- Email: [email protected]
Reaction: Technological Aspects of Molecular Electronics
underlying reactions and uncover important details of the most basic processes of life, which will immediately invite intense research activities.
ACKNOWLEDGMENTS This work was supported by the Ministry of Science and Technology of China (grant 2017YFA0204901) and the National Natural Science Foundation of China (grants 21373014 and 21727806). 1. Waldrop, M.M. (2016). Nature 530, 144–147. 2. Aviram, A., and Ratner, M.A. (1974). Chem. Phys. Lett. 29, 277–283. 3. Reed, M.A., Zhou, C., Muller, J., Burgin, T.P., and Tour, J.M. (1997). Science 278, 252–254. 4. Reecht, G., Scheurer, F., Speisser, V., Dappe, Y.J., Mathevet, F., and Schull, G. (2014). Phys. Rev. Lett. 112, 047403. 5. Xiang, D., Wang, X., Jia, C., Lee, T., and Guo, X. (2016). Chem. Rev. 116, 4318–4440. 6. Jia, C., Migliore, A., Xin, N., Huang, S., Wang, J., Yang, Q., Wang, S., Chen, H., Wang, D., Feng, B., et al. (2016). Science 352, 1443–1445. 7. Thiele, S., Balestro, F., Ballou, R., Klyatskaya, S., Ruben, M., and Wernsdorfer, W. (2014). Science 344, 1135–1138. 8. Natterer, F.D., Yang, K., Paul, W., Willke, P., Choi, T., Greber, T., Heinrich, A.J., and Lutz, C.P. (2017). Nature 543, 226–228. 9. Guo, C., Wang, K., Zerah-Harush, E., Hamill, J., Wang, B., Dubi, Y., and Xu, B. (2016). Nat. Chem. 8, 484–490. 10. Zhang, Y., Luo, Y., Zhang, Y., Yu, Y.J., Kuang, Y.M., Zhang, L., Meng, Q.S., Luo, Y., Yang, J.L., Dong, Z.C., and Hou, J.G. (2016). Nature 531, 623–627. 11. Aragone`s, A.C., Haworth, N.L., Darwish, N., Ciampi, S., Bloomfield, N.J., Wallace, G.G., Diez-Perez, I., and Coote, M.L. (2016). Nature 531, 88–91. 12. Wen, H., Li, W., Chen, J., He, G., Li, L., Olson, M.A., Sue, A.C.-H., Stoddart, J.F., and Guo, X. (2016). Sci. Adv. 2, e1601113. 13. Jia, C., and Guo, X. (2013). Chem. Soc. Rev. 42, 5642–5660. 14. Lo¨rtscher, E. (2013). Nat. Nanotechnol. 8, 381–384. 1Beijing
National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, P.R. China *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chempr.2017.08.006
376
CATALYSIS
Reaction: Technological Aspects of Molecular Electronics Emanuel Lo¨rtscher1,* Dr. Emanuel Lo¨rtscher received his PhD with distinction (summa cum laude) in 2006 from the University of Basel in Switzerland and a master in physics in 2003 from the Swiss Federal Institute of Technology. He has been a research staff member at IBM Research since 2008. He has received various prizes, including the Swiss Physical Society Award for Applied Physics in 2007 and the Faculty Prize of the University of Basel in the same year. His current research focuses on fabrication, electronics, optics, and chemistry at the nanoscale for applications in novel computing applications, biology, medicine, and the Internet of Things.
Moore’s Law—or the miniaturization of the complementary metal-oxide semiconductor (CMOS)—is about to end because performance no longer benefits from downsizing. Consequently, semiconductor research now focuses on employing novel materials and architectures to fulfill specifications requested by applications. This situation provides opportunities for fundamentally different electronic components, among which molecular electronics (ME) is considered again.1 Challenges of CMOS Technology Transistor scaling slows down because miniaturization has reached integration densities that cause unbearable thermal loads (averaged 200–300 W/cm2, locally >500 W/cm2), and future nodes are becoming unaffordable2 in terms of fabrication investments. Hence, getting smaller is no longer the maxim.
Chem 3, 373–379, September 14, 2017 ª 2017 Elsevier Inc.
Gate leakage, short-channel effects, and parasitics must be kept low so as not to consume energy when idle, for example. The semiconductor industry has faced similar issues during its life cycle—causing low yields, inhomogeneities in doping levels and driving voltages, reduced reliability, etc.—but it has always been able to overcome them, for example, by Si-on-insulator (SOI), strained Si, fin field-effect transistors (FinFETs), nanowires, high-k dielectrics, airgaps, 3D stacking, multicores, III/V integration, etc. Hence, CMOS will change, but it will remain the predominant technology for the coming years. A successor must not only provide the same specs as International Technology Roadmap for Semiconductors (ITRS) and non-ITRS anticipated CMOS nodes but also outperform prevalent technology in at least one key aspect: power consumption, fabrication costs, and/or performance. The argumentation is therefore not solely technical but also economical to realign a multibillion-USD industry. The successor can leverage CMOS platforms by being co-integrated wherever such novel components outmatch Si but only if they are process compatible. Topdown fabrication can be pushed to <5 nm where quantum effects and uncontrollable doping levels are detrimental for Si. At such dimensions, however, various bottom-up approaches might enter the game. State of the Art in ME The promises of ME are intrinsic functionalities embodied in a molecular backbone, creating electrical responses, e.g., conductance, current rectification, resistance switching, etc. Such ultimately scaled building blocks can be synthesized identically with picometer control. Negative differential resistance, Coulomb and Kondo effects, quantum interference, and switching are only some of the interesting functionalities demonstrated.
Gaining control over these effects, however, is difficult, and large sampleto-sample fluctuations appear. Converging agreement between theory and experimentation has been reached for simple systems; most structure-property relationships, however, are deduced empirically, mostly from large statistical datasets generated by mechanical manipulation, rather than predicted. CMOS in return is predictable because devices can be designed and properties can be tailored by simulations. Besides ME, however, no other nanoelectronics candidate has reached this maturity so far. Compared with Si as the only modestly modified CMOS core material, the limitless library of potentially suited molecular compounds is huge, and there is little consensus on future screening methods. The quantum-mechanical nature of molecules is considered advantages, e.g., as a result of quantum interference. Transport through molecular orbitals, however, is mediated by a variety of mechanisms, all of which are of low conductance, especially in comparison with ballistic transport offered by layered materials. Additionally, most molecular orbitals are misaligned with Fermi energies and require high voltages for resonant transport—the opposite of the ideal switch’s imperatives. Even for fully conjugated molecular wires with hybridized moleculemetal interfaces, molecules must be categorized as insulators rather than conductors, rendering a single-molecule device unviable because heat dissipation will be higher at comparable integration densities. Furthermore, experiments revealing distinct features are mostly done at ambient or cryogenic temperatures. Although there will be exceptions where cooling is affordable, e.g., a quantum computer, most electronics will be operated above 300 K. At typical temperatures of 85 C, many effects discovered in single-molecule junctions will fade out, and compounds can degrade ther-
mally. Last, assembling multiple compounds to a circuit requires site-selective depositions. Although molecular selfassembly and recognition are powerful bottom-up mechanisms, none of these has made it out of the lab. It is therefore premature to discuss any cost advantages. Toward Technology-Relevant Devices With these challenges ahead, why should ME still try to become a nanoelectronics alternative? One motivation arises from novel computing paradigms such as quantum, neuromorphic, and cognitive computing. Among the breadth of molecular compounds, some compounds might be ideally suited for non-linear workflows because of intrinsic sublayers and backpropagation or some supramolecular compounds might have multiple ports and voltage-added responses that behave like synapses. Furthermore, metalorganic compounds could provide abrupt electronic features addressable in two-terminal geometries thanks to redox- and spin-based mechanisms3 while simultaneously aligning molecular orbitals with microelectronics voltage ranges. Besides logic and memory applications, molecules could be also excellent candidates for sensing tasks in emerging Internet of Things (IoT) applications. ME concepts can also be applied in optical or plasmonic hybrids as electrically driven singlephoton emitters or sensitive electro-optical detectors based on molecular transitions. To make that happen, however, we must target methods that reliably and non-destructively integrate molecules into a solid-state platform. Current experiments are mainly based on single-molecule experiments with mechanical actuation. Although those test beds are excellent research tools, their very-large-scale integration is difficult, expensive, and not scalable. Instead, one could start leveraging
early ME platforms4,5 that are technology driven but have fallen into disgrace: nanopores and cross-bar structures with sandwiched molecular ensembles are considered technologically relevant because variations are small as a result of multi-molecule averaging and constrained geometries. Ensemble performances are mostly superior (e.g., for diode rectification ratios 1,0003 larger6) over single-molecule ones because they can benefit from heterogeneous compositions and related effects (image charges, dipole-induced barrier lowering, etc.). Graphene, carbon nanotubes, or low-mobility metal electrodes and direct grafting on semiconductors achieve better controllable and more stable interfaces. Molecules should be assembled in compartments and protected during fabrication and operation against environmental effects and chemical degradation. The architectures mentioned above provide affordable footprints, relevant current levels, methods for site-selective assembly, and compatibility and scalability to wafer-scale integration— a feasible pathway for post-CMOS.
1. Xin, N., and Guo, X. (2017). Chem 3, this issue, 373–376. 2. Cross, T. (2016). After Moore’s Law. The Economist, March 12, 2016. http://www. economist.com/technology-quarterly/201603-12/after-moores-law. 3. Schwarz, F., Kastlunger, G., Lissel, F., EglerLucas, C., Semenov, S.N., Venkatesan, K., Berke, H., Stadler, R., and Lo¨rtscher, E. (2016). Nat. Nanotechnol. 11, 170–176. 4. Chen, J., Reed, M.A., Rawlett, A.M., and Tour, J.M. (1999). Science 286, 1550–1552. 5. Collier, C.P., Wong, E.W., Belohradsky, M., Raymo, F.M., Stoddart, J.F., Kuekes, P.J., Williams, R.S., and Heath, J.R. (1999). Science 285, 391–394. 6. Chen, X., Roemer, M., Yuan, L., Du, W., Thompson, D., Del Barco, E., and Nijhuis, C.A. (2017). Nat. Nanotechnol. 12, 797–803. 1IBM
Research – Zurich, Science & Technology Department, Sa¨umerstrasse 4, 8803 Ru¨schlikon, Switzerland *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chempr.2017.08.016
Chem 3, 373–379, September 14, 2017
377
ACKNOWLEDGMENTS This work was supported by the Ministry of Science and Technology of China (grant 2017YFA0204901) and the National Natural Science Foundation of China (grants 21373014 and 21727806). 1. Waldrop, M.M. (2016). Nature 530, 144–147. 2. Aviram, A., and Ratner, M.A. (1974). Chem. Phys. Lett. 29, 277–283. 3. Reed, M.A., Zhou, C., Muller, J., Burgin, T.P., and Tour, J.M. (1997). Science 278, 252–254. 4. Reecht, G., Scheurer, F., Speisser, V., Dappe, Y.J., Mathevet, F., and Schull, G. (2014). Phys. Rev. Lett. 112, 047403. 5. Xiang, D., Wang, X., Jia, C., Lee, T., and Guo, X. (2016). Chem. Rev. 116, 4318–4440. 6. Jia, C., Migliore, A., Xin, N., Huang, S., Wang, J., Yang, Q., Wang, S., Chen, H., Wang, D., Feng, B., et al. (2016). Science 352, 1443–1445. 7. Thiele, S., Balestro, F., Ballou, R., Klyatskaya, S., Ruben, M., and Wernsdorfer, W. (2014). Science 344, 1135–1138. 8. Natterer, F.D., Yang, K., Paul, W., Willke, P., Choi, T., Greber, T., Heinrich, A.J., and Lutz, C.P. (2017). Nature 543, 226–228. 9. Guo, C., Wang, K., Zerah-Harush, E., Hamill, J., Wang, B., Dubi, Y., and Xu, B. (2016). Nat. Chem. 8, 484–490. 10. Zhang, Y., Luo, Y., Zhang, Y., Yu, Y.J., Kuang, Y.M., Zhang, L., Meng, Q.S., Luo, Y., Yang, J.L., Dong, Z.C., and Hou, J.G. (2016). Nature 531, 623–627. 11. Aragone`s, A.C., Haworth, N.L., Darwish, N., Ciampi, S., Bloomfield, N.J., Wallace, G.G., Diez-Perez, I., and Coote, M.L. (2016). Nature 531, 88–91. 12. Wen, H., Li, W., Chen, J., He, G., Li, L., Olson, M.A., Sue, A.C.-H., Stoddart, J.F., and Guo, X. (2016). Sci. Adv. 2, e1601113. 13. Jia, C., and Guo, X. (2013). Chem. Soc. Rev. 42, 5642–5660. 14. Lo¨rtscher, E. (2013). Nat. Nanotechnol. 8, 381–384. 1Beijing
National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, P.R. China *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chempr.2017.08.006
376
CATALYSIS
Reaction: Technological Aspects of Molecular Electronics Emanuel Lo¨rtscher1,* Dr. Emanuel Lo¨rtscher received his PhD with distinction (summa cum laude) in 2006 from the University of Basel in Switzerland and a master in physics in 2003 from the Swiss Federal Institute of Technology. He has been a research staff member at IBM Research since 2008. He has received various prizes, including the Swiss Physical Society Award for Applied Physics in 2007 and the Faculty Prize of the University of Basel in the same year. His current research focuses on fabrication, electronics, optics, and chemistry at the nanoscale for applications in novel computing applications, biology, medicine, and the Internet of Things.
Moore’s Law—or the miniaturization of the complementary metal-oxide semiconductor (CMOS)—is about to end because performance no longer benefits from downsizing. Consequently, semiconductor research now focuses on employing novel materials and architectures to fulfill specifications requested by applications. This situation provides opportunities for fundamentally different electronic components, among which molecular electronics (ME) is considered again.1 Challenges of CMOS Technology Transistor scaling slows down because miniaturization has reached integration densities that cause unbearable thermal loads (averaged 200–300 W/cm2, locally >500 W/cm2), and future nodes are becoming unaffordable2 in terms of fabrication investments. Hence, getting smaller is no longer the maxim.
Chem 3, 373–379, September 14, 2017 ª 2017 Elsevier Inc.
Gate leakage, short-channel effects, and parasitics must be kept low so as not to consume energy when idle, for example. The semiconductor industry has faced similar issues during its life cycle—causing low yields, inhomogeneities in doping levels and driving voltages, reduced reliability, etc.—but it has always been able to overcome them, for example, by Si-on-insulator (SOI), strained Si, fin field-effect transistors (FinFETs), nanowires, high-k dielectrics, airgaps, 3D stacking, multicores, III/V integration, etc. Hence, CMOS will change, but it will remain the predominant technology for the coming years. A successor must not only provide the same specs as International Technology Roadmap for Semiconductors (ITRS) and non-ITRS anticipated CMOS nodes but also outperform prevalent technology in at least one key aspect: power consumption, fabrication costs, and/or performance. The argumentation is therefore not solely technical but also economical to realign a multibillion-USD industry. The successor can leverage CMOS platforms by being co-integrated wherever such novel components outmatch Si but only if they are process compatible. Topdown fabrication can be pushed to <5 nm where quantum effects and uncontrollable doping levels are detrimental for Si. At such dimensions, however, various bottom-up approaches might enter the game. State of the Art in ME The promises of ME are intrinsic functionalities embodied in a molecular backbone, creating electrical responses, e.g., conductance, current rectification, resistance switching, etc. Such ultimately scaled building blocks can be synthesized identically with picometer control. Negative differential resistance, Coulomb and Kondo effects, quantum interference, and switching are only some of the interesting functionalities demonstrated.
Gaining control over these effects, however, is difficult, and large sampleto-sample fluctuations appear. Converging agreement between theory and experimentation has been reached for simple systems; most structure-property relationships, however, are deduced empirically, mostly from large statistical datasets generated by mechanical manipulation, rather than predicted. CMOS in return is predictable because devices can be designed and properties can be tailored by simulations. Besides ME, however, no other nanoelectronics candidate has reached this maturity so far. Compared with Si as the only modestly modified CMOS core material, the limitless library of potentially suited molecular compounds is huge, and there is little consensus on future screening methods. The quantum-mechanical nature of molecules is considered advantages, e.g., as a result of quantum interference. Transport through molecular orbitals, however, is mediated by a variety of mechanisms, all of which are of low conductance, especially in comparison with ballistic transport offered by layered materials. Additionally, most molecular orbitals are misaligned with Fermi energies and require high voltages for resonant transport—the opposite of the ideal switch’s imperatives. Even for fully conjugated molecular wires with hybridized moleculemetal interfaces, molecules must be categorized as insulators rather than conductors, rendering a single-molecule device unviable because heat dissipation will be higher at comparable integration densities. Furthermore, experiments revealing distinct features are mostly done at ambient or cryogenic temperatures. Although there will be exceptions where cooling is affordable, e.g., a quantum computer, most electronics will be operated above 300 K. At typical temperatures of 85 C, many effects discovered in single-molecule junctions will fade out, and compounds can degrade ther-
mally. Last, assembling multiple compounds to a circuit requires site-selective depositions. Although molecular selfassembly and recognition are powerful bottom-up mechanisms, none of these has made it out of the lab. It is therefore premature to discuss any cost advantages. Toward Technology-Relevant Devices With these challenges ahead, why should ME still try to become a nanoelectronics alternative? One motivation arises from novel computing paradigms such as quantum, neuromorphic, and cognitive computing. Among the breadth of molecular compounds, some compounds might be ideally suited for non-linear workflows because of intrinsic sublayers and backpropagation or some supramolecular compounds might have multiple ports and voltage-added responses that behave like synapses. Furthermore, metalorganic compounds could provide abrupt electronic features addressable in two-terminal geometries thanks to redox- and spin-based mechanisms3 while simultaneously aligning molecular orbitals with microelectronics voltage ranges. Besides logic and memory applications, molecules could be also excellent candidates for sensing tasks in emerging Internet of Things (IoT) applications. ME concepts can also be applied in optical or plasmonic hybrids as electrically driven singlephoton emitters or sensitive electro-optical detectors based on molecular transitions. To make that happen, however, we must target methods that reliably and non-destructively integrate molecules into a solid-state platform. Current experiments are mainly based on single-molecule experiments with mechanical actuation. Although those test beds are excellent research tools, their very-large-scale integration is difficult, expensive, and not scalable. Instead, one could start leveraging
early ME platforms4,5 that are technology driven but have fallen into disgrace: nanopores and cross-bar structures with sandwiched molecular ensembles are considered technologically relevant because variations are small as a result of multi-molecule averaging and constrained geometries. Ensemble performances are mostly superior (e.g., for diode rectification ratios 1,0003 larger6) over single-molecule ones because they can benefit from heterogeneous compositions and related effects (image charges, dipole-induced barrier lowering, etc.). Graphene, carbon nanotubes, or low-mobility metal electrodes and direct grafting on semiconductors achieve better controllable and more stable interfaces. Molecules should be assembled in compartments and protected during fabrication and operation against environmental effects and chemical degradation. The architectures mentioned above provide affordable footprints, relevant current levels, methods for site-selective assembly, and compatibility and scalability to wafer-scale integration— a feasible pathway for post-CMOS.
1. Xin, N., and Guo, X. (2017). Chem 3, this issue, 373–376. 2. Cross, T. (2016). After Moore’s Law. The Economist, March 12, 2016. http://www. economist.com/technology-quarterly/201603-12/after-moores-law. 3. Schwarz, F., Kastlunger, G., Lissel, F., EglerLucas, C., Semenov, S.N., Venkatesan, K., Berke, H., Stadler, R., and Lo¨rtscher, E. (2016). Nat. Nanotechnol. 11, 170–176. 4. Chen, J., Reed, M.A., Rawlett, A.M., and Tour, J.M. (1999). Science 286, 1550–1552. 5. Collier, C.P., Wong, E.W., Belohradsky, M., Raymo, F.M., Stoddart, J.F., Kuekes, P.J., Williams, R.S., and Heath, J.R. (1999). Science 285, 391–394. 6. Chen, X., Roemer, M., Yuan, L., Du, W., Thompson, D., Del Barco, E., and Nijhuis, C.A. (2017). Nat. Nanotechnol. 12, 797–803. 1IBM
Research – Zurich, Science & Technology Department, Sa¨umerstrasse 4, 8803 Ru¨schlikon, Switzerland *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chempr.2017.08.016
Chem 3, 373–379, September 14, 2017
377