRIL
2(7, 103'3
A study of the effect of ruhidiurn chloride on t h e tr31isiti011temperature of cesium chloride has been made by means of the high temperature X-ray caniera. It has been estahlishFd t h a t increasing amounts of added rubidiutn chloride cause a spectacular lowering of the transition temperature and that, at room teinperzture, u p t o 30 inole (.% of the high temperature forin of cesium chloride may remain in the rubidium chloride in the form of 3 solid solution.
I n a previous publication2 i t was shown -that addition of cesium bromide to cesium chloride
raises markedly the temperature a t which the Pm3rn-tFm3m transition begins and that the
( I ) Taken in p a r t from theses presented t o the faculty of St. 1 , o u i q University for the degree of Master of Science.
( 2 ) L J. Wood, \Vm. Secunda and C . H. Llc13ridp. Tnrs J O U R N A L . 80, 3 0 i (1978).
Dec. 5, 19.59
RE.~CTIONS BETWEEN DRYINORGAPI‘IC SALTS
resulting Fm3m phase consists of a solid solution of Fm3m cesium chloride and a previously unknown Fm3m form of cesium bromide. In this work study has been made of the changes in the Pm3m+Fm3m temperature caused by increasing amounts of rubidium chloride added to cesium chloride. I n 1910 Zemczuzny and Rambach3 reported breaks in cooling curves for mixtures of cesium chloride with small amounts of rubidium chloride which can now be correlated with the Fm3m+Pm3m transition. n’ith increasing amounts of rubidium chloride, these breaks in the cooling curves became smaller and occurred a t lower temperatures up to about 10% rubidium chloride after which they could no longer be observed. The results presented below, obtained by means of the X-ray camera, will show t h a t the transition temperature lowering does not stop a t 10% rubidium chloride and t h a t the Fm3m phase t h a t is stable at room temperature consists of a solid solution of Fm3m (high temperature form) cesium chloride in rubidium chloride. Bridgman4 has shown t h a t a t sufficiently high pressures rubidium chloride undergoes a discontinuity in its density curve a t room temperature, and Paulingj has shown t h a t this discontinuity may correspond to a transition from Fm3m symmetry to Pm3m symmetry. Jacobs6 confirmed the discontinuity in the density curve of rubidium chloride, reported by Bridgman, but using X-ray analysis a t high pressures failed to find any Pm3m diffraction lines. However, Wagner and Lippert? have shown t h a t rubidium chloride can exhibit Pm3m symmetry when sublimed onto Pm3m thallous chloride a t -190’. It therefore appears possible that if rubidium chloride can, in fact, exist in the Pm3m form a t low temperatures, it might form a solid solution with Pm3m cesium chloride. Such a solid solution might be expected to undergo the Pm3m-tFm3m transition a t a higher temperature than t h a t of pure rubidium chloride. It has been part of the purpose of this investigation to examine this possibility. Apparatus and Materials The cesium chloride and the rubidium chloride were t h e same as previously All salt mixtures, unless otherwise indicated, were prepared by thoroughly mixing powered samples of the previously melted pure salts. Each mixture was then fused, cooled to room temperature, ground and stored in a desiccator until needed. T h e salt mixtures were examined at high temperatures by means of a furnace built into the powder camera furnished by the General Electric Company for use with the XRD-3D X-ray diffraction unit.8 The sample tube, which had very thin walls and a n outside diameter of 0.3 to 0.4 mm. was mounted on a ceramic shaft in such a way t h a t it could be rotated a t the center of t h e camera. T h e molybdenum Ka doublet was used for all the diffraction analyses. A control thermocouple for automatically controlling the temperature was placed above the sample tube and just outside the X-ray beam. This thermocouple was connected to a thyratron shift-phase control circuit.8 A temperature measuring thermocouple was placed below and just outside ( 3 ) S. F. Zemczuzny a n d F. R a m b a c h , Z. anovg. Chem., 66, 418 (1909). (4) P. W. Bridgman, Phys. Rev., 67, 237 (1940). ( 5 ) L. Pauling, “ T h e N a t u r e of t h e Chemical Bond,” Cornel1 University Press, I t h a c a , New York, 1945. ( 6 ) R. B . Jacobs, Phys. Rev., 54, 325 (1938). (7) ‘2. Wagner a n d L. Lippert, Z . p h r s i k . C h e m . , B33, 297 (1930). (8) C. lf.McBride, M.S. Thesis, S t . Louis University, 1957.
6140
the X-ray beam. This thermocouple was connected t o a type K Leeds and Xorthrup potentiometer. This thermocouple was calibrated by observing its reading when a sample of pure cesium chloride was a t its transition temperature of 472” as recently found by Wood and co-workers.* The transition temperature was approached from the low temperature side, and great care was taken to avoid overshooting the observation temperature.g X-Ray patterns were photographed a t temperatures just above, just below and very near to the Pm3rn e Frn:itn transition temperature. The sample mas exposed to X-rays for 9 hr., and the temperature was checked a t 1 5 min. intervals for the entire time. By adjusting the automatic temperature control when necessary, t h e over-all temperature variation was held t o about one degree centigrade. This same exacting procedure vvas followed, when necessary, for all observations reported below. Efforts to calihrate the temperature measuring thermocoupie by comparison with a supplementary thermocouple placed a t the same position were not successful. The reasons for this failure are important but cannot be discussed here.
Results and Discussion The unit cell edge of Fm3m cesium chloride was determined a t 475 and 480’ (Table I) by comparison with the unit cell edge of aluminum and with that of Pm3m cesium chloride.’O The unit cell edge of aluminum was obtained a t higher temperatures by the use of a temperature coefficient of expansion determined by Hidnert and Krider (2.57 X 10-5).11 Several unit cell edge values for Fm3m cesium chloride and for rubidium chloride have been calculated for several temperatures t h a t will be involved in the discussion that follows (Table 11). Because of too few data, neither the 4.2 X coefficient of Wagner and Lippert for Fm3m cesium chloride nor the 2.7 X coefficient for rubidium chloride can be said to be rigorously established, but both may be considered adequate for the present use. Both of these coefficients seem reasonable when compared with the very carefully determined temperature coefficient of expansion of 2.18 X for potassium chloride as reported by Glover.12 The relation between these coefficients becomes Fm3m CsCl : Fm3m RbCl:Fm3m KC1 = 4.2 X 2.7 X 10-4: 2.18x 10-4. Using the high temperature X-ray camera and the thermocouples described above, a study was made of the Pm3m-tFm3m transition in mixtures of rubidium chloride and cesium chloride containing increasing percentages of rubidium chloride (Table 111). At sufficiently elevated temperatures (always lower than the transition temperature of pure cesium chloride) these mixtures consisted of a single crystalline phase having Fm3m symmetry. T h e unit cell edge of this phase was always larger than t h a t of rubidium chloride and smaller than that of Fm3m cesium chloride (calculated) a t the (9) This was necessary, because of t h e hysteresis found b y Wood a n d co-workers. Upon cooling f r o m a few degrees above 472O t h e Fm3m P m 3 m transition occurs a t 465’. Preliminary results obtained in t h e author’s laboratory indicate t h a t if t h e 472 degree transition is exceeded only slightly a n d if t h e cesium chloride remains a t this high temperature f o r only a s h o r t time, t h e transition range is narrowed. It appears possible t h a t untransformed nuclei m a y , under these circumstances, remain for a brief time. Such nuclei could seed t h e transition back t o t h e low temperature form as t h e temperature is lowered. These observations. which are n o t y e t conclusive, are t o be reported in a future communication. (10) L. J. Wood and J. Wm. Vogt, THISJOURNAL, 66. 1259 (1944). (11) Peter Hidnert a n d H. S. Krider, J . Reseavch N a t l . Bur. S t a i d ards, 48 (1952). (12) R. E. Glover, Z. Physik, 138, 222 (19.54).
T,. J. ~ ' o o J )C. . SWEENEY, SJ.,.omSR. AT. TIIERESE nrnnns
6150
1'01. SI
TABLE I SOME
UNIT CELL EDGES FOR OQ
Fm3, CESIUM CHLORIDE AND
Reported (A.)temp. (T.)
Observer
FOR
RUBIDIUM CHI.ORIDE Remarks
Aluminurn used as a n internal standard Fm3m CsCl 475 Compared with a0 = 4.123 L.lO for Pnn3rn CsCl at 30" At or near Pm3m Fm3m transition 7.042 482 Afenary, Ubbelohde, Based on one line from a single crystal detemp. (472") n'oodwardl8 termination 7.04 445a Lf'agner and Lippert'd Originally reported as 7.03 kX. units. Fecalcd. by comparison with no = 4.123 A. for Pm3m CsC1 a t 30" 7.05 475 Calcd. by the use of the Pm3m :Fm3ni cell edge ratio = 0.6b 480 7 11 This work Compared indirectly with aluminum a$ a n Fm3m CsCl internal standard A t slightly above the Pnn3m +.Fni3ni 7.10 480 Tliis work Compared with a. = 4.123 a . l o for Pm3m transition temp. (472') CsCl a t 30" 4(;0" 7.10 Originally reported as 7.07 kS. uiiits. Recalcd. by comparison with n u = 4.123 A. for ~ m 3 mCsCl at 30" 6.591 Wood and PaprothZ1 Aluminum used as a n internal standard 30 Fm3m RbCl This work 6.645 260 Aluminum used as a n internal standard 6.658 2 59 This work Compared with a0 = 6.591 for RbCl a t 30" 3.74 - 190 Wagner and Lippert' KbCl deposited on Pm3m TiCl a t - lS(1° Pm3m RbCl Extrapolated from - 190" to 30°d 3.83 30 3.95 Calcd. by the use of the Pm31ri:FinOni cell 30 edge ratio = 0.6* a Work in progress in the author's laboratory by G. A. McLaren on the cesium chloride-potassium chloride system has shown t h a t as little as one mole '$6 of potassium chloride present in cesium chloride as impurity could be expected to lower the transition by 10 to 15'. This small amount of potassium chloride would lower $e unit cell only a few units ill t h e third decimal place and easily could go undetected. An extrapolated value of 4.23 ..2. was first obtained for 475" by the use of what appears to be an accurately determined temperature coefficient by Menary, Ubbelnhde anti Woodwardi3 chic11 was then divided b y the Pm3m: Fm3m cell edge ratio.* This calcillation cannot be espected to yield innre than approximate results. This temperature is 15 degrees above the transition temperature reported by Wagner and Lippert. The results reported in this work confirm the rather sudden large increase in unit cell edge immediately above the transition temperature reported b y Wagner and Lippert. Precision measurements or the temperature coefficient of expansion of F,,,3,,,cesium chloride in the first few degrees above the transition temperature might make a very interesting study. -1temperature c i ) efficient of 3.8 X was used for this calculation which was based on the TlCl unit cell edge value of Wagticr and 1,ippi'rt a t - 190" and at room temperature. The 3.82 value must be considered to be only an apprositnation. 7.046 7.043
475
This work This work
-f
TABLE I1 TABLE 111 UNIT CELL EDGES AT VARIOTSS TEMPERATURES AS CALCU- COMPLETIOS TEMPERATURES FOR THE CESIUM CII1,OKII)E I.ATED B Y b l E A N S O F THE LIXE.4R COEFFICIENT O F EXPANPm3m -+ F I d m TRANSITION IN CESIUMCHLORILXRUBIDICM CIILORIDEI ~ I X T U R E S ~ SION A N D AS OBSERVEDCUBEEDGEAT SOMEKson-N Completiim TEMPER.4TtIRE Max. temp. for hIin. temp. f u r temp. for F m 3 m CsCl Fm3m R b C l 5101e rG 1'm:im f F m 3 m piire Frn3n1, I'in3in Vm:3111 at 7.10 + 4.2 X 10-4 ( I = ~ c,.m+ 2.7 x i n - 4 I~1,CI oc. oc. OC.
-
(1
(Wagner and Lippertla)
T , 'C.
a0
(A.)
-
( t - 200) ( T h i s wurk)
- 4SO)
T , "C.
a0
('1 1
525 480 obsd. 365 260
7.119" 7.10 7.052 36.5 6.680 7.008 2 60 6 . 6 5 2 ohscl. 30 6.911 30 6 , 5 9 1 ohstl. a T h e previously reported relative value of 7.075 was not calibrated by means of a n internal standard and was not corrected for opacity of the sample. This value, on a n nbsolute basis, is somewhat too low.
a.2
corresponding temperature (Table 11), and this unit cell edge decreased for larger additions of rubidium chloride. In view of these observations i t seemed reasonable, indeed necessary, to conclude that in the region of the single crystalline phase, all of the cesium chloride existed in the Fm3m form in solid solution with the rubidium chloride. (13) J. VV', M e n a r y , A. R. Ubbelohde a n d I. Woodward, Pmc. R o y . SOC.(London), 8208, 158 (1951). (1.4) G. T a g n e r a n d I,. Lippert, Z . Physik. Chem., B31. 263 fl ( , ; ,,!28. 1 I2i f l R i O ' 1
these lower temperatures. The composition of the Fm3m solution of Table I V was determined by testing a new series of mixtures containing increasing amounts of rubidium chloride (Table V). As the percentage of rubidium chloride increased from 50yG,the intensity of the Pm3m pattern decreased progressively, until a t 70 mole of rubidium chloride the Pm3m pattern could no longer be observed even after thorough annealing. When as little as lp0 of pure cesium chloride was mixed mechanically with this 70yGsample, the Pm3m pattern again could be seen easily. The observations show that the solubility of cesium chloride in rubitiiuiii chloride a t room temperature cannot be far from 30 mole 5; and that the resulting solid solution has Fm3m symmetry. TABLE Y ROOMTEMPERATURE X-RAYASALXSISOF MIXIURESC O S TAISING 50-70 MOLEToRbCl Composition of mixt. in mole R CSCl RbCl
Description of P m 3 m G cesium chloride pattern
50 50 Strong and definite 35 65 Definite 66'/3 n'eak b u t defiiiite 33 '/3 312/3 ij81/3 IYeak-only one Pni31ri h i e 30 70 S o Prn3tn lines obsd a Each mixture produced a strong Fm3in patterii.
If Vegard's law be assumed, the composition of this so!id solution and the unit cell edge value of 6.653 A. shown in Table IV can be used for calculating a hypothetical room temperature unit cell edge for Fm3m cesium chloride. The value obtained is 6.803 A$. which is appreciably less than the extrapolated value of 6.911 A. shown in Table 11. The agreement is as good as could be expected, considering the uncertainty of the extrapolation and the possibility that Vegard's law may not hold rigorously. X series of observations was made on an 8.5: lC5 molar mixture of rubidium chloride and cesium chloride over a wide range of temperature. At all temperatures only an Fm3m solid solution was found having about the unit cell edge and teniperature coefficient of expansion to be expected for such a mixture. I n view of the various observations described above, it does not seem unreasonable to postulate the possibility of an Fm3m+Pm3m rubidium chloride transition in the presence of PmXm cesium chloride a t lower temperatures, although no such transition was observed a t temperatures included in this investigation. It is hoped that quantitative iiteasureinents inade in the low tentperature range can be reported a t some future time. S T 1,0111q
I,?rfK5scirTI?I
