·Enantiomers, Racemates,
`• and Resolutions/
`
`Andre Collet
`Centre National de Ia Recherche Scientifique
`• College de France
`Paris
`
`Samuel H. Wilen
`Tlze City University of New York
`The City College
`. New York, NY
`
`A WILEY-INTERSCIENCE PUBLICATION
`
`JOHN WILEY & SONS
`
`New York · Chichester· Brisbane ·Toronto
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 1
`
`

`

`-~ ·,.-. :' ·:· .,
`
`(
`
`·~:·
`
`\
`
`1981
`
`~ /"'· \
`G
`
`I
`
`/
`·•
`
`Copyright© 1981 by John Wiley & Sons, Inc.
`
`All rights reserved. Published simultaneously in Canada.
`
`Reproduction or translation of any part of this work
`beyond that permitted by Sections 107 or 108 of the
`1976 United States Copyright Act without the permission
`of the copyright owner is unlawful. Requests for
`permission or further information should be addressed to
`the Permissions Department, John Wiley & Sons, Inc.
`
`Library of Congress Cataloging in Publication Data:
`Jacques, Jean.
`Enantiomers, racemates, and resolutions.
`
`"A Wiley-lnterscience publication."
`Includes index.
`2. Racemization.
`I. Stereochemistry.
`I. Collet, Andre,
`3. Chirality.
`4. Isomerism.
`joint author.
`II. Wilen, Samuel H., joint author.
`Ill. Title
`
`541.2'23
`QD48l.J26
`ISBN 0-4 71-08058-6
`
`81-1604
`AACRl
`
`Printed in the United States of America
`
`10
`
`9
`
`8
`
`7
`
`6
`
`5
`
`4
`
`3
`
`2
`
`I
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 2
`
`

`

`250
`
`Resolution by Direct Crystallization
`
`[tis known that certain racemates may be resolved by inclusion in an enantio·
`to Mislow,
`the above results
`morphous crystal (Section 5.1.8). According
`correspond to the inverse of this type or resolution since racemic 7 constitutes the
`host molecules which are resolved by inclusion of optically active guest molecules
`in their lattice.
`
`REFERENCES 4.4
`
`J. H. Van't Hoff, Die Lagenmg der Atome im Raume, Vieweg, Braunschweig, 2nd ed.,
`1894, p. 30; 3rd cd., 1908, p. 8.
`2 K. Amaya, Bull. Chern. Soc. Jpn.,I96I, 34, 1803.
`3 H. Q, Jones, Proc. Cambridge Phi/os. Soc., 1907, 14, 27; quoted by L. Ebert and G.
`KortUm (ref. 5); see also E. SchrOer, Ber,, 1932,65, 966.
`4 H. Goldschmidt and M. C. Cooper, Z. Phys. Chern., 1898, 26, 711.
`5 L. Ebert and G. KortUm, Her., 1931, 64, 342.
`6 B. Bomich and D. W. Watts,J. Am. Chem. Soc.,1968, 90,6228.
`7 K. Mizumachi,J. Coord. Chern., 1973,3, 191.
`B M. Yamamoto andY. Yamamoto,/norg. Nuclear Chern. Lett., 1915, II, 833.
`9 F. S. Kipping and W. J. Pope, Proc. Chern. Soc. London, 1897-1906, 113.1. 01em. Soc.,
`1898, 73, 606.
`10 L Ostromisslenskii, Ber., 1908, 41, 3035.
`II A. LUttringhaus and D. Bcrrcr, Tetrahedron Lett., 1959, I 0.
`12 M. B- Groen, IL Schadcnbcrg, and H. Wynberg, J. Org. Chem., 1971, 36, 2797.
`13 D. H. R. Barton and G. W. Kirby,J. Chern. Soc., 1962, 806.
`(a) L. Addadi and M. Lahav, J. Am. 01em. Soc., 1978, 100, 2831. (b) ibid., 1979, 101,
`14
`2152. (c) Pure Appl. Own., 1979, Sl, 1269.
`J. van Mil, E. Gati, L. Addadi, and M. Lahav, submitted for public;.~tion 1 1981). We are
`greatly indebted to Professor M. Lahav for permitting U'l to cite this work prior !o publi·
`cation.
`J. L. Purvis, U.S. Patent 2,790,001 (1957); 01em. Abstr., 1957,51, 139lla.
`16
`11 B.S. Green and L. Heller, Science, 1914, 185, 525.
`18 K. S. Hayes, W. D. Hounshell, P. Finocchiaro, and K. Mislow, 1. Am. O~em. Soc., 1977,
`99,4152.
`
`IS
`
`CHAPTER 5
`Formation and Separation
`of Diastereomers
`
`In the preceding chapter we have principally examined methods for the resolution
`of enantiomers that do not require the intervention of chiral agents. We shall now
`examine those processes which depend on the fonnation of diastereomeric com·
`pounds derived from the enantiomcrs to be separated. Unlike enantiomers, dia·
`stereomer pairs may have significantly different physical properties which may be
`the basis of their separation from one another. We consider, in particular, crystal·
`line diastereomeric compounds. We examine two broad categories in succession:
`dissociable compounds, or complexes, and covalent compounds. Tills classification
`is convenient even if somewhat arbitrary.
`The number or resolutions mediated by diastereomers described in the litera·
`ture is quite large, and we have not felt it necessary to cite all examples known to
`us, The cases cited are representative and cover the principal resolving agents and
`functional groups.
`While covalent diastereomers are increasingly separated by chromatography,
`the separation of other types of diastereomeric substance depends entirely on
`crystallization techniques that are based upon differences in solubility. Thus, in this
`chapter we apply several or general concepts developed earlier and, in particular,
`those involving the use or phase diagrams.
`Before taking up these matters, let us brieny examine the nontrivial matter of
`diastereomer specification and of the way in which the different mixtures derived
`from diastereomers may be distinguished.
`The bimolecular combination of two cltiral substances A and B may lead to
`four diastereomers (Scheme I). We have adopted the terminology in which the
`letter p is used to designate the Uiastereomers resulting from reaction of the two
`constituents having like sign of rotation and the letter n to designate the diastereo(cid:173)
`mers formed from constituents of unlike sign. This convention stems from a
`suggestion made by [. Ugi (Z. Naturforsch .• 1965, 20b, 405) for covalent corn·
`pounds possessing but two chiral centers and which is based on the nomenclature of
`Calm, Ingold, and Prelog: RR = SS = p and RS = SR = n. In our convention,
`which is applicable to all types of dissociable as well as to covalent diastercomers,
`no account is taken of the absolute configurations of the chiral centers, however.
`
`251
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 3
`
`

`

`252
`
`Formation and Separation of Diastereomers
`
`5.1 Dissociable Compounds and Complexes
`
`253
`
`·~The p and IJ designations take into account only the signs of the rotations of species
`A and B. • In the case of diastereomeric salts, for example, reference may be made
`to Tables I and 2 in Section 5.1, which give the signs of the rotatory power of the
`principal alkaloids and the naturally occurring acids used in resolutions.
`
`SCHEME I
`
`5.1 DISSOCIABLE COMPOUNDS AND COMPLEXES
`
`The most widely used resolution method remains the formation and separation of
`crystalline diastereomeric salts between racemic substrates and optically active
`resolving agents. Other usable dissociable crystalline combinations do exist, none(cid:173)
`theless; these are Lewis acid-base complexes, inclusion compounds. and quasi·
`racemates. These diverse addition compounds have in common their ease of usage.
`They are obtained generally by simple mixing of the constituents in an appropriate
`solvent. Regeneration of the constituents is most often immediate and the resolving
`agent is almost always recovered in a form that allows its reuse.
`The resolution method consisting of the formation of a salt between a racemic
`acid and an optically active base was discovered by Pasteur 1
`3 in 1853 1
`-
`:
`
`I have shown that the absolute identity of the physical and chemical proper(cid:173)
`ties of nonsuperposable right and left bodies ceased to exist when these sub(cid:173)
`stances were put in the presence of [optically 1 active bodies. Thus, the right
`and left tartrates of the same [optically I active organic base are entirely
`distinct in their crystalline forms, in their solubility, etc ... ; it was thus to be
`hoped that one could take advantage of this difference to isolate the two tar(cid:173)
`taric acids whiCh comprise the racemate: after much fruitless research
`attempted on various bases, this is the service done by the two bases quinicine
`and cinchonicine. When, for example, one prepares the racemate of cinchoni(cid:173)
`cine (i.e., in modern terms the cinchotoxine salt of racemic tartaric acid],
`then for a given concentration of the solution it is always the case that the
`first crystallization consists for the most part of left tartrate of cinchonicine;
`the right tartrate remains in the mother liquor. A similar result is obtained
`with quinicine; however, in this case it is the right tartrate which deposits at
`first. Thus, when a binary composition analogous to that of racemic acid be
`suspected, its resolution should be attempted by placing it in the presence of
`an (optically) active product which, as a consequence of the necessary dis(cid:173)
`similarity of the properties of the combinations which it will be possible to
`make from the components of the complex group, will allow the separation
`of the latter.
`
`The process may be summarized by Scheme 1, which corresponds to the treatment
`of a racemic acid dlAH with an optically active base to form saltsp and nor to its
`inverse (racemic base and active acid).
`
`SCHEME I
`dB
`d /AH - {dA-,dBH+}+ {IA-,dBH+}
`n salt
`p salt
`
`When they are prepared separately from previously resolved components, salts
`n and p have different crystalline fonns and frequently also different degrees of
`solvation. The possibility of separating such diastereomeric salts when they are
`allowed to crystaJlize from a mixture of a racemate and an optically active resolving
`agent presupposes the occurrence of a number of conditions all present together
`which we examine in the following sections: salts p and n, or at least one of these,
`
`CD dAdB = P+ (or P-)
`CD
`/AlB = P- (or p.)
`
`n+ (or n_)
`
`n_ (or n+)
`
`Q) dAJB
`0 /AdB
`Given the above, the four diastereomers [AB] are found to consist of two
`enantiomeric p compounds, P+ and P- and, by the same token, two enantiomeric n
`compounds, n+ and n_. It should be evident that the sign of rotation of a given
`diastereomer p or n is not necessarily related directly to those of its constituents A
`and B. In Scheme 1, the diastereomers dAdB and dAIB have been arbitrarily desig·
`nated asP+ and n+; they could just as well (experimentaUy) have been found to be
`P- and n_, or even p_ and n+> for example.
`In a resolution, which brings into play a racemic substrate and a resolving
`agent which is by definition a single enantiomer, the formation of diastereomers
`leads to a mixture of only two compounds: p and n. It is important to observe, as
`Scheme 2 makes clear, that the mixtures derived from racemic A and optically
`active B (case <D) are not identical to those derived from the inverse operation,
`namely. racemic Band optically active A (case (l)).ln one case, the diastcreomcrsp
`and n have the same sign, while in the other they have unlike signs.
`
`SCHEME 2
`
`CD d/A (
`
`dB
`
`dAdB +/A dB
`
`(e.g., P+> n_)
`
`IB
`
`dAIB +/AlB
`
`(n+. P-)
`
`dA
`
`dBdA+IBdA
`
`(p+. n+)
`
`lA
`
`dB/A + /BIA
`
`(n_, p_)
`
`CD d/B
`
`(
`
`The consequence of this lack of symmetry between the two cases is taken up in
`Section 5.1.16. In the sections which follow, we generally deal with mixtures of
`diastereomeric pairs (p,n) without need of further specification of their sign of
`rotation.
`
`• In those cases- fortunately relatively rare- in which the s~n inverts upon a change in
`solvent, it is necessary to stipulate the solvent used (preferably that solvent in which the salts
`are best formed).
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 4
`
`

`

`254
`
`Fonnation and Separation or Diastereomers
`must be crystalline; their solubilities must differ; they must not cocrystallize (form
`solid solutions); and they must not form double salts (addition compounds fp, 11] ).
`While any optically active acid or base is in principle usable, its use as a resolv(cid:173)
`ing agent is limited by its availability and cost.
`The naturally occurring alkaloids, which were virtually exclusively utilized for
`about 100 years, are still much utilized in the resolution or adds. They are
`gradually being displaced by synthetic bases and by derivatives of natural products.
`The totally synthetic bases are, most often, primary amines and consequently are
`stronger bases than the common alkaloidal resolving agents, which are all tertiary
`amines and this may in some instances facilitate salt formation. On the other hand,
`synthetic bases all have the disadvantage that they themselves need to be resolved.
`Titis resolution, however, furnishes both enantiomers, which are thus available for
`use as resolving agents, quite unlike what obtains with alkaloids.*
`The use of diastereomeric salts in resolutions of acids or bases in preference to
`that of covalent diastereomers is traditional. It stems from the simplicity with
`which diastereomeric salts are formed and from the ease of their cleavage to
`resolution substrates.
`The frequency of use of the several basic resolving agents is quite unequal. For
`some 230 cases of resolutions of acids described in the literature between 1960 and
`1970,4 about one third were carried out with brucine and quinine. During the same
`interval, tartaric acid and its derivatives accounted for about half of all resolving
`agents used in the resolution of bases.
`The cost of a resolving agent is also of some interest, although it is clearly not
`an independent variable since one does not necessarily have the option of choice.
`Brucine, cinchonidine, cinchonine, stryclmine, dehydroabietylamine, (+)-and(-)(cid:173)
`(+)-
`(+)(cid:173)
`ephedrine,
`(-)-2-amino-1-butanol,
`and
`(-)-<r-methylbenzylamine,
`amphetamine, and (+)-deoxyephedrine are the least expensive bases available com(cid:173)
`mercially; each cost less than $100 per mole in 1979.
`Comparable least expensive resolving aciUs commercially available are: (+)(cid:173)
`camphor-10-sulfonic
`acid,
`(+)-camphoric
`acid,
`(-}dibenzoyltartaric acid,
`diacetoneketogulonic acid, (+ )- and (-)-mandelic acid, (-}-malic acid, and (+)-and
`(-}tartaric acid. Each of these cost less than $.90 per mole in 1979.
`Tables 1 and 2 list the principal bases and acids used as resolving agents
`through salt formation. Virtually all are commercially available. The leading firms
`furnishing organic compounds for laboratory use also supply most of the resolving
`agents on a relatively small scale and at relatively high prices. There are also
`suppliers more or less specialized in certain types of compounds (e.g., alkaloids,
`camphor derivatives, or synthetic amines for pharmaceutical use) who may furnish
`kilogram quantities of some resolving agents at lower prices.
`The practical aspects of the use of resolving agents, which are only briefly given
`in the following sections, are taken up in Chapter 7. The purification of resolving
`agents and the cleavage of diastereomeric salts are discussed in Section 7.4.
`
`•Quinine and quinidine, on one hand, and cinchonidine and cinchonine, on the other, may be
`considered as pairs of quasi-enantiomers. Their use in the resolution of acids so as to lead to
`both substrate enantiomers is discussed in Section 5.3.2.
`
`•
`•
`•
`
`•
`
`•
`•
`'
`
`;
`
`•
`
`'
`
`'
`
`11-
`3
`" .!!.
`
`-§, ;
`0
`"'
`" 0 • §
`"'
`
`"-
`0
`
`" E • z
`
`0 z
`" 0
`E
`0 "'
`
`" .~ ;;;
`E
`.£
`~
`" ·;:
`
`~ " l\ •
`'" " :E
`0 E
`~ "
`~
`~
`" "' -;;
`" d:
`
`0.
`'()
`
`" :;; ,::
`
`·;;
`~ ·;:
`.!2
`~
`0
`0 .;;
`_...._ .E:
`
`::;. :: :: :: ~ i" = ---
`
`l: -(cid:173)
`uuuuaa~o~=oo~~~~--~-­
`··..,... oo--(cid:173)
`.... :c:c:c
`<»O
`Cu~;...t.;~;:;:l ~:E tl) . : ; ; . - -.. ~- .. ~ ~ ~
`~~~~-~~~~~ ~ ~l:~e~
`~m~omm•N- ~-~----­
`NM-MONOM•~~-N~•oNm
`---N-N--9~-MNN-·--
`il+l++i++<+•+•+•fl+•+•+l~l
`
`9W.WWW·M~NNNNMNNNN
`~~~~~~~~~~~~~ci~~~~
`MMMMNNMNN---NN _ __ _
`m~NNmm~••~•M-•~··~~
`
`6q·6
`a:o.~c:i>=~:q,q,c[o
`0
`~~~~~~~~Z~~z~~zz~~
`££-"-''~~'££;,;'; --,;: , '£ , =,;..:(cid:173)
`;;-~-!?, _"!:::.::. .• ~ -~:;:<> ::;:::::::,..:;:"'
`:-:
`,.!-';!
`vvuvuuuuuuuuuuvuuv
`I
`+
`0
`~
`0
`7 [
`
`-
`
`'
`
`0
`
`~~ +M
`~-i-­
`::a~J..
`<:l ==,.....::
`§ J:l ~
`g- Q.(cid:173)
`Q. e.£
`,..; ·a -g
`::;:~ 3. IT
`>._-;' e:
`-;:--::
`§09-l++
`..::.51"'1..~~~
`.-.
`§-E":".-::~~
`..!.--.~~~-!.·g·g
`i:!
`.-::.!... 7,s ~ S.s.s
`·- ..... + As .. c..·e~~
`E+~+·E· -5-5
`~ --:~--606~~~
`1 •• ]IIJt~tlll
`•
`~:E.E.S ~ g g-g_:-l<~<::!-5_-g_
`.5
`III~§§IE.!J~~u~~~~·~
`·-~~---~~1-~····~-
`2~33.51Bt~BE~~~,;Q
`~~oouu~~o~o<~~~doo
`
`_,.....f'fl .... <o'(lCI'o~= f'fl..,.ll')-o.-..0::10-N
`
`__ ......... __ ,.....,.....,.....
`
`255
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 5
`
`

`

`CHAPTER 7
`Experimental Aspects and
`Art of Resolutions
`
`The information and concepts· found in the preceding chapters on the nature and
`properties of enantiomers, diastereomers, and their mixtures should permit a better
`understanding of, and ability to carry out, resolutions. The practical details about
`how one actually begins the resolution and how one controls and monitors the pro·
`gress and carries the fmal purification to a successful conclusion are given in this
`chapter. This should nonnally suffice to achieve success, as would be true in any
`other unit process associated with the practice of organic chemistry.
`Yet many practitioners believe that there is a mystique and an aura of "art"
`to resolutions. While we do subscribe to the proposition that such techniques as
`recrystallization require some practice to be carried out well, we believe, and think
`organic chemists would generally agree, that virtually anyone can learn to carry out
`a .Purification implicit in a recrystallization. So, too, with resolutions.
`We hope that this chapter will complete the demystification. In our experience,
`there need be few - if any - failures in intelligently and systematically executed
`resolutions.
`
`7 .I CHOICE OF RESOLUTION METHOD
`
`Even if it is difficult to predict with certainty how the resolution of a racemate
`should be undertaken in order to achieve success efficiently, a number of elements
`must be taken into consideration quite early in the design: (a) the quantity and (b)
`the structure of the compound to be resolved.
`
`7 .1.1 Choice of Method as a Function of Scale
`
`It must be evident that on a scale smaller than lOOmg, a new resolution carried to
`completion using crystallization techniques may be achieved, but rarely without
`difficulty. Chromatographic resolution tecfmiques are more suitable here. The
`choice to be made is that of a covalent diastereomeric derivative from which
`regeneration of the enantiomers is easy and attended by little risk of racemization.
`378
`
`7.1 Choice of Resolution Method
`
`379
`
`With the chromatographic techniques presently available, few mixtures will with(cid:173)
`stand separation trials carried out with perserverance.
`For amounts ranging from one to hundreds of grams, cluomatography is less
`useful in resolutions, although. at the lower end of the scale, techniques such as
`flash chromatography 1 are applicable in many cases. We remain doubtful of the
`claim made by Pirkle to the effect that "most 'first time' resolutions of enantiomers
`will soon be effected almost solely by liquid chromatographic techniques." 2 Many
`laboratories are neither equipped with the necessary facilities for carrying out
`preparative liquid chromatography nor do they have the experience required to
`quickly achieve such separations successfully. The investment of time required for
`a satisfactory chromatographic resolution on a modest scale would then be such as
`to lead many researchers to undertake traditional resolutions involving crystallization
`first.
`On the intermediate scale indicated, separation of dissociable diastereomeric
`compounds by crystallization would seem, for the time being, to offer the best
`chance of success and convenience. On this scale, the choice of resolving agent is
`still either influenced by their cost or by the need to eventually recover them.
`for quantities of substance which range from a kilogram to tons, use of
`diastereomeric salts in resolutions becomes in turn less and less attractive as a
`consequence of the relatively high cost of resolving agents (bases in particular).
`Only resolution through direct crystallization of enantiomers is likely to answer to
`industrial requirements.• his necessary, however, Utat the racemate employed (or
`one of its derivatives) be a conglomerate.
`
`7 .1.2 Choice of Method According to the Structure of the Substrate
`
`It is clear that the structure of the substrate has at least as much importance in the
`choice of the resolution method as does the scale.
`Use of covalent derivatives in the resolution of chiral acids and bases is relatively
`rare. Salts are easy to prepare, and their components are easy to regenerate on
`virtually any scale. In contrast, resolution with covalent diastereomers is accompa·
`nied by additional risks (racemization, low yields) in connection with their
`formation and decomposition. Such derivatives are preferred in the resolution of
`alcohols, phenols, ketones, and so on, where salts cannot be directly formed. More(cid:173)
`over, an advantage of covalent Jiastereomers is that they may frequently be sepa(cid:173)
`rated either by chromatography or by crystallization.
`Other structural peculiarities also figure in the choice of resolution method.
`In the case of enolizable ketones a detour is sometimes advantageous: they may be
`optically activated by resolution of the corresponding alcohol.
`
`• In some well-defined ca.~es, such as the amino acids, fermentation processes and even asym(cid:173)
`metric synthesis are applicable on an industrial scale.
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 6
`
`

`

`380
`
`REFERENCES7.1
`
`W. C. Still, M. Kahn, and A. Mitra,J. Org. Chern., 1978,43,2923.
`2 W. H. Pirkle and J. R. Hauske,J. Org. Chem., 1977, 42, 1839.
`
`7.2 OBTAINING CRYSTALLINE DIASTEREOMERS
`
`Once it has been detennined that a resolution requires the formation of dia(cid:173)
`stereomers, there still remains the matter of choosing a resolving agent from among
`the many available. For covalent diastereomers, the choices are quite varied, and it
`is difficult to give more specific advice than that which may be gleaned from
`Section 5.2.
`When a suitable derivative has been prepared, the separation (by chromatography
`or by crystallization) with which one is confronted is fundamentally no different
`from that of most mixtures ordinarily encountered in the practice of organic
`chemistry. It is otherwise if the projected resolution requires the crystallization
`of diastereomeric salts.
`Finding the .. best" diastereomeric derivative constitutes the first step and is
`undoubtedly the most difficult to rationalize in a resolution.
`
`7 .2.1 Systematic Trials. Choice of Resolving Agent
`
`The search for a suitable resolving agent in separations mediated by diastereomeric
`salts may be carried out in a systematic manner ,1 as a result of the ease of formation
`and decomposition of this type of diastereomer. Moreover, the crystallization and
`separation of iliastereomcric salts of a wide range of acidic and basic resolving
`agents are operationally quite similar. We do not consider it to be a good strategy
`to begin a resolution by combining a relatively large quantity of racemate, even as
`little as I g, with a single resolving agent chosen more or less at random or on the
`basis of its immediate availability and to follow this with the persistent and
`repeated recrystallization of the salt which has hopefully formed. While this way of
`proceeding may occasionally lead to the expected result, it is statistically much
`more likely to result \n a product that is optically impure or even completely
`racemic.
`Taking advantage of the experience of several groups of specialists in this type
`of separation, notably that of Fredga and his collaborators (these being perhaps
`responsible for the largest number of published resolutions of acids), it is possible
`to systematize the process that leads most rapidly to the selection of usable
`diastereomer salts.
`It is essential that laboratories where resolutions are carried out have available
`at least a half·dozen optically active bases and as many acids with which systematic
`trials with small quantities of substrate may be undertaken. 1l1c process comprises
`three steps:
`
`Experimental Aspects and Art of Resolutionli
`
`7.2 Obtaining Crystalline Diastereomers
`
`381
`
`Table 1 Measurement of rotations of small samplesa
`
`Case a
`Molecular weight 150: (a] 0 = + 20°
`Substrate taken: 150 mg (0.001 mol)
`Substrate recovered from the diastcrcomer mixture: 30mg
`(40% of 0.0005 mol maximum)
`Rotation observed: <>o = + 0.300° (c = 0.030 g/2.0 mL = 1.5 g/1 00 mL) if
`optically pure
`Case b
`Molecular weight 450; [o:J 0 = + 20°
`Substrate taken: 45 mg (0.0001 mol)
`Substrate recovered: 9 mg
`Rotation observed: a 0 = + 0.090° (c = 0.009 g/2.0 mL) if optically pure
`
`0 S. H. Wilen, A. Collet, and J. Jacques, Tetrahedron, 1977,33,2725. Reproduced
`hy permission of Pergamon Pmss Ltd.
`
`Combining 10-~ to 10-3 mole of the racemic substrate with an equivalent of
`optically active base or acid in the presence of a quantity of solvent of the
`order of I mL (see Section 7.2.2). The same operation is simultaneously
`carried out with several other resolving agents.
`2 When a crystalline salt is obtained (see Section 7 .2.3), it is isolated, dried, and
`weighed. In order for the test to be significant, the quantity of crystals obtained
`must obviously not exceed approximately one half of the total weight of the
`two diastereomers that may be formed; otherwise, one must use more solvent.
`3 After having decomposed the salt, the liberated substrate is isolated to permit
`the measurement of its specific rotation.
`
`The possibility of working with such a small amount of substance (of the order
`of 100 mg substrate) is tied to the sensitivity of modern photoelectric polarimeters
`(minimum ±0.005°) which is quite adequate for the rotations typically observed
`during the course of resolution trials (Table 1 ). Since the specific rotation by itself
`does not generally reveal the optical purity of the product, other criteria of purity
`of the supposedly resolved substance should be applied: melting points (or more
`usefulJy the fusion interval, measured e.g., through differentiol scanning calo·
`rirnetry) in particular.
`The results are compiled into a crystallization-rotation matrix (for example,
`Table 2), which provides a semiquantitative sununary of the resolution trials. These
`results may be classified into three situations; (a) no crystalline diastereomers
`obtained; (b) crystalline salt obtained giving evidence of weak or no resolution; and
`(c) crystalline salt obtained corresponding to a net res.olution which may be fair
`or good.
`While the first of these situations precludes a resolution in the short term, it
`does not necessari1y mean that the mixture would not form a crystalline product
`
`DR. REDDY’S LABS., INC. EX. 1063 PAGE 7
`
`

`

`382
`
`Experimental Aspects and Art of Resolutions
`
`Table 2 Resolution trials for the phenylglyceric acidsa
`
`_0--cH-CH-COOH
`I
`I
`x
`OH OH
`
`X
`
`I
`
`2
`
`3
`
`4
`
`L 0
`0
`0
`
`Oil
`
`Threo
`
`Erythro
`
`H
`o-Ct
`m-Cl
`p-CI L
`H
`o-CI
`m-CI
`p-CI
`
`0
`
`I
`0 L
`d
`
`Oil
`
`5
`
`0
`
`0
`Oil
`Oil
`I
`
`6
`
`0
`
`0
`
`7
`
`8
`
`9
`
`10
`
`II
`
`12
`
`I
`
`0 Oil
`I_
`
`4
`ii
`
`0
`d
`
`Oil
`
`4
`T
`0
`d
`
`13
`
`0
`
`0
`
`a Samples of acid (J 00 mg, ca. 5.0 mmol) are treated with I equivalent base in I mL
`ethanol. The crystals obtained are directly de<:.omposed and the rotation of the acid
`measured. Key: 0 =no resolution; d or /, 4 or 1. 4 or 1 =weak, fair, or good
`I::::={+) a-metltylbenzYJamine; 2 = (+)-threo-1-p(cid:173)
`resolution. From
`ref. 13.
`3 = ( + )-threo-1-p-nitrophenyl-2-dimethyl(cid:173)
`nitrophenyl-2-amino-1 ,3-propanediol;
`amino-1,3-propanediol: 4 =(-)-ephedrine; 5 = (-)-deoxyephedrine; 6 = (-)(cid:173)
`amphetamine; 7 = Jehydroabietylamine; 8 = funtumine {Ja:-amino-Sa-prcgnan-
`20-onc); 9 =quinine; I 0 =cinchonine; ll =cinchonidine; 12 =strychnine; 13 =
`hrucine.
`
`were one to continue waiting for a longer period of time or, for example, if the
`trial were carried out in another laboratory.
`The second situation may correspond to the formation of a 1-1 addition com·
`pound between salts p and n (whence [a] 5ubsrrute = 0'\ to their cocrystallization
`(formation of a solid solution), or to an iilsufficient difference in solubility. All
`three of these possibilities render the resolution impractical.
`The last of these situations constitutes the favorable case, since it indicates
`that the diastereomer mixture has a eutectic sufficiently removed from an equimolar
`composition so as to make separation of the diastereomers possible.
`Table 2 gives some examples of the use of such resolution trials. Moreover,
`the problem is somewhat different if one is dealing with an isolated case or with a
`series of related substances. In the latter case, it is sometimes possible to limit the
`number of systematic trials to one or two members of the series and subsequently
`to orient the choice of resolving agent used with the other members of the series
`on the basis of the first results (see, for example, the threo-phenyiglyceric acids
`in Table 2). However, as we have seen in connection with the results of Winther and
`Werner (Section 53.1), this .. limited" approach may not always be successful.
`Table 2 shows that of the 33 trials, nine give usable results. By the same token,
`the 112 trials cited by Matell and collected in Table I of Section 5.3.1 furnish 26
`favorable cases. It is not possible to generalize the proportion of success (which
`
`7.2 Obtaining Crystalline Diastereomers
`
`383
`
`approximates one fourth in the two series of trials), particularly because one rarely
`fmds descriptions of resolution failures in the literature.
`Although resolution trials such as those described in Table 2 suggest that even
`for closely related compounds it is difficult to find any pattern in the choice of
`usable resolving agents, nonetheless analysis of reported resolutions as in Tables of
`Resolving Agents and Optical Resolutions3 improves the probability of success
`substantially over the random use of resolving agents in the trial matrix illustrated.
`For example, brocine is the preeminent resolving agent for the successful resolution
`of numerous alcohols via their hydrogen phthalate esters.
`While it remains true that this first step in resolutions is still virtually totally
`unpredictable, experience shows that if one conducts the search patiently and
`systematically, it is quite rare not to finally achieve success.
`Finally, we mention a variant in the way in which resolving agents are chosen
`in some specific cases. It consists in the use in the first trials not of the racemic
`substrate but of one of its eoantiomers. Tills is possible, for instance, when a total
`synthesis of a natural product leads to a racemate which is less accessible than one
`of the enantiomers.1 However, note that while the formation of a relatively insolu(cid:173)
`ble crystalline salt with this pure enantiomer and some resolving agent is a necessary
`condition for the sucess of the resolution, it is not a sufficient one. It is necessary
`that the racemate and the same resolving agent lead neither to an addition compound
`between the salts nor to their cocrystallization.
`
`7 .2.2 Choice of Crystallization Solvent
`
`We have already seen (Section 5.1) that the nature of the solvent has only a small
`effect on the ratio of solubilities of the two diastereomers, except in the case of
`differential solvation of the p and n diastereomers. One also knows that good
`laboratory practice calls for the use of crystallization solvents in which the solute
`is neither too soluble nor too insoluble. Given the ionic nature of the salts used in
`resolutions, one can predict that the best solvents will be polar ones. While there
`are exceptions to this generalization for diastereomeric salts. and it is probably less
`true for covalent diastereomers to be separated by crystallization, the solubility of
`salts in nonpolar solvents such as hexane or benzene is generally too low to permit
`their use. Moreover, such solvents sometimes even prevent the formation of salts;
`certain alkaloids crystallize from them in the pure state even in the presence of
`acids.
`Table 3 gives a statistical analysis of solvents used in over BOD resolutions 3
`involving the formation of salts. It is immediately evident that alcohols and acetone
`(anhydrous or aqueous). water. and mixtures of solvents containing an alcohol
`figure in about 90o/o of the cases. The distinction between anhydrous and aqueous
`solvents is important to the extent that one knows that the tendency of forming
`solva.tes is greater with water (hydrates) than with alcohols and other solvents.4
`Surprisingly, some resolutions take a different course according to whether they
`are carried out in absolute ethanol or in 96% ethanol, for example. Moreover,
`hydrate formation sometime