THE COLOURING MATTERS O F CAMWOOD, BARWOOD, ETC. 125

By PAULINE O'NEILL and A4n~~wrc GEORGE PERKIN.

OF ihe four well-known I' insoluble " red dye-woods, sanderswood, barwood, caliaturwood, and camwood,':' t he first-named has been most closely examined, and although barwood and caliaturwood have received somu attention, no description of the colouring matter of camwood appears t o be on record. The close similarity in the dyeing properties of these woods, and the fact tha t all are substantive to wool, suggests tha t the dyes present are chemically allied, and it is indeed conceivable, from the very similar shades given by sanderswood, 73arwooc1, and caliaturwood tha t the same colouring matter niay be present in each case. Camwood, on the other hand, gives with mordanted wool somewhat bluer tones,

Chroniium. Aluminium. Tin. Iron. Camwood . . . .. . . .. Reddish-violet Red Bluish-red Violet Sanderswood . . . . . , Brownish-red Orange-red Red Naroon

and its colouring matter, according to dyers, is more readily dis- solved by water thaa tha t of the other dye-woods of this class. A chemical comparison of this with thz santalin of sanderswood appeared desirable, t he latter having somewhat recently been closely studied by Caiii and Siinonsen (T., 1912, 101, 1061). These authors have suggested for santalin the formula C,,N,,O,, also considered probable by Weyermann and Haff ely ( A nnalen, 1850, 74, 226), and find tha t this may be expressed as

Of the numerous derivatives of this colouring matter described in their paper, santa,lin dimethyl ether, C,,TT,0,(OMe)3, is specially interesting in that, by oxidation, veratric, anisic, and butyric acids can be prepared from i t . That sanderswood contains, in addition t o santalin, a second colouring matter appears possible from the work of Weidel (Ze i t s ch . fur Cherri., 1870, 6 , 83), who obtained from it, in addition to his colourless, crystalline santal, C,H,O,, a bright red substance, C,,H,,O,, soluble i n alkaline solu- tions with a purplish-red colour. Curiously enough, in the more recent accounts of sanderswood, this paper appears to have been lost sight of. The present investigation, although commenced a considerable

* A less known insoluble red wood is the Narra wood of the Philippine Islands, Pterocarpus, spp., which has been investigated by Brooks (Zoc. ci t . ) .

voL. CXIII. F

C,,H,O,(OH),(OW.

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online / Journal Homepage / Table of Contents for this issue

126 O’NEILL AND PERKIN : THX COLO-URING MATTEES OF

time ago, is of necessity somewhat incomplete, and must remaiii so for the present, on account of other work whii-3h in the present ciroiimstances is of greater i1nport;t1mt-e.

I n order to obtain some idea as to the solverlt properties of the constituents of camwood, the fractional extraction of a few grams of the powdered wood in a Soxhlet apparatus was first carried out, and for the sake of comparison, samples of barwood and sanders- wood were similarly treated. Employing in the order given, benzene, ether, ethyl acetate, acetone, and alcohol, fractions very similar in appearance were obtained from each of these woods, as i f , indeed, all contained the same coiietituents. From the behaviour of the solvents, however, i t became apparent that more than one colouring matter was present, for, after the removal of coIourIess substances by means of benzene, ether somewhat slowly dissolved a red colouring matter, arid when this ceased to pass into solution, a considerable amount of a brownish-red substance, evidently the main product, was extracted by means of commercial ethyl acetate. Subsequently, by nieans of acetone, a small amount of a similar compound was isolated, whereas the final alcoholic extract contained traces of a darker and somewhat; uninviting- looking product.*

I n working wit’h camwood on the larger scale, such a process was impracticable, aiid an alcoholic extract was first prepared. I n order t o isolate the mixed colouring matters, the precipitation of these as lead salts by means of lead acetate according to Meier’s method (L4rc7~. Yharnz., 1845, 55, 285) was at’ first adopted, but the subsequent filtration processes were so extremely tedious that a more rapid method was devised. This consisted in the addition of concentrated barium hydroxide solution to the wood extract in boiling 80 per cent. alcohol, there being thus formed a precipitate of the barium salt, which filtered rapidly, and was subsequently washed wi th 80 per cent., and finally with boiling alcohol. The reddish-brown precipitate suspended in water was decomposed by means of dilute hydrochloric acid, the cruds colouring matter collected, and, in order to remove as far as possible obstinately retained mineral matter, boiled two or three times with 5 per cent. hydrochloric acid solution. When cold, the product; was repeatedly ground with water to decompose completely the oxonium hydro- chloride present.

* By exhaustion with alcohol, camwood game 16 per cent. and sanderswood 18.4 per cent. of extract.

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOD, BARWOOD, AND SANf)ll:RS\VOO I). 124

The dried substance was now dissolved in a little alcohol, the :jc>llltion stirred with pui3ied and, thB dcohol remo-Jed hy evttpora - tion, t he residue finely ground, and extracted in a large Soxhlct apparatus with coniinercial ethyl acetate. At the eiid of t he operation some quantity of the colouring matter had separated out (A4), arid from this the supernatant liquid was decanted, evaporated t o a small bulk: and poured into ether, causing the depositioiz of some qnantity of brown precipitate, which was added t o A . That fraction soluble in ether is referred to later as B.

A was dissolved in absolute alcohol, the solution treated a t about G O o with alcoholic potassium acetate solution, and the volnininous precipitate of the potassium salt collected and washed with alcohol (coinpsre Caiii and Simoziseii, 7oc. c i t . ) . Itl was decomposed iii war111 aqueous suspension with dilute sulphuric acid, t h o reddish-brow11 product, after washing, allowed to dry on porous porcelain, dissolved in a little alcohol, and reprecipitated by pouring into ether. Finally, i t was digested with boiling purified ethyl acetate, which removed a trace of more readily soluble substance, and repeatedly uashcd with the same solvent.

Foulld: C-65-68, 65.49; 11-5.47, 5.52. C,,Te,,Q, requires C = 65.69 ; H

This substance, for which the nanle i~osnutdiri is proposed, consists of a dark chocolate-coloi~red powder which o n grinding becomes redder in appearaiice. As delmsiited from dilute alcohol, i t could iiot be obtained iiz a definitely crystalline condition, and although when isolated fmni ethyl acetate the product frequently exhibited minute serrated edges, detached crystals have as yet iiot been observed. When heated, it’ showed no signs of melting, had darkeiied a t 280°, and was fully decoinposed at. 290-300°, beiiig then a carbonaceous powder. It’ is readily soluble in boil- ing methylated spirit, butt with hot absolute alcohol, it resiiiifies and does noL readily pass into solution. The general properties of this substance are described later in comparison with those of s a 11 tali n .

Although the clementary aiialysis of isosan talin was in agree- ment with the €ormula @1511,,0, proposed for saiitalin by Caiii slid Simonsen, Zeisel determinations ‘k of the former showed tha t it could not possess this formula, which would require, as deter- mined by these authors, CH3==5*4’i per cent. Indeed, the best expression fo r the figures obtained is given below, as the result of the analysis of numerous preparations dried ah 160O.

* These were carried out hy Perkin’s method, with and without acetic nnhydride, and employing as a preca,ution biilbs containing a suspension of amorphous phosphorus.

F 2

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

128 O’NEILL AND PHBKITU’: THE COLOUSINQ MATTERS OF

Found: CH,=6.82, 7.01, 7.03, 6.85, 7.30. C,,H,GO,(O*CB,), recjuircs C = 65.75 ; R 7 5.02 ; 6: If, - 6.85 per

cent.

The potassium salt, prepared with boiling alcoholic potassium acetate, was repeatedly washed with alcohol, and consisted of an amorphous, reddish-brown powder. I n case the isosantalin is impure, the precipitate is not granular at, this lempernture, hnt congeals to resinous lumps.

Found: K=2.71, 3.021, 3 * O S , 3.06, 2.26. *

(C,,R,,O,K)(C,,PP,,O~)~ requires K== 2.18 per cent. C7,HG5OC4 requires K = 2.88 per cent.

These figures were again distinct from those obtained by others for the potassium salt of santctlin, C30H2705K, which requires K = 6.76 per cent.

To decide between the two forniulz given above Is not easy, for although the latter agroes best with the analyses obtained, isosantalin obstinately retains traces of ash which naturally accumulate iii the precipitate, and, again, this ainorphous potassium salt is not easy t o wash. When an alcoholic solution of ismantdin is treated with a mineral acid., a ,n i:iteizse reddish-violet solution is obtained, which is evidently due to the fornzstiom of alp oxoniux salt. These compounds, owing to their ready solubility in alcohol and acetic acid, could not be sstisfaciorily isolated, although the hydrobromide, which is fairly shble, can be obtained on evapora- t ing its alcoholic solution as a black, glassy iiiacs. They are also produced t o some extent in the insoluble lorm by the action of hot dilute aqueous acid, arid are gradually decomposed i n contact with water, with regeneratioil of the reddish-brown colour of the original dye.

A cetyl Der.ivntive.--The substance was digested with boiling acetic anhydride and a trace of pyridine for three hours, and the solution poured into water. The precipitated product when dry was dissolved in benzene, the solution filtered if iiecessary, and poured into light petroleum. The deep salmon-coloured powder was finally washed with a little ether.

Found : C = 63.54 ; H =4*95. C,4H,80s(CaH30)4 requires jY = 63.36 ; H = 4.95 per cent.

A n acetyl determination by the ethyl acetate method gave C,H40, = 39.67, whereas the above formula requires C2H,0, = 39.60 per cent. These figures are very similar to those found by Cain and Simonsen for acetylsantalin.

It did notl possess a definite melting point, bu t gradually decom- posed without fusion between 250° and 280°, forming a carbon-

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOI), BARWOOD, AND SANDERSWOOD. 129

aceous powder. It was readily soluble in benzene, sparingly so in alcohol, and has notl been obtained as yet in a definitely crystal- liile condition. h molecular weight determination, employing naphthalene as

solvent, gave the following restilt: 0.633 in 13.5 naphthalene gave A4 = - (1.14~. This small depression did not appear t o arise from a separation

of the substance during the crystallisation of the naphthalene. The result is iiileresting, as it approximates to the formula [C,,I-p,,O7(6,H30),I,, M.W. = 2360, which accords with t h e forma- tion of tho potassium salt, C90HS7032R, previously referred to. 'tVhereas the substaiice 011 the one hand may be x p r d e d as itself highly associated, it is quite probable that the association occurs in the solution of naphthalene or that a colloidal soliat'ion is thus pro- duced. Coinplez salts, again, of the character of potassium isosantalin cs?n be obtained from simple substaiiceS, such as gall- acetophenone (T., 1903, 83, 1311, which with aqueous potassiuni acetate f o n m potassium t!.i~~allacetopheiin!ie, C

M.W. == 2344.

Whereas wsantalin is almost entirely precipitated when its concentrated alcoholic solution is poured into ether, a second colour- iiig ma tt'c-1. is present, which in these circiimstances remains dis- solved. The residue obtained hy evaporation of the extract' gave C = 66.64, 1% = 5.61 , CH,- 8.06 per cent., figures indimting in the light of later results -the presence of isosantalin and a third sub- stance of high methoxy-content as impurities. With the object of removing the f oriner, a concentrated alcoholic soliition was poured into n:uch benzene, the resinous precipitate, which separated in siiiall quantity, removed., and the highly fluorescent liquid gradu- d l y cva poraked. Any deposit which quickly separated was neglected, the clear liquid concentrated until bright red flocks separated, aiid tlneas were collected and washed with benzene.

A more rzpid method of isolating this substance, although in R cruder form, was t o pour a concentrated alcoholic extract of the wood into boiling benzene, and to sitbinit the liquid t o fractional evaporation. The product, could be obtained purer by a second t,reatment in the same manner. The bright scarlet, powder obtained lsy khesi:, methods gave CI!L;--8.10 per cent., a figure showing 110 reasonable connexion with isosantaliii, and further purification was evidently necessary. This was effected by treat- ing a cancentrated solution of $he colouring matter in alcohol with

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

130 O'NEILL AND PXEKIN : THE COLOURING MATTERS OF

a little Gyrirobromic acid, and repeat'edly shaking the liyuid with ether, in which the hydrobromide is insoluble." The solution of the latt,er was now poured into water, and in order completely t o remove hydrobromic acid, the mixture was gently warmed with sodium acetate solutioii. Finally, the compound thus isolated was dried on sand, extracted with ether, the liquid diluted with a little benzene, and the red deposit which separated during empora- tioii washed with benzene.

Found: C=67.63; H=5-33; C€I,= i -33 , 7.43.

Such a formula, which represents this compouiid as isosantaliii less one atom of oxygen, but richer in hydrogen, suggests t he replacement of z CO-grouping in the latter by CHI2, but as this for the1 present must remain hypothetical, the compouiid will be tentatively referred t o as deoxyisosantdiiz.

It consists of a scarlet, amorphous powder which has n o t yet been obtained in a Clefinit ely crystalline coudilioii. \Vhen heated, the preparation did iiot sllow aiiy well-defincd riieltiiig poirit, bu t decomposed a t 160 - l G ; j O with the evolution of gas, It is readily soluble in pure ethyl acetate, sparingly so in ether, and is agaiii distinguished from isosaiitaliii by the redder s l ides it gives on dye- ing and the redder toile of its allraliiie aolutions. A fd le r account oE these is given later. Wheii a solution of the substance in absolute alcohol is treated with potassium acetate, n o immedia: c, precipitation of poGsssiuin salt occurs as happens in the case of i$o:;antah, and only when excess of the reagenf, is employed is a gelatiiious deposit formed. The presence of this is n o dcubt ~espousible for the resj nous condition of impure potassiuiii santnlin when this is pre- pared in the preseiice of boiliiig alcoliol. It was acetylated by boiling with acetic aiihydride niid pyridine, and the ljroduct. isolated by means of water, ~ 7 a s dissolved j i i benzene, precipit'atecl by light petroleum, aiid subsequently washed with p u r e ether.

Fouild : C= 64.31, 64.61 ; M - 5 4 0 , 5-29 ; C,II,O, - - 33-67, 39.94. C.,,I3~,0~(C,II~O), I T ~ I I ~ ~ ~ S C -= 64.86 ; 13 = 5.40 ; CJ,M,O,=- 40.63

C,,H,,07 requires G = 67-92 ; H = 5-63 ; CR, = 7-08 per cent.

per cent. A cet!iTdeoxyisosnn tolit7 thus coiitaiiis the same nnmller of acetyl

groups as acetylisosaiitaiiu. bi consists, when dry, of an almost colourless pow d c r, di fTer i I i g cc) I i , i tl PI' ably i TI appw ran ce f roi n ace t'y1- ismantaliit, :tnd is E L ~ S U wore wticlily soliible i l l l ~ e n / ~ ~ i e . \ V ~ P I I I t e a t d , il f'ris~d aiid decolkipo~eil ;it 17'0 1 7 5 ' ) .

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOD, BARWOOD, AND SANDEBSWOOD. 131

0.7485 in 12.13 naphthalene gave A t = - 0.325O.

Acetyldeoxyisosantalin in these circumstances thus shows approximately one-half the molecular weight of acetylisosantalin, a property again distinguishing these two substances.

M.W. = 1324. (C32H32011)2 requires M.W. = 1184.

Although isosantalin closely resembled in its general reactions the santalin of sanderswood, the formula C,,H803(O*C'H3)2 assigned t o the latter by earlier workers nevertheless pointed to a consider- able difference between the two substancs. For purposes of com- parison, an alcoholic solution of saiiderswood was prepared, the colouring matter isolated by means both of lead acetate! and baryta, and the product worked up by identical methods to those previously employed with camwood. As a result, it was found tha t t he phenolic constituents of both woods closely resembled ~ i i e another i n general properties, and two distinct compounds, santalin and deoxysantalin, corresponding with iso- and deoxyiso-santalin, were isolated. The sanlalii? was finally purified by extraction with purified ethyl acetate, and could be snhsequently crystallised from dilute aicohol.

FouI?~: C=65.70; $1-4.90; C€1,=-6.81, 7.09. C,,HI,Q, requires c"= 65.7 ; 11 = 5-11 ; @R,= 5-47 per cent. C,,H,,Q, C = 65.8 ; I€ = 5.02 ; CR,= 6.85 per cent. ,,

Santalin is therefore isomeric with isosantalin, and cannot possess the formula C,,H,,O,. It consisted of a chocolate- coloured powder, very similar iii appearance to isosantalin, readily soluble in h o t inethy1ate:l spirit, although with absolute alcohol i t fesinified and dissolved with difficulty. When heated, it commenced to soften a t 243O, decomposed a t 250-26Qo, and a t 270° had become a honeycombed, carbonaceous mass. Cain and Simon- sen (Zoc. (. i t .) describe saiitalin as melting a t 223--226O, b u t from the exainination of the colouriiig inatter isolated according t o their direction and in ot?her ways, it seemed certain that) santaliiz deconi- poses a t about' the temperature stated above, and does not possess a distinct ineltiiig point. The decorupsition of this sub- stance, Iioweve!~, OVP!II*S st i! ~ 1 1 ~ i ~ 1 1 lowt~r tenipernCurc t h a n in the case of imm,nldi t l . It, was tirst poilitfed rtut ljrp one of ais (T., 1895, '75, 443) tl-ial t hr- t-oi[iiner(li:tl sxntallrl of Merck yielded with ; ~ l ~ ~ o l ~ o l i c poiw+iii III ;!c.t.tate ;I, itiaiuuir-cc,Jo~ 1 d i , which on arialysis gave K :- 6-94, corresponding with the fo~rnu1:t C,,Z1,70,0K. Although in this ease, as was indeed indicated, the reaction was

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

132 O'NEILI, AND PERKIN : THE COLOURING MATTERS OF

only given as a general illustration of the potassium acetate method, this result appeared t o be correct, for not only was it confirmed by Gain and Simonsen, but Brooks again (Philippine J . Sci., 1910, 5, 448) obtained from narrin an allied colouring matter of narra wood, Pterocarpws, spp ., a copper salt, (C15H1305)2Cu. Experi- ments now carried out wit,h pure santalin gave a very different result, the potassium salt prepared by means of boiling potassium acetate giving K=2*45 per cent., a figure very similar t o that yielded by potassium isosantalin.

There was some reason t o suspect that the commercial santalin sample just referred to represented the precipitate derived by acidification of an alkaline extract, of the wood, and as it is known that in this way some alteration of the colouring matter, or at least a portion of it, occurs, it was interesting to examine the potassiuni salt obtained from this crude material. Such a pre- paration, although very thoroughly washed, contained free mineral matter, and the potassium salt derived from it gave K==4*01 per cent. This result, although' higher than that given by pure santalin, is not sufficiently so to explain the analysis of the original product, which mould thus appear to have contained much free mineral matter.

The potassium salt of sanialin may thus be represented, as in the case of the isosantalin compound, as either C72€165(924K or

Santalin in alcoholic solution shows evidence; of the formation of soluble oxonium salts with rniiieral acids, and although i t closely resembles isosnntalin, gives colour reactions which indicate that it is not the same substance.

C96H87032K.

Santalin. isoSantalin. Alcoholic hydrobromic acid ............ Crimson Reddish-violet Diluto sodium hydroxide ............... 11~11 red Dulf violet Alcoholic farric chloride.. ................ Violct Blnish-violet

Again, the solution of isosantaliii in alcohol is much browner than that of saiitalin. Somewhat more marked are the differences in shade given by these colouring matters in similar circumstances, especially when mordants are employed. For dyeing the colouring matter, 0.25 grani in alcoholic solution wts added to the water in the dye-bath, and the dyeiiigs were carried out ( n ) employing moo1 alone, ( h ) employing wool iilorle and snbseyneiitly sadcleiiiiig wit)h bichrome, ( c ) employing wool morclantcd with biihrome and cream of tartar, and (d) employing woo4 inordaiited with bichrome and sulphuric acid. The 37esr2lt:: obtxinclcl w ' e r ~ as follows :

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOD, BARWOOD, AND SANDERSWOOH). 133

( u ) ( 6 ) (fJ) (d ) :3antsliii ... Pale dull red Yull reddish- Yale redtlisli- Pale reddish-

isoSantalin Palc violet-red Dull violet - Violet -red Violet-red, brown pink pink

maroon wea Ire r t 11 an

Acetylsantalin, prepared in the same way as acetylisosantalin,

Found : C2H40, = 39-58, 38.96. C,,H,,O,(C,H,O), requires C,13,02 = 39.6 per cent.

It consisted of a deep salmon-coloured powder, which commenced to decompose a t about 225O, and at' 255-260° swelled u p to form a carbonaceons mass :

(c) .

closely resembled the latter in appearance.

0.7258 in 12.03 naphthalene gave A t =1 -- O . 1 6 5 O . Tho small depression thus obtained is similar t o tha t given by

AZ.lV. = 2558.

o cetylisosantalin, and approximates to the formula (C32H30012).L.

D eo xysan f din.

Sanderswood resembles camwood in containing a second, more readily soluble coloiiring matter, f o r which the' nttmr3 deoxysantalist is suggested. It may be prepared according to the method given for t h e isolation of deoxyisosantalin.

Found : C = 67.59, 68.10 ; H = 5.08, 5-70 ; CH3= 7-21. C",,M,O, requires C = 67.92 ; R = 5.65 ; CB, = 7.08 per cent.

It consists of a bright red powder, the ethereal solution of which possesses a greenish fluorescence, bu t it has n o t yet been obtained in a crystalline condition. Alcoholic potassium acetate gives no immediate precipitate of potassium salt, bu t aftel some time or by means of excess of the reagent a gelatinous deposit separates. A comparison of the colour reaction of deoxysantalin (r l ) and deoxy- isosantalin (6) are given in the following table. These, compared respectively with santalin and isosantdin (we earlier). possess much yellower tones.

Alcoholic solution ........................... Orange-brown Orange Alcoholic hydrobromic acid ............... Scarlet Bright crimson Dilute sodium hydroxide .................. Scarlet Crimson-scarlet

( a ) ( b )

Alcoholic ferric chloride.. ................... Maroon Violet

The dyeing experiments were carried out by the processes which were adopted with santaliii and isosantalin.

Deoxysantalin.. . Red Dull bluish- Dull crimson Red ( a ) ( b ) ( c ) (4

red Deoxyisosantalin Pale reddish- Red-puce Dull reddish- Pale dull red-

violet violet dish-violet F"

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

134 O'NElLL AND PERKIN: THE COLOURING MATTERS OF

As a result, it, was found that although in general character the shades given by saiitalin and deoxysantalin are similar, the latter is by far the better dye, possessing a stronger aud brighter appear- ance. The difference between isosantalin and cleoxyisosaiitalin i n this respect was not so marked, and, if anything, m7as in favour of the former.

Acet?l~deoxysai~tailivL was obtained as an almost colourless amorphous powder, readily soluble in benzene.

Found : C- 64-43 ; H = 5-50 ; C,H402= 39.30. C,,H,,0,(C,H30), requires C =- 64.86 ; p4 - - 4.50 ; C,H,O, = 40*52

per cent.

Deoxysaiitalin is thus isomeric with deosyisosanlalin. The first isolation of a colouring iiiatter from sanderswood was

carried out by Msier (loc. c i t . ) , who employed ether as a solvent, and describes this as red, microscopic prisms melting a t 104-105°, but did not analyse his product. A repetition of Meier's process failed t o give the substance in a crystalline condition, as, indeed, Weyerrnann and Haffely (Zoc. cit.) found, but the product was interesting in that itl evidentdy consisted mainly of deoxysantalin, being bright red and giving no immediate precipitate with alcoholic potassium acetate. This was n o t t o he wandered at, as pure santalin is not soluble in pure ether.

I n the hope of obtaining the, colouring matter in a crystalline form, wool was dyed with sanderswoocl in the usual manner, and attempts were made to extract it by means of solvents from the fibre, but without success. Although the exact rezson for its in- solubility wzs not determined, it seems likely that it is there fixed as ail acid calcium salt, because only by steeping the dyed material in 33 per cent. hydrochloric acid overnight did it then dissolve in the usual way. The material, subsequently well washed and dried, was extracted with ether, and the fluorescent extract! was evaporated t o dryness and the residue washed first wit'h benzene 2nd then with water t o remove acid. I t was again extracted with ether, benzene added to the extract,, and the red flocks which separated on evapora- tion were washed with benzene. The small amount of bright red powder thus isolated had the properties of deoxysantalin, and has hitherto failed to yield crystals. A methoxy-determination gave C:EI,=6*58 per cent'. It thus seems that the dyeing property of sanderswood is due rather t o the deoxysantalin than to santalin itself, and again, although sanderswood is specially rich in colouring mathr , itl is astonishing how Iittle of this comes into play in the dyeing process.

Some proof of this was obtained by dyeing wool with 60 per cent.

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOD, BARWOOD, AND SANDERSWOOD. 135

of sanderswood, and a t the eiid of the operation employing the residual wood twice for dyeing fresh material. The first pattern possessed t,he usual colonr, the second was pale pink, and the third was practically uiidyed, indicating that, all the colouring matter soluble in water had then been removed. The woody matter, how- ever, now yielded to boiling alcohol 13-8 per cent'. of extract com- pared with the 19.4 per cent. originally present, and this consisted of a dark red, resinous mass a i d contained much santalin. Ev011, therefore, if i t be presumed thatl all the soluble matter re- moved from the wood in the dye-bath is colouring matter, i t is evideiit that the larger proportion present therein, owi~ig to its illsoluble nature. remains undissolved and takes no part in this operatioil,

Weidel (loc. c i f .), by rexiloving the colouring matter from sanderswootl with dilute alkali hydroxide, precipitating with acid, and extractiiig the prodtrct, with ether, isolated two crystalline com- pounds, namely, santal, C,-H,O,, and a red substance, C14H120a, very sparingly soluble in the ordinary solvents. Experiments to is01 ate these compounds by Tt'eidel's proem liavc hitherto failed, althoygh, as shown later, they can be obtained frotrm barwood. As in any case they are difficult to prepare, especially by his method, a definite statement as to their non-existence i n sanderswood cannot at present safely he inade.

On the other haiid, saaderswood coiitaiiis, like camwood, a colour- less, phevlolic suhstaiice ilot removed from ether by means of hydro- bromie acid, which does riot appcay t o c r y s t a l h e readily.

A comparison of the colouring matter of this wood with those of sanderswood a i d camwooti appeared to be interesting, but owing to lack of opportunity its examination in this respect could not be carried out a t all completely, and the sxtrset, from only 450 grams of the wood has as yet been employed.

The oiily important work on thiq subject is that of Anderson ( J o u m . Ckenz. Soc., 1876, ii, 582), which, curiously enough, has been ignored in the later literature. He obtained froin the wood colourless crystals, which he termed baphiin, and from this product by means of alkali solution prepared baphic acid, soluble in this reagent, and baphinitin and baphinitone, insoluble, all of which were also crystalline. According to Ryan and Fitzgerald (PTOC. Royal Irish Acad., 1913, 30, 106), baphinitone is identical with hornopterocarpin, and possesses the formula CI7Hl,O,. Anderson also considered the wood to contain three colouring matters, one

P" 2

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

136 O'NEILL AND PERKIN: THE COLOURPNG MATTERS OF

soluble in ether, a secoiiri soluble in alcohol ;tnd iiisoluhle ill etttiJu, and a third which is crystalline.

The alcoholic extraci+< of barwooci wits worked up iii the same manner as that of camwood and sanderswood, and a colouring matter corresponding with the santalin and isosantalifi of the latter was iyolated. After treatment with pure ethyl acetate, it consisted of a chocolate-red powder indistinguishable in appearance from santalin. The analysis of this substance and of its potassium salt gave the following result :

F'OUII~ : C= 65.8 ; 13 = 4.6 ; CI13 = 7.43 ; R = 3-06. On heating, it commenced t o soften at 240°, apparently decom-

posing, and a t 270° had the appearance of a honeycombed, carbon- aceous mass, a behaviour very similar to that given by sant'alin; its cdour reactions, which were violet with alcoholic ferric chloride, crimson with alcoholic hydrobromic acid, and dull red w i t h dilute sodium hydroxidel solution, were the same as those given by this colouring matter. It, thus appears extremely probable that this is identical with santalin.

The ethereal liquid obtaiirecl during the isolation of this sub- stance;, which corresponded with that termed B in connexion with camwood, was evaporated to dryness, the residue again dissolved in a little alcohol, and poured into ether. The solution decanted from a tracel of resinous precipitate was washed once with water and treated drop by drop wi th fuming hydrobromic acid until all the colouring matter had been removed from t'he ether as hydro- bromide (C). After washing with water, the ether, on evaporation to dryness, gave an almost colourless, crystalline residue. This was digested with benzene to remove a trace of a substance resembling pterocarpin, -f dissolved in hot alcohol, and the solution diluted with a little water. The large leaflets which separated 011

cooling and became opaque on drying, owing to loss of water of crystallisation, were recrystallised from abeolute alcohol. The yield was 0.7 gram.

Fouud : C--G3*95; 51-4-20; GH3=5.04. C,,H,,O, requires C = 64.0 ; H == 4.0 ; CX13 = 5.0 per cent.

This compound appears to be the santal described by Vl'eidel (loc. c i t . ) as existing in sanderswood; he assigned t o it the formula C,H,O,,~H,O, and obtained from it a dibromo-compound,

* Thc residue obtained by evaporating tho alcoholic extract to dryness on solution in 75 per cent. alcohol gives gradually a considerable deposit of this product.

f The complete removal of this substance from the colouring matter by either the lead acetate or bsryta method of precipitation is djficult to effect.

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOI), BARWOOD, AND SANDXRSWOOD. 137

C,EI,0,Br2. Weidel, however, did not ascertain tho presence of a methoxy-group, and its analysis in this respect indicates that it possesses the formula C,,H,0,(O*C13c3). Ssntal melts a t 222--223O, is readily soluble in dilute alkali hydroxide, and sparingly so in absolute slcohol, from which it crystallises in thin plates or flat needles. With alcoholic lead acetate, it gives a colourless pre- cipitate, and with alcoholic ferric chloride a violet-black colora- tion, although this, according t o Weidel, is dark red. On acetyl- ation, it gives a well-defined, crystalline acetyl derivative, and seems to be very similar t o Anderson’s baphic acid (loc. cit.). The hydriodic acid liquid from the Zeisel determination contained crystals, and these, after removal of the acid, separated from dilute alcohol in sniall, colourless, flat needles melting atl 270-273*, apparently with decomposition. This iiew compound, for which t8he name santol is proposed, would appear t o possess the formula C15€I~,,0~, and this it is hoped t o confirm a t a later period. I t s solution in dilute alkali hydroxide, at first) almost colourless, rapidly develops a reddish-violet tint, and the liquid on acidification deposits yellow crystals. If the acid mixture is boiled, these become colour- less, apparently with regeiieration of santol, for when collected and washed, the product’ can again be made to produce the same changes, Santol thus appears to contain a lactone group. With lead acetate, it gives a colourless precipitate, and with alcoholic ferric chloride a brownish-black coloration.

The hydrobromic acid liquid (C) obtained dinring the isolation of santal, decanted from any insoluble matter, was diluted with water, the precipitate repeatedly treated with water to decompose the hydrobromide, and the product, when dry, dissolved in a little ethyl acetate, the solution being allowed to evaporate spontaneously. After several days, minute specks were present in the viscous fluid, and this was now stirred with a mixture of pure ether and ethyl acetate, the insoluble product being collected and washed with ethyl acetate I n order t o remove santal, which was found to be present, it was digested at the boiling point for one hour with absolute alcohol, and the glistening iiiass of bright red crystals (0.4 gram) analysed.

This had the reactions of the compound Cl,H120, (@=68.8; H=4.9) described by Weidel as existing in sanderswood, but prob- ably contained still a little santal, and this seemed certain by a crystallisation of the substance from alcohol, when the 0.0756 gram thus obtained gave the higher figures, C=:G7*56; H-4.84 per cent. A further purification has not been possible, but there can be little, doubt that this coinpound still retained santal, and that when COE-

pletely freed from this, Weidel’s formula, C,,IT,,O,, represents its

(Found: C=67.3; H=4.8; CH,=9.77.)

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

138 O’NEILL AND PERKIN : THE COLOURING MATTERS OF

composition. On t’he other hand, the result of the Zeisel cleter- mination indicahs that Weidel’s formula can scarcely be correct, and that the true formula, based on his figures, would appear t o lw C,,H,,0,JO*CH3), or C2,H,,0,(0*@H3),, the former of which rcquii-es C = 69.1 ; H = 5.06 ; CH,= 10.3 per cent. I n case this is correct, this compound, for which the name santalorne is proposed, has the com- position of deoxyscmtn7in nzonohtethtyl ether, and, indeed, its general properties are in harmony with this suggestioii.

Santalone is sparingly Foluble in most solvents, and crystallises from alcohol in small, glistening red leaflets. It readily dissolves in alcoholic hydrobromic acid with a bright scarlet, in dilute alkalis with a fairly permanent red, whereas alcoholic ferric chloride pro- duces a violet colour. Alcoholic potassium acetate does not appear t o cause the separation of a potassium salt When heated, the purest preparation obtained up t o the present d;jdcenecl ahout 280° and melted at 300°.

The hydriodic acid liquid fr(Jlr1 the Zeisel determination coiitaine. I a blackish-green, nietallic-looking resin, which was soluble in alkali hydroxide with a violet colour, rapidly becoming brown in the air and closely resemhling the proclu.ct given in the same way by cleoxysaiitalin.

Several years ago 1-Ierz and Peykin (Thorpe’s I‘ Dictiunary of Applied Chemistry,” 1913, Vol. IV. , G29) isolated both these compounds from barwood by Weidel’s method in very small amount, but in order to obtain any quantity of this rare material, the process of extracting lttrge quantities of resitlous colouring matter with ether for several days proved impracticable. On the other hand, the operation was not a l ~ a y s successful, and this appeared to depend on the alteration of the main bulk of the coiistituents of the wood soluble in ether, other than santal and santalone, during the alkaline ex1 raction into resinous products insoluble in this medium.

The mother liquors from which the santaloiie had I1ee11 isolated contained a c onsi cl eralol e a mount of insoluble col our in g ni at t er resembling deoxysantaljii, which, however, was not closely examined.

The results of this iiiveskigatjon indicate that t h e inole insoluble colouring matter of camwood, here tlerrned isostrutaTi?7, is most simply expressed as C2,H,,Q,(OMe),. It is isomeric with the santalin of sanderswoocl, and for this the same formula is therefore proposed in place of C,4H,104(01CIe) suggested by Cain and

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

CAMWOOD, BARWOOD, AND SANDERSWOOD. 139

Simoriveri . Although santlalin from sanderswood appears to be identical with the colouring matter which can be similarly isolated from barwood, the isosantalin of camwood differs from both these in that it possesses not only a much higher decomposition point!, but gives distinctly bluer shades on rnordanted woollen cloth. isosantalin and santalin are respectively accompanied in the dye- wood by the more readily soluble colouring matters deosyisosantalin ancl deorysantalin, which again are isomeric, and for which the formula C&J3,605(OMe)2 or C,,PI,,0,(OMe)2 is proposed. The con- nexion between the santalins and deoxysantalins is a t present un- certain, and the acetyl determinations do not favour the view which at first suggested itself, that the former differ from the latter in possessing an ad ditioiial hytlroxyl group. There is some proba- bility that the deouy-coinpoiincis are converted by oxidation in alkaline solution inlo ths sap? taliiis, but expel-irrients on this point itncl on their behaviour towards 21-benzoquinotte in alcoholic solu- tion have €or the present been laid aside. Noiie of' these compounds has as yet been obtained in a definitely crystalline condition, ancl the opinion is expressed, a t least as regards santalin, that, if this is to be effected sonie special process, perhaps of the character by which crystalliiie Jixrnatein is prepared from logwood, will have to be devised Meier's saiitalin (loc. c i f . ) , which it is suggested is rather deoxysantalin than Cain and Fh~onsen's saiitalin, was described by this author as crystalline, and i t is quite possible that both deosy- and deoxy iao-santalin, here obtained, contain still soine iinpusity which renders crystallisation more difficult. The potassium salts of iwsnntaliiz and santalin possess the formula Ci2H650211C or C9,€-I,,0,,K, and this fact, coupled with the exceed- ingly small depressioii of the freezing point given by their acetyl compounds in nsphthalma, indicates that either these colouring matters possess a very high molecular weight or that this depression is abnormal. The effect of other solvents remains to be investi- gated, bnt as Cairi and Simonsen, employing phenol, obtained with santalin a iriolecdar weight of 257, the latter suggestion would a,ppear t o be correct.

Santal and Weidel's substance, CL41112Q4 ( ?>, nosh' termed santalone, have been found t o exist, in barwood, and the formuh considered applicable to these conipouncls are respectively Cl,H,0,(811Ze) rind Cp,R, jO,(OMe),, the latter being possibly a t l eoxysnn fu l i i~ nLonomPth!/l e ther . These compouiids will be further examined when opportunity arises, and it' is poesible that by their study some clue t o the constitution of both santalin and ismantaliti will resnlt ,

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online

140 SENTER AND TUCKER:

N o t e o n S t . John’s Wort,

According t o Keegan (Chem. iVews, 1915, 111, 290), St. John’s wort, Hypericum perforaturn, contains much flavone, which seemed to be similar to gossypetin. Desiring a further supply of this colouring matter, previously oiily known to be present i n quantity in cotton flowers, the yellow colouriiig mat’ter of St. John’s wort was submitted t o examination. A concentrated alcoholic extract of the dried material was poured into water, t he mixture shaken with ether to remove wax, etc., and the clear liquid digested a t the boiling point with a little sulphuric acid. When cold, the pre- cipitate was collected, and in this condition dissolved in alkali with a green shade somewhat similar to tha t given by gossypetin. -After removal of impurity by solution in ether, the yellow product was crystallised froin dilute alcohol, and on treatment with alkali hydr- oxide now gave a pure yellow tint. The acetyl compound melted a t 194-196O, and when mixed with acetyl quercetin did not depress the melting pointl of the latter. The benzoyl compound melted a t 185-186°, and this was identical with tha t of pentabenzoyl quercetin obtained by means of benzoic anhydride. The colouring matter of St. J~hii’~ wort consists, therefore, of quercetin, and no gossypetin is present’.

CLOTHWQRKERY’ RESE-YRCli LABORATORY, DYETNG DEPARTMENT,

THIG UNIVERS~TY, LEEDS. [Received, January 3rd, 191 5.3

Publ

ishe

d on

01

Janu

ary

1918

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 25

/10/

2014

14:

45:5

8.

View Article Online