[ C i ) \ r R I l t L 1IO\ TKOM ISt E'r\KLMP \ I ( J fC H P M I S I R Y , I j V I V t R S I I T OF II,I I\OIS]
Fluorination of Methoxydichlorophosphine' B Y DONALD RAYRfARTIN ,\ND PHXLIP J. PIZZOL4TO'
The. application of the Swarts reaction to the fluorination of inorganic halides is well Imowri . Only in recent years have there been any attempt5 to evaluate the effect upon the fluorination reaction of the substitution of organic radicals fur halogen atoms. Booth, et al., have fluorinated some alkyl chlorosilanes2 and Emeldus and Heal' have studied the effect of fluorination upon the properties of the organic derivatives of the halides of sulfur, selenium and telluriuni. Alkoxy derivatives of phosphorus(V) sulfochloride have been synthesized by Booth, Martin and KendalL5 I t was decided to study the fluorination of an alkoxy halide of phosphorus(II1) in order to extend the knowledge obtained in the studies mentioned above, and to compare the results with those obtained in the fluorination of phosphorus(TI1) halidw Experimental Preparation and Purification of Methoxydichlorophosphme Methoxydichlorophosphine was prepared by L( modification of the procedure used by Kowalewski One mole of anhydrous methanol was added dropwise to one mole of phosphorus( 111) chloride with vigorous stirring 'it a reaction temperature ne11 below 0'. The metganol R J S added a t a ratt of 33 ml per hour. The reactor was rt three-necked flask fitted nit11 .L stirrer, drying tube cont,tining barium oxide, arid 'I dropping funnel. Upon thc completion of the additioii of ihe methanol, the reaction products were allowed to riw 4owly t o room temperature to facilitate in the evolution of the dissolved hydrogen chloride. Thc reindining mixtnre of reaction product> 'it ~5 fractionally distilled t w ICC contliicted in a fractionating The column packed with glass helircq r , 111 i. d.) i i i ~ ~ohtaincd 111 A purified m e t l i o ~ y d i c h l o r o p h ~ ~ p l iW'Iyield of about fill:(. Fluorination of Methoxydich1orophosphine.-Thc method and apparatus used for thr stepwise fluorination by meanq of the Swarts rextion was the same as previously described,6 except that no catalyst was required. The temperature of the fluorination was maintained below 10' and, for the hest yields of the chlorofluoride and the difluoride, the pressure within the generator was maint k i e d helox 50 mni. Apparently little or none of the chlorofluoride is produced when the pressure within thc 111111 I3rlov 25 inin. there appears incr~dsei n the yield with diminution may be due to tlie incredwd tendency tor the dichloridc to volatihzci :tlorig with the products of i h t fluorii~~tiori Tvpic.il L \ I X riiii( i i t i gi\11 thc yiclrli 5hon 11. (1) Presented hetore the D~\iriori of I'h) 5icnl aud Inorganic Chemistry of the American Cheinic.il Society 01 .\pril 17 1950 a1 Detroit, hIichigan ( 2 ) The Solvay Process Company, Hopewell Virmnla (9) II P Booth an(l P E4 Cum11 Tms J a ) i R V \ I 68, ?BJO2 t 6 2 (19.181, e l seg 4 1 15. J EinelCui and 1r C, Tied J C/I,-J~Ih c 1126-1131 11'316). I i) H
S Booth, 11 K Martin nid F I , Kendall T S J JT o~r IO, 2527-2525 (10.18) 6? TT Ih'>i 1
\
Koirale i - h
/
h r ~ i
\ -
29
K'.
11
217 2 2 2
The products of fluoriiiation were fractiorlally distilled in a modified Dufton column as previously described by Martin arid Fauit,' escept that in the electronic circuit for operating the controls of the column a Radio CorpoEttion of America vacuum tube (OA4G) was employed i l l stead of the. General Electric thyratron tube (GB7). Methoxychlorofluorophosphine and methoxydifluorophosphine were purified by redistillation in the :hovementioned column under the conditions shown CHIOPCIF CIIJOI'L'Z
Pressure withiu the column, min. 85 740 Temperature of head of column, "C. - 13.0 - 16.3 Analyses.--The samples were collected and weighed iii sealed glass ampoules of approximately 1 ml. capacity. These samples were hydrolyzed in 125 ml. of :L solution containing sodium hydroxide in slight escess over t h e quantity required for complete hydrolysis. The conceiitration of caustic was maintained as low :ts possible :is :i precautiori against obtaining silica in the solution, which would interfere with the analysis for phosphorus. The chlorofluoride and difluoride were hydrolyzed instantanc-ously, whereas the dichloride sample required :is long a' twenty minutes. Chlorine was determined by the Volhard procedure, the necessary precautions being exercised to prevent the phosphite from reducing the silver( I) ion .* Phosphoru., was determined gravimetrically as magnesium pyrophosphate after oxidation of the phosphite to phosphate by means of bromine and nitric acid, followed by a double precipitation as magnesium ammonium phosphate. CarI ~ o nand hydrogen were determitied by meaiis of the standard methods of microanalysis. Some difficulty w a i experienced in the determination of carbon in methoxydifluorophosphine. However, aeparate analyses for thv rnethoxy group checked very well with each other aiid with the theoretical value. The methoxy analysis was coiiducted on 7.5-3.0 g. samples. These were hydrolyzed mid the methanol fractionally distilled into it tared receiver.bJi) From the weight and density (d2",) of the distillate the precentage of methoxy group was Qualitative tests established the presence of fluorine in the chlorofluoride and the difluoride. -4 summary of the analyses is contained in Table I. Determination of the Physical Constants (see Table I). --Methoxydichlorophosphine and its chlorofluoride and difluoride derivatives are all colorless gases which condense to colorless liquids. Upon freezing, the dichloride and difluoride are transformed to white solids, but the chlorofluoridc forrn.z a trausparent glass. The freeziiig points were obt:iiiied by t h e procedure described by Booth and Martin.lS The vapor pressures w r c determined by the static niethod of Booth, Elsey : t i i d B ~ r c h f i e l d ,and ' ~ the nrce517) I). K. M a r t i n ; i u d J . 1'. F;iitst. J . / ' i i ~ . s ('011. Chein., 63, 1253 1202 (1949). ,S! H. S. Booth and A . K. Bozdrth. Tms J O U R X A L , 61, 2927-293 I 1R39). i o ! Clark Microanalytical Laboratory, Urbana, Illinois. ,IO1 0.v. Lupin, 2. onal. C h m . , 97, 210-220 (1931). i,11) 'A', Dittmar and C. A Fnwsitt, Traits. Roy. SOC.Ediicbargl!. 33, 909-534 (1887). , 12; Ti. S. Booth and 11 R M a r t i n , 'l'nis .JOURNAL,64, ?l98-22Oi I
;ind 1'. E
Harchfirlcl, zhid , 67,
FLUORIN.\TION OF ~\IETI-IOXYDICIILOKOPMOSPIIIEE
Oct., 19qYJ
4-5sc5
Although Emeldus and Heal4 observed that the displacement of halogen atoms by fluorine reduced the stability Of and phenyl derivatives Of sULfu~,selenium and tellurium halides, no evidence was obtained in this investigation in suppofi of this trend. ~ ~work ~ l with phosphorusCV) ethoxysulfodichloride and the chlorofluoride and difluoride derivatives thereof indicated that thermal stability increased with fluorine c ~ n t e n t . ~ Emeleus and Heal4 also observed that thermal TABLE I stability increased with the number of alkyl or CHsCH1OPClt OPCIF CHaOPI‘r aryl radicals introduced into a compound in place I3oiling point, ‘ C , 93.0 38.9 -15.5 of halogen atoms. Dimethoxychlorophosphine Heat of vaporization, g . cal./ was not prepared in this investigation, but earlier mole 8295 7077 5722 work on the derivatives of the reaction of ethylene ’l‘routon’s constant 22.7 22.7 22.2 7.8300 7.8362 7.7335 chlorohydrin with boron trichloride gave evidence \lapor pressure constants”{; 1812,6 1546,5 1250,5 that thermal stability increases with the number of 1,iquid density, dra 1.406 1.318 ... . ... alkoxy groups which replace halogen atoms. Freezing point, * 0 . 4 O -91.0 Glassb -117.3 is more stable than (C1C2H40)aThus, (C1C%H40)3B 100.01 132.93 116.47 . . . . . . . 115.2 100.6 BC1 which is more stable than C ~ C Z H ~ O B C ~ ~ . ’ ~ 9.03 10.31 12.01 Calculated It was observed that the substitution of a 9.03 10.14 11.04 Found methoxy radical for a chlorine atom in phosphorus2.27 2.60 3.02 (111) chloride produced a more reactive molecule. IIyrlrogen, 2.33 2.73 2.98 Calculated 23.31 26.60 30.98 Methoxydichlorophosphine is more easily fluoriPhosphorus, %{ Found 23.20 26.82 31.3 nated than phosphorus(II1) chloride a t compar53.35 30.44 . . . . . . ., Chlorine, ~ ( C a l c u l a t e d able temperatures. This is demonstrated by the ’ Found 53.24 30.24 . . . . .. . fact that phosphorus(II1) chloride requires a Calculated . . .. . .. . . . , 31.05 hlethoxy, $71 ... . . . . . .. 30.9 Found catalyst, while methoxydichlorophosphine does not. Similarly, i t is observed that a lower tema Constants in the equation: log pcmrn.) = A - BIT. Glass forms a t some temperature below - 79 perature is required to fluorinate phosphorus(V) ethoxysulfodichloride than is required to fluorinate Discussion phosphorus(V) sulfochloride. l5 The boiling point of the methoxydichlorophosThe ease of hydrolysis is greater with methoxyphine prepared in this investigation was observed difluorophosphine than with methoxydichloroto be 93.0°, as compared with a value of 94’ re- phosphine, as discussed above. This behavior is ported many years ago by Kowalewski.6 Al- the opposite of that demonstrated by the halides though the replacement of a chlorine atom in phos- of the non-metals which contain no organic radiphorus(II1) chloride by a methoxy radical lowers cals, but is analogous to the behavior of the phosthe molecular weight very slightly, the boiling phorus(V) ethoxysulfodihalides toward hydrolypoint of the methoxy derivative is slightly higher. sis, since it was observed that phosphorus(V) ethOther related properties such as the heat of vapori- oxysulfodifluoride is more easily hydrolyzed than zation and Trouton constant are changed corre- phosphorus(V) ethoxysulfodichloride.5 spondingly and indicate a trend toward the values for methanol, as shown below: Summary Heat of 1. Methoxydichlorophosphine has been synTrouton’s thesized by the reaction of methanol with phosCompound B. p., OK. c2af”zle constant phorus(II1) chloride (yield about 60%). PCll 349.2 7060 20.2 2. Methoxydichlorophosphine has been fluoriCHaOPCI? 366.2 8295 22 7 nated by means of the Swarts reaction to give CH9OH 337.8 8420 24 9 methoxychlorofluorophosphine and methoxydiStepwise replacement of the chlorine atoms by fluorophosphine. fluorine atoms in phosphorus(II1) chloride results 3. The above compounds have been purified in the lowering of the boiling points of the result- and characterized by determination of their molecing compounds by the increments shown below. ular weights, percentage compositions, freezing The methoxy derivatives exhibit the same trend, points, vapor pressures, heats of vaporization, but with more nearly constant increments. boiling points, Trouton’s constants and liquid densities. 4. The physical and chemical properties of PC1, 76.0 62 CHaOPCI? 93.0 5 4 . 1
sary temperature correctioFs have been applied. The values for the boiling points, heats of vaporization and Trouton’s constants were calculated from the equations representing the straight lines obtained when the vapor pressure data were plotted as log fi P I S . I j T . The vapor hensities were determined by the-R-egnault procedure. Aside from the hydrolytic reactions requisite t o analyses, no chemical properties of these compounds were studied. I t was observed, however, that the materials, when dry, did not attack glass. . The dichloride slowly attacked mercury upon standing, and all of the compounds exerted 5ome solvent action upon Lubriseal stopcock grease.
,
ydW(:;:
,
,,
,
,
,
,,
,
O.
PCI?F 13.9 81 PCIF? - 47 3 53 9 PFo -101.2
CHzOPClF CHsOPFl
:3 9 -10
5
;4 i
*
(14) D R hlartin and Leone S. Mako, unpublished laboratory notes. (15) H. S. Booth and M. Catherine Cassidy, THISJ o u R v A r , , 62 2369-2372 (1940).
these compounds have been compared with those
fluorophosphines, and other similar compounds.
of phosphorus(II1) chloride, methanol, the chloro-
RECEIVEDAPRIL7, 1950
[CONTRIBl:TION
FROM THE
DEPARTMENT OF CHEMISTRY, USIVBRSITY OF
NOTRE DAME]
The Preparation and Properties of Some Thienyl Butenols'" IPORGF 'I. G i i I r T m l h
li-hen a-thienylmagnesi urn 1m i 1lit1e wa5 ;t I lowed to react with butadiene nmnoxide there was obtained a hutend ( I ) which was teiitatively assumed to be I-(ct-thieriy1)-buten-%-01-1.This assumption was based on the preparatioii by Senieniuk and Jenkins' of a series of butenols (to which was assigned the general structure of R --CH2---CH=CH-CH20H (XI) by the reaction of various Grignard reagents with butadiene monoxid?. Thev favored the reaction course AJ,
N --hfgS
+ CH,
-
CH -CH-CH,
OhfgX I
+
HLO Hi - ------+ I3 -CH=CHz --+ K--CI12--CH=CH--CH20H (S)
K---CH?-cH---CH=CH, 12 - CI-Iz-CHOH
---+
an explanation of the formation of (X) rather than the expected R-CH2-CHOH-CH=CH2. They did not exclude however, the possibility of * ' 1,A-addition" of the Grignard reagent to butadiene oxide. Evidence for assigning the general structure (X) t o this series of compounds rested solely on the identification, by physical constants of the product obtained from the reaction of 11I ethylmagnesium b romi d e with butadiene mot 1oxide as penten-2-01-1. The possibility that isomeric products3 might result from the reaction of (;rignard reagents with butadiene monoxide made it desirable to cunfirni the tentative structure assigned to (I). Confirmation was obtained in the following manner, I-(cr-Thienyl)-buten-3-ol-2 (II) and 1- (a-thienyl)-buten-3-ol-l (III)* were prepared by the reaction of a-thienylsodium with butadiene monox3s
(la) Abstracted in part from the dissertation presented to the Graduate School of the University of Notre Dame bv Ceorne T Gmitter. ilb) General 'lire Research Fellow 194fi 1949. Present ?ddress Velsicol Corporation, Chicago, Illinois. C2) Semeniuk and Jenkins, J . A n Phnrrii 4ssnc, .5cz E d , 97, 118 (1948). , 3 ) Since the conipletion of our tnvesttgation, Gaylord a n d Beckcr (J.Org. Chcm., 16, 305 (1960)) have reported the formation of 141naphthyl)-buten-3-01-2 when butadiene oxide reacted with l-naphthylmagnesium bromide Jenkins and Semeniuk,* on the other hand, report having obtained otily I-( l-naphthyl)-hiiten-2-ol-4 from the same reaction. (4) Isobutylene oxide and styrene oxide have been reported by Henry (Compl. f r n d . , 146, 21 (19907)) and Tiffeneau and Foumeati tbid., 146, 647 (1908)) t o reart %I>isobutyraldehyde and phenyluxtafdebyde, respectively, with Grignard reagents. (111) was synthesized to exclude the possibility t h a t butadiene monoxide had re rcred with a-thirnrlni3jinP.ium bromide i n d 4 m i l n r farhiori
lvr) F?, T , m R~xror;
ide arid allyliiiagnesiuni hroniide with ?-thiophene aldehyde respectively. Alttempts to synthesize (11) by 'tn alteniate route, the reaction of a-thienylmethylmagnesiuni chloride with acrolein and the Barbier modification5 of this method yielded. as the only isolable product,6 sym-di-(a-thienyljethane. A comparison of the physical constants of (It with those of IT and I11 indicated (I) is not identical with either compound, a fact which was further confirmed by the difference in melting points of the corresponding 3,3-dinitrobenzoatesa XIthough the three hutenols (I), (TI) and (111) underwent cleavage on ozonizationi indicating the presence of an olefinic linkage8 in their structures. formaldehyde (isolated as a 2,4-dinitrophenylhy-drazane J was obtained as an ozonization product only in the case of (11) and (111) but not in the case of (I) a fact indicating that the olefinic linkage in (I) is not terminally located. Failure to obtain acetaldehyde as an ozonization product of (I) was considered to exclude the possibility that (1) possessed the structure Th-CH(0H)-CH=CH---CH3 which could conceivably result if butadiene monoxide rearranged to crotonaldehyde which then underwent a normal reaction with the Grignard reagent. Dehydration of trj, (TI) and (111) yielded ail unsaturated intermediate (Ivj which, without further purification9 was converted in each case to the same maleic anhydride addition compound !V j according to the Diels-Alder method. lo If the possibility of migration of a thiophene ring from its original position on the butene chain during dehydration is excluded, then the dehydration product (IV) must be 1-(a-thienylj-butadiene-l,3,the maleic anhydride addition compound (Vi, X(m-thienyl\-14-tetrahydrophthalic anhy" I , ciride, the carlxm skeleton of (I),Th--C-&--C-C
i
l
l
1
arid the origirial butenols must differ only in the (,?) 13arhier, Compl rcnd., 148, 110
(1899).
(8) Delaby (CoinP1 r e n d , 194, 1248 (1932)) reported t h a t the reaction of benzylmagnesium bromide with acrolein gave only a 6% yield of benzylvinylcarbinol ' i )Thiophene under identical conditions was not oxidized. 8) Attempts to establish the presence of an olefinic linkage in compounds (I), (11) and (111) hy catalytic hydrogenation were unsuccessful. The compounds failed to add hydrogen. (9) The tendency for a-thienyl butadienes t o polymerize was so pronounced t h a t attempts t o purify these compounds by distillation m vaczio and at temperature- as low a? 40-.50° led t o their extensive polymerization. 'IO) Dirls, Alder ani1 Priri, B o , 62, 2081 (1429).
