3 Replies Latest reply: Jul 8, 2011 11:17 AM by desiraju RSS

    Are polymorphs obvious?  What is obvious about crystallization and why this really, really matter$. Part 2.


      Every compound has different polymorphic forms and the number offorms known for a given compound is proportional to the time and energy spentin research on that compound"

                  McCrone’s provocative comment1 is very well known within the crystal engineering community and it is fair to assert that it has inspired research into the phenomenon of polymorphism.  However, one cannot accept it to be true since it is impractical to validate even in the era of robotic high throughput crystallization. Simply put, it would be practically impossible to conduct an infinite number of crystallization experiments on every compound that is not yet proven to exhibit polymorphism. Additionally, there are compounds that have been crystallized many times, as exemplified by naphthalene and sucrose, that are not yet known to exhibit polymorphism under ambient conditions.2,3Interestingly, aspirin used to be in this category but now there is a second polymorph of aspirin,4 although the nature of aspirin polymorphism remains a touchy subject.5,6 It would be nice to have the resources to go hunting for new crystal forms of naphthalene and sucrose but it would behard to justify funding for such a project.

                  Why does McCrone’s comment matter greatly? Much ofthe recent interest in polymorphism can be attributed to the legal implicationsof polymorphism in the context of the pharmaceutical industry. In particular,since a polymorph is not obvious until it has been made, new polymorphstypically meet the criteria of novelty, utility and lack of obviousness neededfor patentability in the United States. This issue was brought to the fore by thediscovery of form II of ranitidine hydrochloride or Zantac®, the #1 blockbusterdrug at the time of litigation (see SOTW 3). However, in Europe the equivalent criterion for issuing a patent is “inventiveness” rather than “obviousness” and this means that polymorphs discovered through high throughput screening might henceforth be difficult to patent because McCrone’s comment is regarded by some as scientifically accepted.  Ironically, at least from the perspective of physicochemical properties, this is a “much ado about nothing”since polymorphs are often similar enough to each other in aqueous solubility (but not always) to make them bioequivalent.7 Therefore, this matter is primarily a legal and regulatory issue rather than a matter of whether thedrug works.

                  From my perspective it is not yet clear that polymorphism is a rule rather than an exception. What we do know is that of the 233,346 “organic only” crystal structures archived in the Cambridge Structural Database (see SOTW 1), there are only 9870 entries for polymorphs, .i.e. less than 9870/2 compounds out of (233,346-9870/2)or ca. 2.2% are presently known to be polymorphic.  For a more in-depth perspective on polymorphism, including its legal implications, see JoelBernstein’s recent article on the subject of polymorphism.8


      (1) W. C. McCrone, inPhysics and Chemistry of the Organic Solid State, eds. D. Fox, M. M. Labes andA. Weissberger, Interscience Publishers, London, 1965, vol. 2, pp. 725-767.

      (2) S. Block, C.E. Weir, G.J. Piermarinin, Polymorphism in Benzene, Naphthalene, and Anthracene at High Pressure”,1970, Science, 169, 586-587.

      (3)  T. Lee, G. D. Chang, “Sucrose Conformational Polymorphism: AJigsaw Puzzle with Multiple Routes to a Unique Solution”, 2009, Cryst. Growth Des., 9, 3551–3561

      (4) P. Vishweshwar, J. A. McMahon, M.Oliveira, M.L. Peterson, M. J. Zaworotko, "The Predictably Elusive Form IIof Aspirin", 2005, J.Am. Chem. Soc. 127,16802–16803.

      (5) A.D. Bond, R. Boese, G.R. Desiraju, "On thePolymorphism of Aspirin: Crystalline Aspirin as Intergrowths of Two"Polymorphic" Domains", 2007, Angewandte Chemie, International Edition, 46, 618–622.

      (6) E. J.Chan, T.R. Welberry, A.P.Heerdegen, D.J. Goossens, “Diffuse scattering study of aspirin forms (I)and (II)”, 2010, ActaCrystallographica, Section B: Structural Science, B66,  696-707.

      (7) M. Pudipeddi, A.T.M.Serajuddin, “Trends in Solubility ofPolymorphs”, 2005, Journal of PharmaceuticalSciences, 94, 929-939.

      (8) J. Bernstein, “Polymorphism - A Perspective, CrystalGrowth & Design”, 2011, 11,632-650.

        • Are polymorphs obvious?  What is obvious about crystallization and why this really, really matter$. Part 2.

          Dear Michael,


          thanks for your interesting thoughts about polymorphism and McCrone. What do you think about the computation of energy landscapes?


          Nowadays accurate computation of energy landscapes becomes more and more efficient, see the last blind test about crystal structure prediction and the publications of M. Neumann [1-6].


          The outcome of nearly every simulation results in several energetically closely related crystal structures / packing motifs within ~5 to 10 kJ/mol. Of course, one must keep the accuracy of these calculations in mind, but the accuracy will increase in future.


          In my opinion the results from the calculations published so far reinforce the hypothesis of McCrone. Maybe we are simply unable to find routes for the crystallization of these crystal structures up to now and need to search for different novel crystallization paths?


          I think only time can answer this rather philosphical question, but applying computational chemistry with the power of modern quantum chemical calculations may help to provide deep insights into inter- and intramolecular forces and packing motifs and may help to understand polymorphism.




          [1] W. D. S.Motherwell and et al., Acta Crystallogr., Sect. B: Struct. Sci.,

          2002, 58, 647–661

          [2] G.M. Day and et al., Acta Crystallogr., Sect. B: Struct. Sci., 2005, 61,


          [3] G.M. Day and et al., Acta Crystallogr., Sect. B: Struct. Sci., 2009, 65,


          [4] M. A. Neumann, F. J. J. Leusen and J. Kendrick, Angew. Chem., Int.

          Ed., 2008, 47, 2427–2430.

          [5] M. A. Neumann and M.-A. Perrin, J. Phys. Chem. B, 2005, 109,


          [6] M. A. Neumann, J. Phys. Chem. B, 2008, 112, 9810–9829.

            • Are polymorphs obvious?  What is obvious about crystallization and why this really, really matter$. Part 2.

              Dear Juergen:


              Thank you for your thoughtful response to my question.  I believe that your comments raise important questions because of the implications of crystal structure prediction.  Indeed, my next post was going to ask the following question, which as you have stated is now a rhetorical question: are crystals structures predictable?  I am old enough to remember reading John Maddox’s 1988 Nature editorial on the subject of crystal structure prediction. It served as a source of inspiration to the crystal engineering community. In particular, his provocative assertion that "one of the continuing scandals in the physical sciences is that it remains impossible to predict the structure of even the simplest crystalline solids from a knowledge of their composition"1 helped to spur a new generation of crystal engineers. However, although crystal engineering (which is focused upon the generation of new compounds) is not synonymous with crystal structure prediction, I agree with you that the science of crystal structure prediction has advanced to the point that polymorphs can be predicted before they have been made.  In my personal experience this is exemplified by aspirin form II2 and 5-fluorouracil.3 The power of modern crystal structure prediction methodology begs three further questions:


              - What does crystal structure prediction mean for the patentability of new polymorphs? For example, what if a particular crystal packing in a specific space group is reported in a manuscript before it has been made. Since a PXRD pattern can be readily calculated from a simulated crystal structure can one claim PXRD peaks or unit cell parameters in a patent if theyare already disclosed?


              - How close a match does a predicted structure have to be to the real structure to be considered correct? Presumably this is a matter that should be quantifiable.


              - How reliable is crystal structure prediction methodology for more challenging compounds such as solvates/hydrates(for which composition is often variable) and larger molecules (especially those with numerous hydrogen bonding sites and/or flexible functional groups).


              1. Maddox, J. Nature, 1988, 335, 201.

              2. Ouvrard, C.; Price, S.L. CrystalGrowth & Design, 2004, 4, 1119-1127.

              3. Hulme, A.T.; Price, S.L.; Tocher, D.A.  Journal of the American Chemical Society, 2005, 127,  1116-1117. 

            • Re: Are polymorphs obvious?  What is obvious about crystallization and why this really, really matter$. Part 2.

              Dear Mike,


              I doubt whether anyone would find a normal pressure polymorph of either naphthalene or sucrose that easily. If at all. And yes, I do feel that there has been a tendency, in some quarters, to assume that McCrone's statement is always applicable. This statement was over used by many of us around a decade ago (perhaps like a joke in talks) and it sort of got stuck in the general perception. It would be truly unfortunate if this statement was accepted in legal matters as a scientific truth.