Mechanochemical carbon–carbon bond formation that proceeds via a cocrystal intermediate

Stipe Lukina, Martina Tirelia, Ivor Lončarića, Dajana Barišića, Primož Šketbc, Domagoj Vrsaljkod, Marco di Michiele, Janez Plavecbcf, Krunoslav Užarević*a and Ivan Halasz*a
aRuđer Bošković Institute, Bijenička c. 54, 10000 Zagreb, Croatia. E-mail: krunoslav.uzarevic@irb.hr; ivan.halasz@irb.hr
bNational Institute of Chemistry, Slovenian NMR Center, Ljubljana, Slovenia
cEN-FIST Center of Excellence, Ljubljana, Slovenia
dFaculty of Chemical Engineering and Technology, University of Zagreb, Zagreb, Croatia
eESRF – the European Synchrotron, 71 Avenue des Martyrs, 38000 Grenoble, France
fFaculty of Chemistry and Chemical Technology, University of Ljubljana, Ljubljana, Slovenia

Received 1st October 2018 , Accepted 31st October 2018

First published on 31st October 2018


We report the first cocrystal as an intermediate in a solid-state organic reaction wherein molecules of barbituric acid and vanillin assume a favorable orientation for the subsequent Knoevenagel condensation.


The Knoevenagel condensation is an important carbon–carbon bond forming reaction. More than a hundred years after the original report by Knoevenagel,1 Suzuki2 and Kaupp3 demonstrated an efficient and quantitative Knoevenagel condensation in the solid state achieved by milling. Other studies of solvent-free Knoevenagel condensation reactions soon followed.4–10 The reaction of barbituric acid (barb) and vanillin (van) was even used as a model mechanochemical organic reaction for assessing energetics of milling,11,12 to test twin-screw extrusion for solid-state organic synthesis,13 and latest, to reveal a peculiar deviation of solid-state reaction kinetics from the one observed in solution, stemming from changes in the rheology of the milled sample.14 However, studies of barb–van Knoevenagel condensation were thus far limited to ex situ reaction monitoring by, e.g., solution UV-Vis11 or NMR spectroscopies.14

In this work, we employ real-time in situ Raman spectroscopy monitoring15,16 to reveal that the solid-state Knoevenagel condensation (Scheme 1) of barb and van proceeds through a cocrystal intermediate. In the cocrystal, packing of barb and van is such that molecules of barb are suitably positioned for the nucleophilic attack to the carbonyl C-atom of van, which is the first step of the Knoevenagel condensation. Furthermore, solid-state ab initio calculations offer a plausible reaction path for the C[double bond, length as m-dash]C bond to form within the cocrystal. Liquid-assisted grinding (LAG) with ethanol considerably accelerated the reaction in comparison to LAG with non-protic liquids or neat grinding. On the other hand, LAG with a strong non-nucleophilic base provided a reaction pathway that circumvented the cocrystal intermediate.


image file: c8cc07853j-s1.tif
Scheme 1 Solid-state Knoevenagel condensation of van and barb under mechanochemical milling conditions. The new C[double bond, length as m-dash]C bond is shown in red.

As far as we are aware, the Knoevenagel condensation of barb and van studied here is the first mechanochemical organic reaction where a cocrystal of reactants is formed in situ during milling before the targeted reaction and the formation of a new covalent bond occurred. Nevertheless, cocrystals as multicomponent solids have been recognised as a promising medium to conduct solid-state reactions efficiently and selectively.17–19 As such, they have been used at length in crystal engineering to design and construct solids to enable specific and usually photochemical reactivity.20–23 Targeted use of cocrystals to achieve thermal solid-state reactions has also been explored, albeit to a lesser extent.24–29

Milling of 1 mmol of barb and 1 mmol of van was performed using a vibratory IST500 ball mill operating at 30 Hz, and equipped with 2 stainless steel balls of 7 mm (each weighing 1.4 g) in diameter as the milling media. In situ monitoring of neat grinding revealed only mixing of reactants during the first 70 minutes of milling after which a new intermediate phase started to emerge, and it was followed by a slow formation of the product I over a period of several hours (Fig. 1). According to solution NMR, I was the expected Knoevenagel condensation product and was amorphous as evidenced by PXRD (Fig. S2, ESI). In a repeated experiment, we used in situ monitoring to isolate the intermediate by stopping the reaction at an appropriate time and solved its crystal structure from laboratory PXRD data (Fig. 2a), revealing it to be a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 cocrystal of barb and van.


image file: c8cc07853j-f1.tif
Fig. 1 In situ Raman monitoring of neat grinding of barb and van. Raman spectra of barb, van, the cocrystal, and I are given above the 2D plot. Formation of the cocrystal can be monitored by the Raman band at 1735 cm−1 that corresponds to the C[double bond, length as m-dash]O stretching of barb molecules in the cocrystal. The Raman band of the newly formed C[double bond, length as m-dash]C bond at 1542 cm−1 indicates the formation of I. Intensity colour code: blue – low, red – high.

image file: c8cc07853j-f2.tif
Fig. 2 (a) Rietveld plot for the 1[thin space (1/6-em)]:[thin space (1/6-em)]1 barb:van cocrystal structure determined from powder diffraction data collected using CuKα X-rays. A high-angle region is enlarged to reveal more details. Refinement parameters: Rp = 2.3%, Rwp = 3.1%, RBragg = 1.0%, and gof = 1.91. (b) Ribbons of barb and stacks of van molecules in the cocrystal. Orange circles highlight the closest distance between the barb methylene group and the van carbonyl group. The 3.7 Å is the distance between the corresponding carbon atoms. Dashed bonds represent hydrogen bonds. (c) Two Knoevenagel condensation product molecules and one water molecule in the asymmetric unit of II. Arrows point to newly formed C[double bond, length as m-dash]C bonds.

The cocrystal exhibits ribbons of barb molecules while molecules of van pack in stacks (Fig. 2b). Strong N–H⋯O hydrogen bonds connect molecules of barb within the ribbons, leaving one carbonyl O atom to be an acceptor in a hydrogen bond with the hydroxy group of van. Having both reacting molecules within a crystal brings to mind the topochemical principle which anticipates that the reacting centers should be close.30 The topochemical principle may hardly be applicable here since the average crystal structure needs not to reflect the changes occurring at the molecular level.31 However, we do note that the arrangement of molecules in the barb:van cocrystal is such that the barb methylene group and the aldehyde group of van are separated by only ca. 3.7 Å and suitably orientated for the addition of the barb nucleophile to the carbonyl C atom of van (Fig. 2b).

We were thus inspired to examine the potential topochemical reaction in the cocrystal by using solid-state ab initio calculations. Indeed, we have been able to identify a plausible reaction path for the barb methylene group to attack the aldehyde C atom of van and to create a new C[double bond, length as m-dash]C bond within the cocrystal. In the process, one water molecule simultaneously leaves, and the calculated energy barrier is below 350 kJ mol−1 (Fig. 3). This, however, is an overestimation of the energy barrier resulting from the formation of one product molecule out of the four pairs of barb and van in the cocrystal unit cell. Periodic repetition of the unit cell prevents relaxation of the crystal geometry as it would be the case if only one molecule of the product would form in a bigger crystal. Calculations of a single pair of reactant molecules under vacuum gave an energy barrier of around 130 kJ mol−1 (Fig. S1, ESI).


image file: c8cc07853j-f3.tif
Fig. 3 Minimum energy path for the reaction of one molecule of barb and one molecule of van within the unit cell of the cocrystal. The dashed line is cubic spline interpolation and is a guide to the eye.

Since the neat grinding reaction takes over 14 hours to complete, we tried to accelerate the product formation with the use of liquid-assisted grinding.18,32,33 Liquid additives are known to accelerate and also selectively direct mechanochemical reactions,34,35 or even may act as an acid–base catalyst.16 Here, we have explored LAG with the addition of 20 μL of acetonitrile (MeCN), nitromethane (MeNO2), ethanol (EtOH) and N,N-diisopropylethylamine (dipea).

LAG with MeCN or MeNO2 resulted in fast cocrystal formation which started 3 min into milling (Fig. 4a and Fig. S4, S5, ESI). Surprisingly, the subsequent C[double bond, length as m-dash]C bond formation was slowed down revealing a prolonged life-span of the cocrystal under those reaction conditions (Fig. 4). This can be explained by the stabilisation of the cocrystal with the liquid additives, as was previously observed in related systems.35,36 However, the amorphous product I remained stable after 50 hours of neat grinding (Fig. S3, ESI), and prolonged milling in LAG reactions resulted in transformation of I to new forms of the Knoevenagel condensation product, II and III, observed after ca. 15 and ca. 20 hours of continuous milling, respectively (Fig. S4 and S5, ESI).


image file: c8cc07853j-f4.tif
Fig. 4 (a) Formation kinetics and life-span of barb:van cocrystal in neat grinding and LAG reactions, derived from the cocrystal Raman band at 1735 cm−1, and (b) kinetics of C[double bond, length as m-dash]C bond formation derived from changes in intensities of the band at 1542 cm−1. LAG with EtOH yielded III while other liquids yielded I. The intensity of the C[double bond, length as m-dash]C band is lower in III than in I. Note the different scales in (a) and (b).

The new forms II and III are crystalline, as opposed to the amorphous form I (Fig. S2, ESI), and each has a distinct Raman spectrum (Fig. S24, ESI). According to solution NMR, both II and III are the expected Knoevenagel condensation product (Fig. S2, S12–S14 and S24, ESI), and thermal analysis suggested that I and II are hydrates while III is likely anhydrous (Fig. S25–S28, ESI). Formation of a hydrate product was previously suggested3 and is not surprising since water is a byproduct of Knoevenagel condensation. This is corroborated by the crystal structure of II, also confirming the molecular structure of the Knoevenagel condensation product (Fig. 2c). Possible tautomeric variations in II and III are excluded based on 15N solid-state NMR spectra which exhibited the same chemical shifts for nitrogen atoms (Fig. S20, ESI).

The reaction using EtOH as a protic liquid additive revealed fast cocrystal formation, but also a significantly increased rate of C[double bond, length as m-dash]C bond formation, which was complete within ca. 2 hours (Fig. 4 and Fig. S6, ESI). Since other studied reactions were rather slow, we have chosen this reaction for a monitoring experiment using in situ synchrotron PXRD,37–39 in tandem with Raman spectroscopy.35,40 PXRD showed persistent but diminishing diffraction signals from reactants, and a weak signal coming from the cocrystal. The Knoevenagel product, however, for a limited period of 70 min milling, remained X-ray amorphous, but Raman spectroscopy revealed its formation soon into milling, as evidenced by the emergence of the C[double bond, length as m-dash]C Raman band at 1542 cm−1 (Fig. S31, ESI).

With dipea, a non-nucleophilic base, the reaction proceeded directly from reactants to I (Fig. S7, ESI) circumventing the cocrystal. Bulky isopropyl and ethyl groups of dipea possibly could not achieve interactions that would stabilise the cocrystal, as was the case with other liquids. As a base, it also likely facilitates proton abstraction from barb and thus the nucleophilic attack on the carbonyl group of van, resulting in a faster reaction rate than in neat grinding.

Being able to isolate the barb:van cocrystal as a pure phase, we were also interested in thermal initiation of Knoevenagel condensation in the cocrystal.17,24 Heating the cocrystal to 150 °C resulted in complete conversion to the anhydrous product III (Fig. S2, ESI). Interestingly, our attempts to synthesise the cocrystal from solution failed (Fig. S30, ESI), indicating that it may be accessible only by milling.34,41

Finally, we note a striking difference in the observed reactivity of barb and van in our hands and in a recent report by James and coworkers.14 Namely, our reactions did not change the physical form of the reaction mixture and sticking of the reaction mixture to the milling media. A possible cause for this difference stems from the use of two small milling balls in our case, as compared to the use of one large (13.6 g) ball by James. Two balls, with their mutual collisions, could possibly clean one another, but more importantly, one large milling ball significantly heated the reaction mixture,14 while two small milling balls cause a minor increase in temperature.42,43 We believe strong heating in combination with a low melting point of vanillin (ca. 82 °C) could have caused eutectic melting of the reaction mixture, which could have readily changed the physical form of the reaction mixture. In addition, lower temperature in our milling setup also explains slower reactions, since even a small or moderate increase in temperature can dramatically influence mechanochemical reaction rates.42,44

In summary, we report, as far as we are aware, the first cocrystal as an intermediate in a mechanochemical organic reaction. Remarkably, the cocrystal suitably positioned the reacting centers for the subsequent carbon–carbon bond formation. Coupled with crystal engineering, this observation reveals possibilities for controlled use of cocrystals for targeted reactivity in milling reactions. Moreover, with the choice of liquid additives, we were able to affect cocrystal formation and its stability under milling. We also highlight the value of in situ reaction monitoring which enabled identification and isolation of this unique and easily overlooked intermediate. Currently, we are undertaking a detailed in situ kinetic study of this model organic reaction with the intention to elucidate the influence of liquid additives.

We are grateful to Mr Vitomir Stanišić, Mr Ivan Kulcsár and the team at the fine-mechanics workshop of the Ruđer Bošković Institute for their continuous support. We thank Tomislav Stolar for critically reading the manuscript. Croatian Science Foundation (Grant No. UIP-2014-09-4744) is gratefully acknowledged. IH is grateful to the Adris Foundation for financial support. SL is supported by the Croatian Science Foundation. PŠ and JP acknowledge support by the Slovenian Research Agency, grant number P1-0242. IL was supported by the European Union through the European Regional Development Fund – the Competitiveness and Cohesion Operational Programme (KK.01.1.1.06) and the H2020 CSA Twinning project No. 692194, RBI-T-WINNING.

Conflicts of interest

There are no conflicts of interest to declare.

References

  1. E. Knoevenagel, Ber. Dtsch. Chem. Ges., 1898, 31, 2596–2619 CrossRef CAS.
  2. S. Wada and H. Suzuki, Tetrahedron Lett., 2003, 44, 399–401 CrossRef CAS.
  3. G. Kaupp, M. R. Naimi-Jamal and J. Schmeyers, Tetrahedron, 2003, 59, 3753–3760 CrossRef CAS.
  4. R. Schmidt, C. F. Burmeister, M. Baláž, A. Kwade and A. Stolle, Org. Process Res. Dev., 2015, 19, 427–436 CrossRef CAS.
  5. D. V. Demchuk, M. N. Elinson and G. I. Nikishin, Mendeleev Commun., 2011, 21, 224–225 CrossRef CAS.
  6. R. Trotzki, M. M. Hoffmann and B. Ondruschka, Green Chem., 2008, 10, 767–772 RSC.
  7. L. Rong, X. Li, H. Wang, D. Shi, S. Tu and Q. Zhuang, Synth. Commun., 2006, 36, 2407–2412 CrossRef CAS.
  8. S. Mashkouri and M. R. Naimi-Jamal, Molecules, 2009, 14, 474–479 CrossRef CAS PubMed.
  9. S. Haferkamp, F. Fischer, W. Kraus and F. Emmerling, Beilstein J. Org. Chem., 2017, 13, 2010–2014 CrossRef CAS PubMed.
  10. S. Haferkamp, W. Kraus and F. Emmerling, J. Mater. Sci., 2018, 53, 13713–13718 CrossRef CAS.
  11. A. Stolle, R. Schmidt and K. Jacob, Faraday Discuss., 2014, 170, 267–286 RSC.
  12. C. F. Burmeister, A. Stolle, R. Schmidt, K. Jacob, S. Breitung-Faes and A. Kwade, Chem. Eng. Technol., 2014, 37, 857–864 CrossRef CAS.
  13. D. E. Crawford, C. K. G. Miskimmin, A. B. Albadarin, G. Walker and S. L. James, Green Chem., 2017, 19, 1507–1518 RSC.
  14. B. P. Hutchings, D. E. Crawford, L. Gao, P. Hu and S. L. James, Angew. Chem., Int. Ed., 2017, 56, 15252–15256 CrossRef CAS PubMed.
  15. D. Gracin, V. Štrukil, T. Friščić, I. Halasz and K. Užarević, Angew. Chem., Int. Ed., 2014, 53, 6193–6197 CrossRef CAS PubMed.
  16. M. Tireli, M. Juribašić Kulcsar, N. Cindro, D. Gracin, N. Biliškov, M. Borovina, M. Ćurić, I. Halasz and K. Užarević, Chem. Commun., 2015, 51, 8058–8061 RSC.
  17. M. Etter, G. Frankenbach and J. Bernstein, Tetrahedron Lett., 1989, 30, 3617–3620 CrossRef CAS.
  18. T. Friščić and W. Jones, Cryst. Growth Des., 2009, 9, 1621–1637 CrossRef.
  19. J. W. Lauher, F. W. Fowler and N. S. Goroff, Acc. Chem. Res., 2008, 41, 1215–1229 CrossRef CAS PubMed.
  20. L. R. MacGillivray, J. L. Reid and J. A. Ripmeester, J. Am. Chem. Soc., 2000, 122, 7817–7818 CrossRef CAS.
  21. L. R. MacGillivray, G. S. Papaefstathiou, T. Friščić, T. D. Hamilton, D.-K. Bučar, Q. Chu, D. B. Varshney and I. G. Georgiev, Acc. Chem. Res., 2008, 41, 280–291 CrossRef CAS PubMed.
  22. A. Sokolov, D.-K. Bučar, J. Baltrusaitis, S. Gu and L. MacGillivray, Angew. Chem., Int. Ed., 2010, 49, 4273–4277 CrossRef CAS PubMed.
  23. A. V. Trask and W. Jones, in Crystal Engineering of Organic Cocrystals by the Solid-State Grinding Approach, ed. F. Toda, Springer, Berlin, Heidelberg, 2005, pp. 41–70 Search PubMed.
  24. M. L. Cheney, G. J. McManus, J. A. Perman, Z. Wang and M. J. Zaworotko, Cryst. Growth Des., 2007, 7, 616–617 CrossRef CAS.
  25. J. A. Perman, A. J. Cairns, L. Wojtas, M. Eddaoudi and M. J. Zaworotko, CrystEngComm, 2011, 13, 3130–3133 RSC.
  26. J. A. Perman, K. Dubois, F. Nouar, S. Zoccali, u. Wojtas, M. Eddaoudi, R. W. Larsen and M. J. Zaworotko, Cryst. Growth Des., 2009, 9, 5021–5023 CrossRef CAS.
  27. J. H. Kim, S. M. Hubig, S. V. Lindeman and J. K. Kochi, J. Am. Chem. Soc., 2001, 123, 87–95 CrossRef CAS PubMed.
  28. J. H. Kim, S. V. Lindeman and J. K. Kochi, J. Am. Chem. Soc., 2001, 123, 4951–4959 CrossRef CAS PubMed.
  29. J. Schmeyers, F. Toda, J. Boy and G. Kaupp, J. Chem. Soc., Perkin Trans. 2, 1998, 989–994 RSC.
  30. G. Schmidt, Pure Appl. Chem., 1971, 27, 647–678 CAS.
  31. I. Halasz, Cryst. Growth Des., 2010, 10, 2817–2823 CrossRef CAS.
  32. N. Shan, F. Toda and W. Jones, Chem. Commun., 2002, 2372–2373 RSC.
  33. A. V. Trask, W. D. S. Motherwell and W. Jones, Chem. Commun., 2004, 890–891 RSC.
  34. S. Karki, T. Friščić, W. Jones and W. D. S. Motherwell, Mol. Pharmacol., 2007, 4, 347–354 CrossRef CAS PubMed.
  35. S. Lukin, T. Stolar, M. Tireli, M. V. Blanco, D. Babić, T. Friščić, K. Užarević and I. Halasz, Chem. – Eur. J., 2017, 23, 13941–13949 CrossRef CAS PubMed.
  36. A. M. Belenguer, G. I. Lampronti, A. J. Cruz-Cabeza, C. A. Hunter and J. K. M. Sanders, Chem. Sci., 2016, 7, 6617–6627 RSC.
  37. I. Halasz, S. A. J. Kimber, P. J. Beldon, A. M. Belenguer, F. Adams, V. Honkimäki, R. C. Nightingale, R. E. Dinnebier and T. Friščić, Nat. Protoc., 2013, 8, 1718–1729 CrossRef PubMed.
  38. T. Friščić, I. Halasz, P. A. Beldon, A. M. Belenguer, F. Adams, S. A. J. Kimber, V. Honkimäki and R. E. Dinnebier, Nat. Chem., 2013, 5, 66–73 CrossRef PubMed.
  39. K. Užarević, I. Halasz and T. Friščić, J. Phys. Chem. Lett., 2015, 6, 4129–4140 CrossRef PubMed.
  40. L. Batzdorf, F. Fischer, M. Wilke, K.-J. r. Wenzel and F. Emmerling, Angew. Chem., Int. Ed., 2015, 54, 1799–1802 CrossRef CAS PubMed.
  41. A. V. Trask, J. van de Streek, W. D. S. Motherwell and W. Jones, Cryst. Growth Des., 2005, 5, 2233–2241 CrossRef CAS.
  42. K. Užarević, V. Štrukil, C. Mottillo, P. A. Julien, A. Puškarić, T. Friščić and I. Halasz, Cryst. Growth Des., 2016, 16, 2342–2347 CrossRef.
  43. K. Užarević, N. Ferdelji, T. Mrla, P. Julien, B. Halasz, T. Friščić and I. Halasz, Chem. Sci., 2018, 9, 2525–2532 RSC.
  44. J. M. Andersen and J. Mack, Chem. Sci., 2017, 8, 5447–5453 RSC.

Footnotes

Electronic supplementary information (ESI) available: Experimental and computational details, in situ Raman and PXRD data, thermal measurements, NMR data and FTIR data. CCDC 1869002 and 1818969. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8cc07853j
These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2018