Vol. 82
1502
COMMUNICATIONS T O T H E EDITOR 3-DEOXYHEXOSONES
Acids quantitatively converted 3-deoxyhexosones 284 rnp) , to 5- (hydroxymethyl) -2-furaldehyde (A, The The 3-deoxyhexosones are accepted as intermedi- 2,4-dinitrophenylhydrazone1m.p. 203’. ates in the conversion of sugars by alkali to meta- yields after heating for 2 hours at 100’ in 2 N saccharinic acids,l and also have been postulated as acetic acid were 45, 80, and 0.5% from IIIa, IIIb, intermediates in the acid catalyzed production of and fructose, respectively. Alkali converted the furfurals from sugars2 Yet only the first member 3-deoxyhexosones by an internal reaction of the of the series, 2-oxopropanal, was known. The two Cannizzaro type t o a mixture of the corresponding possible 3-deoxy-~-hexosonesnow have been pre- metasaccharinic acids. Paper chromatography of pared and their behavior studied under alkaline the lactones showed only “a-” and “B”-metasaccharinic lactones. The yield of glucometasacand acid conditions. Di-~-fructose-glycine~ dihydrate4 (Ia) was heated charinic acids6 from IIIa was 2570 after only 5 at 100” for 4.5 minutes in dilute aqueous solution minutes of standing a t 25’ in oxygen-free 0.038 hi at $H 5. Decationization and concentration lime water. The “p” isomer was isolated as its 22’ (cf. Nef’). yielded 3-deoxy-~-erythrohexosone (IIIa) as a calcium salt [ a ]?;D iBecause the 3-deoxyhexosones are converted colorless amorphous solid. Traces of impurities rapidly by weak acid and alkali to 5-(hydroxy(less than lo%), which included glucose (1-2%) and mannose (1-2%) were removed by chromatog- rnethyl)-2-furaldchyde and metasaccharinic acids, raphy on a paper column giving the pure osone respectively, and because they give intense brown[Found: C, 44.53; HI 6.591. Di-D-tagatose-gly- ing with amino acids,a these highly reactive subcine4 (Ib) similarly gave 3-deoxy-~-threohexosone stances may be significant intermediates in various (IIIb). Both osones readily gave osazones (Table types of sugar degradations. I) in the absence of acid. ( 6 ) R’. S f . Corhett ; i n 4 I . Lenner. J . C h m . SOC.,1431 11!l33)
sir:
( 7 ) J. U . S e f , A n u ( h r i n . L i r b i q . ~ 376, , I(I9lO). ( 8 ) E . F.I, .IAnct. . I i i s i . J C ~ C J I12, L . ,491 (1959).
TABLE I SUBSTITUTED P H E N Y L O S A Z O N E S O F 3-DEOXYIIEXOSONESa M.P., O C . Osone Substituent (dec.) [aIz5D
2,4-Dinitr0-~ 206 +80OoC Triacetate 181 +5880d 2,5-Dichloro242 +400”” Triacetate 184 ...., 4-Xtro-b 260-262 +35OoC D-threo 2,4-Dinitro258-259 -5OOoc Tri-acetate 209 ..... 2,5-Dichloro 226 -208OE All were crystalline and gave satisfactory analyses. Identical with the corresponding osazone from 3-de oxy-^g1ucose.J Pyridine-acetic acid 1: 1. Chloroform. e Pyridine. D-evythro-
The formation of a Meoxyosone and the Lobry de Bruyn-Alberdn van Ekenstein transformation are considered to proceed through a common intermediate, the 1,?-enediol of the sugar.’ Dehydration of the sugar l,%enediol t o the 3-deoxyosone is normally a side reaction,’ but in the case of the diketose-amines the position is reversed and only traces of the epimeric aldoses are formed + I-I-N(R)CH:COO-
GI
CH @OH
HCPOH I
&(R)CH,COO-
11
CH I 3(‘OH
C8H;Oj I (Enol form)
II
H~IRKHKOOCHO
I
CH
CHI
C iH;Oi I1
C.jH,O : I11
1
(1) J. C. Speck, J r . , Advaitces in Cavbohydrale Cheii?.,13, 63 (1958). (2) &I. L. Wolfrom, R. D. Schuetz and L. F. Cavalieri, THISJOURN A L . 70, 514 (1948). (3) 1.l’-(Carborymcthyliminn)~his-[l-deoxy-~fr~5e]. (1)I < . 1; I.. J. h n e t , Ai(s1. J . Chcirz , 12, 280 (I%!]). (3) E. F. L. J . h n e t , unpublished results.
DIVISIUSOF C.S.I.K.O.
I ~ C ~ OI’RESERVAI‘IOS D ASD ‘rRASSPORT
E. F. L. J. ANET
HOMEBUSIf, ~ . ~ , !.US . CRALIA ~ i ~ .
RECEIVED FEBREARY 8,1060
TRANSITION METAL CATALYSTS. 111. NATURE OF T H E ACTIVE SITE I N ORGANOMETALLIC CATALYSTS
Sir: There is considerable interest in defining the active species in olefin polymerization catalysts formed from an organometallic reducing agent and a transition metal compound. Since two different coinpounds are employed and the compounds are usually electron deficient molecules, earlier investigators postulated, and later isolated, bimolecular coordination complexes having the halogen bridge structure’, p
x
\
I’
hfi ,/
k~
‘>-
j
X’
hI1 = Electropositive metal from organo-
/
metallic reducing agent such as aluminum.
\
i L I 2 = Transition metal of Groups IV-VI of the Periodic Chart (Ti, V, Cr).
Mr
Regardless of whether such complexes are essential or not, there is the separate problem of which metal is the site of chain propagation. Our copolymerization studies are pertinent to this problem. Ethylene and propylene were copolymerized in cyclohexane using a ”l., vigorously stirred autoclave4 at 30 p s i . and 70”. Catalyst was injected (1) G. S a t t a , P. Pino, G. X a z z a n t i , E. Rlantica a n d hl. Peraldo, J. P o l y n w Sci., 26, 120 (195i). (2) G. N a t t a , P.Corradini a n d I. W. Bassi, TIIISJ O U R N A L 60 , , 7.55 (1938). (7) \V.I. C a r r i c k . tbid , SO, 043.3 (1038). 4) U S u t h c r l a n d a n d j 1’. IicKenzic, Ii;t1, E i i , . (1101: , 46, 17 (l!IX).
March 20,1960
COMMUNICATIONS TO THE EDITOR
into the pressurized reactor by hypodermic syringes and the pressure was maintained a t 30 p.s.i. with no monomer vented. Copolymer propylene contents were measured by infrared absorption a t
7.25 u. Propylene is less reactive than ethylene with all the catalysts employed and it accumulates in the reactor causing a linear increase in the average propylene content of the gross copolymer with increasing yield of copolymer (Fig. 1). Ethylene and propylene show distinctly different relative reactivities with catalysts prepared from a cominon reducing agent, Al(i-Bu)3, and different transition metal compounds. Propylene relative reactivity increases in the series HfC14 < ZrC14 < T i c & < v o c 1 3 < VCL. This increase appears to correspond to increasing electronegativity of the transition metal center. However, there is no significant difference in the relative reactivity of the two monomers when the reducing agent is changed from A ~ ( ~ - B uto ) ~Zn(C6HS)2, , Zn(n-Bu)2, or CH3TiC13,6 using a common transition metal compound, V C 4 . These reducing agents differ widely in general steric configuration and in bond hybridization, electronegativity, valence, and size of the metal ion. The common denominator in this series is the reduction of vc14 wholly, or partially, to the divalent ~ t a t e .Significantly, ~~~ the two transition metal compounds (CH3TiC13 vc14), form a highly active catalyst and the dependence of copolymer composition on polymer yield is that for vc14 catalysts rather than for the Al(i-Bu)3-TiC14 catalyst .' Selection between two monomers a t the growing chain end is a sensitive measure of the structure and polarity of the active propagation site, and a change in this selectivity implies a change in the nature of the active site. The fact that the relative monomer reactivities are altered by changes in the transition metal center but not by changes in the reducing agent implies that propagation occurs a t the transition metal center, with no direct participation by the reducing agent.3 This further suggests that the formation of bimetallic complexes involving the metal of the reducing agent is not an essential feature of the catalytic function. Natta and co-workers also have observed differences in comonomer reactivity ratios using catalysts composed of aluminum trihexyl and certain transition metal halides,* but they did not vary the reducing agent. The available data do not allow a rigorous comparison of the present data with the published results; however, a qualitative comparison indicates that the propylene content of the initial copolymer in their case is higher than the values reported here for the three transition metal compounds common to both studies (TiC14, VOC13, vC4) by a factor of 0.1-0.5. No specific significance is attached to the fact that the two sets of data are not identical since the experiments
+
( 5 ) W. L. Carrick, A. G. Chasar a n d J. J. S m i t h , Paper No. 4, Polymer Division, 134th Meeting of t h e American Chemical Society, Chicago, Illinois, September 7-12, 1958. ( 6 ) Belgian P a t e n t 533,447. (7) Experimental points are not given for t h e AlRa-TiCla case since more t h a n one active species is present and reproducibility is poor. ( 5 ) C. N a t t a , J . Polymer S c t . , 34, 21-48, 88-92 (1959).
lo
1503
r VCI.
L
e
4
6
8
10
12
14
Copolymer yield in aromr.
Fig. 1.-Effect of catalyst structure on copolymer composition: in all experiments the monomer feed contained 28% propylene, 69% ethylene and 3% ethane (area per cent. b y vapor chromatography).
were carried out with different reaction diluents, pressures, temperatures, degree of monomer dispersion, and possibly different degrees of reduction of the transition metal. I t is significant that the trend is the same in both cases. The published opinion of Natta and co-workers's8 favors propagation from an aluminum center in a bimetallic complex, although their copolymerization data are more consistent with our view above3. RESEARCH DEPARTMENT WAYXEL. CARRICK J. KAROL UNIOXCARBIDEPLASTICS COMPANYFREDERICK L. KARAPIXKA DIVISIONOF UNIOXCARBIDECORP. GEORGE JOSEPH J. SMITH BOUNDBROOK, NEW JERSEY RECEIVED JANUARY 26, 1960 STRONGLY BASIC SOLUTIONS I N TETRAMETHYLENE SULFONE (SULFOLANE)
Sir: The small autoprotolysis constant and moderately good ionizing characteristics of sulfolane suggest i t as a solvent in which one could study a very wide range of acidities.' Using solutions of phenyltrimethylammonium hydroxide in aqueous sulfolane, we have been able to measure P K a ' s of neutral proton acids distinctly weaker than any previously evaluated by the Hammett method.2 Table I presents these values along with those of other weak acids obtained by similar methods in aqueous hydrazine3 and aqueous ethylenediamine. H- = 20, our most basic value of the H- function2 (Fig. l), is about 4 units more basic than any proper H- previously e ~ a l u a t e d . ~The , ~ 0.01 d l solution of base in sulfolane containing 5 mole yo water is over lo6 times more basic than the corresponding aqueous solution, i.e., H - is over G units greater. The contribution to this exalted basicity resulting from the reduction of water activity HA
+ OH-=
A-
+ HzO
(1) R . L. Burwell, Jr., and C. H.Langford, THISJOURNAL, 81,3799 (1959). (2) M. A. Paul a n d F. A. Long, Chem. Reuiews, 67, 1 (1957). (3) N. C. Deno, THISJOURNAL, 74,2039 (1952). (4) R. Schaal, J . Chim. Phys., 61, 784, 796 (1955). We believe t h a t certain details of Schaal's procedures subject his values t o substantial error.
