nature materials | VOL 8 | FEBRUARY 2009 | www.nature.com/naturematerials 91

news & views

Winnik, Manners and co-workers have nicely demonstrated that epitaxial crystallization-driven assembly of micellar structures is not only a living process but, like its covalent counterpart, it allows access to a range of new architectures. The controllability of this approach that allows one to not only easily tailor the length but also the nature of self-assembled nanostructures may well revolutionize approaches to controlled nano-assemblies in the same way the development of living covalent polymerization conditions dramatically impacted the accessibility of complex covalent polymer architectures.

There are of course many questions that still need to be addressed. For example, it does remain to be seen how general this crystallization-driven assembly approach is; will it work with other highly crystalline polymers? But the potential is significant and these initial results are a tantalizing hint at the possibilities for this synthetic assembly tool. ❐

Stuart J. Rowan is in the Department of Macromolecular Science and Engineering, Case Western Reserve University, Cleveland, Ohio 44106-7202, USA. e-mail: stuart.rowan@case.edu

References1. Gädt, T., Ieong, N. S., Cambridge, G., Winnik, M. A. & Manners, I.

Nature Mater. 8, 144–150 (2009).2. Darling, S. B. Prog. Polym. Sci. 32, 1152–1204 (2007).3. Zheng, Y. et al. J. Phys. Chem. B 103, 10331–10334 (1999).4. Israelachvili, J. N. Intermolecular and Surface Forces 2nd edn

(Academic Press, 1991).5. Cornelissen, J. J. L. M. et al. Science 280, 1427–1430 (1998).6. Yu, K. et al. Langmuir 12, 5980–5984 (1996).7. Riegel, I. C. et al. Langmuir 18, 3358–3363 (2002).8. Li, Z. et al. Science 306, 98–101 (2004).9. Pochan, D. J. et al. Science 306, 94–97 (2004).10. Cui, H. et al. Science 317, 648–650 (2007).11. Massey, J. A. et al. J. Am. Chem. Soc. 122, 11577–11584 (2000).12. Wang, X. et al. Science 317, 644–647 (2007).13. Szwarc, M. J. Polym. Sci . A 36, ix–xv (1998).14. Edmondson, S. et al. Chem. Soc. Rev. 33, 14–22 (2004).

in the past decade, extensive research has been undertaken to explore the electron spin degree of freedom for the design

of new electronic devices. This has been motivated by the prospect of using spin in addition, or as an alternative, to charge as the physical quantity carrying information, which may change device functionality to an entirely new paradigm: dubbed spintronics1. One of the key requirements for engineering spintronics devices is the efficient injection

of spin-polarized (SP) charge carriers from a terminal (that is, a ferromagnet (FM) electrode) into a semiconductor interlayer. In addition, efficient devices require a long relaxation time for the spin, which in a non-magnetic interlayer is primarily limited by the spin–orbit and hyperfine interactions2. Organic semiconductors (OSECs) are composed of light elements, such as carbon and hydrogen, which have weak spin–orbit and hyperfine interactions for the relevant

electronic states that participate in the electrical conductance process, and are therefore believed to possess long spin-relaxation times3,4. In addition, OSECs are optically active, which contributes to them being good candidates for spintronics and new magneto-optoelectronic devices5.

In previous work6, an organic diode made of Alq3 (8-hydroxy-quinoline aluminium) — a small, organic luminescent molecule — sandwiched between two

sPiNtRoNics

organics strike backThe spin injection efficiency from a ferromagnetic electrode into an organic layer has been successfully probed by two purpose-made techniques. The observed spin diffusion lengths of tens of nanometres hold promise for potential spintronics applications.

Z. Valy Vardeny

Detector 2

∆B

a bFμ1 Fμ2OSEC

Precession

Positrons1

Positrons2

Spin injection

δBB0

S0; E0

S(t)

z0

λS

Detector 1

SPmuons

ΔP Δ

B

Δ ∼ δB

B0

Figure 1 | Low-energy muon spin rotation technique used by Drew and colleagues9. a, Scheme of the experiment (see text for a detailed description). b, The field δB is formed by the injected spins into the OSEC and is obtained from the probability difference, ΔP, between the measured field distribution with the device current turned on and off, using the skewness parameter, Δ. In the case shown here, the field δB is parallel to B0, and therefore the field distribution has a positive Δ, which corresponds to an increase of the local field caused by the injected spins into the OSEC. Note that P is conserved, since the integral, ∫ΔPdB = 0.

© 2009 Macmillan Publishers Limited. All rights reserved© 2009 Macmillan Publishers Limited. All rights reserved.

92 nature materials | VOL 8 | FEBRUARY 2009 | www.nature.com/naturematerials

news & views

FM electrodes (termed a ‘spin valve’) showed a magnetoresistance (MR) of 40% at low temperature. It was hence concluded that SP carriers are injected from the FM electrode into the OSEC with a relatively high efficiency. However, this was recently challenged7,8 owing to the lack of compelling spectroscopic evidence for SP carrier injection into OSEC materials. On pages 109 and 115 of this issue, articles by Drew et al.9 and Cinchetti et al.10 now report strong evidence for high-efficiency spin injection from a FM electrode into an OSEC layer by using properly designed spectroscopic techniques. The importance of these two experiments is that standard spectroscopic techniques used for detecting SP carrier injection into inorganic semiconductors rely on the existence of a sizeable spin–orbit coupling in the material of interest11,12. The weak spin–orbit coupling in OSECs, that on the one hand makes them so attractive, has therefore presented a real obstacle for the detection of spin injection, severely limiting the advance of organic spintronics.

The technique used by Drew et al.9 is based on the fact that the SP current from a FM–OSEC junction creates a magnetic field, δB, in the OSEC layer, where δB decays at an average distance from the FM–OSEC junction that is related to the spin diffusion length, λS. To measure the decay of δB, Drew et al. used low-energy muon spin rotation (μSR), an important tool for measuring local B in magnetic

materials and superconductors13. The muons carry spin (S0), and on implantation their stopping depth (z0) into the OSEC is directly determined by their given kinetic energy, E0 (Fig. 1a)13. After implantation, the 100% SP muons decay with a lifetime of τ ~2.2 μs, during which their spins precess according to the local field B(z0) in the OSEC layer.

The implanted muons decay into two kinds of neutrinos — which are practically invisible to conventional detectors — and positrons, of which emission direction is correlated with the muons’ spin direction at the time of decay, τ (ref. 13). Therefore, the signal from two detectors placed at the device ends (labelled 1 and 2 in Fig. 1a), which are capable of measuring transient positron flux, may be configured to yield the muons’ spin direction at τ; and consequently B(z0) may be statistically inferred. Drew et al. completed all necessary control experiments for validating the technique in a working organic spin-valve device, and from their measurements concluded that the μSR and MR results are in perfect agreement. The measurements yielded a spin diffusion length of λS ~ 10 nm in Alq3 at 90 K.The μSR and MR experiments were also carried out at various temperatures, T; and surprisingly showed that λS in Alq3 decreases substantially with T (ref. 9.)

In the second experiment reported, Cinchetti et al.10 use the fact that light penetration depth, dL, at an impinging laser energy of ℏω = 3.1 eV is much larger than the electron inelastic mean free

path, λinel, which determines the probing depth of photoemission (Fig. 2a). With this in mind, the authors measured spin-dependent photoemission from a FM–OSEC junction in the form of Co–CuPc (copper phthalocyanine). Direct photoemission cannot occur from the OSEC or Co substrate by absorbing a single photon with the laser energy, ℏω. Thus, two 3.1 eV photons are required to ‘liberate’ an electron from the OSEC, with the photons provided by two synchronized femtosecond pulsed laser beams (Fig. 2a). The CuPc layer thickness, d, on the Co spin-injecting FM substrate was systematically varied, and the photo-electron spin and energy distribution were measured by appropriate spin-dependent detectors.

Two-photon photoemission (TPPE) occurs from a very thin layer, on the order of λinel ~ 1 nm (Fig. 2a). Thus if d > λinel, an SP photo-electron is released from the OSEC by absorbing the second laser pulse only if SP electrons are ballistically injected into the OSEC layer from the Co substrate through the Co–CuPc junction as a result of absorbing the first pulse (Fig. 2b). This provides the necessary tool for detecting injected SP carriers into the OSEC. By measuring spin-dependent TPPE from a Co–CuPc junction of finite d, Cinchetti et al. conclude that the spin injection capability, F, of injecting SP electrons into the OSEC through the FM–OSEC interface is quite high (F ~ 85%). Subsequently, the spin polarization of the injected electrons was measured as a function of d; with which

Figure 2 | Two-photon photoemission method used by Cinchetti and colleagues10. a, Space diagram, (see text for a detailed description). b, Energy diagram. The first absorbed photon creates a hot electron in the Co substrate at energy ℏω – EF (~2.7 eV), which is ~50% spin-polarized. The hot electron balistically reaches the OSEC through the interface dipole layer that is formed at the FM–OSEC junction, where the second absorbed photon re-excites it to generate free SP photoelectrons at an energy of 2ℏω – EF above the vacuum level. The highest occupied and lowest unoccupied molecular orbits (HOMO and LUMO) and EF levels of the OSEC are assigned.

Efinal

ħω

Evac

Photoelectrons

EF

fs pulse1

fs pulse2

dL

Spin injection

ħω

Laser beam 1

Laser beam 2

InterfacedipoleCobalt CuPc

Detector

Cobalt CuPc

S

SS

λ inel

d

Space Energy

LUMO

HOMO

a b

© 2009 Macmillan Publishers Limited. All rights reserved© 2009 Macmillan Publishers Limited. All rights reserved.

nature materials | VOL 8 | FEBRUARY 2009 | www.nature.com/naturematerials 93

news & views

Ma

teRia

l wit

Ness

“Should engineers think?” That question, at first glance mildly insulting to the profession, crops up early in a recent volume of Interdisciplinary Science Reviews devoted to ‘Philosophy and Engineering’ (33(3), 2008). But as Natasha McCarthy of the UK’s Royal Academy of Engineering points out in her editorial introduction, “The stereotype tells us that engineers are ‘doers’ and not ‘thinkers’.”

Leaving aside the flaws in the notion of sharp boundaries between scientists (the ‘thinkers’) and engineers, the message is clear: engineers have a job to do, and ‘philosophy’ is rarely seen as relevant to it. But beyond thinking about the technical aspects of the task, should engineers consider matters of cost, safety, risk, environmental, social, ethical and aesthetic impacts? If they don’t, should this be seen as a form of engineering failure?

Of course, few engineering jobs don’t explicitly embrace some of these concerns. Many materials, for example, are developed specifically to improve safety or to reduce harm to the environment. But even then, the issues are not always straightforward. A biodegradable product can’t be assumed to be ‘greener’ without a life-cycle analysis that might lie

beyond the expertise of the product designers. No one foresaw, or could be expected to foresee, or even now agrees on, the social impact of automobiles, computers and mobile phones.

But truly to engage with some of these matters, or even to make seemingly routine assessments of risk and safety, might require that engineers acknowledge a genuinely philosophical dimension of the profession, particularly in the sphere of epistemology, or as McCarthy puts it, what we know and how we know what we know. Partly this is a question of how one obtains reliable information. But there has been increasing recognition that it also bears on what is knowable. Some complex engineered systems show emergent phenomena that can’t be predicted, even in principle, from knowledge of the components. Dealing with that challenge is considered here by W. P. S. Dias of the University of Moratuwa in Sri Lanka, and by Darryl Farber and colleagues at Pennsylvania State University. A related difficulty tackled by the latter group is how to meaningfully forecast performance in the face of incomplete knowledge about what the system will encounter.

Epistemology aside, much of engineering philosophy might be considered to encompass ethics. Here McCarthy takes a stance: an engineer cannot be expected to reflect on aspects beyond the call of duty, which basically means getting the job done efficiently, effectively and safely. Indeed, it might be dangerous to do so without the necessary expertise. But that, she says, doesn’t absolve engineers of moral obligations — for they have a duty to ensure that such broader issues have been given due thought by others suitably qualified.

This sound principle might quickly become a minefield in practice. How can engineers assure themselves that this process has been carried out, and done well? How is the obligation enforced? (Clearly, it isn’t.) In this and other ways, this volume is just the start of the discussion. ❐

What’s philosophy got to do With it?

PHILIP BALL

it was hence found to decrease. From the obtained dependence on d, the λS, at 2.4 eV above the highest occupied molecular orbital level of the OSEC was determined to be ~13 nm in CuPc at room temperature.

The works of Drew et al.9 and Cinchetti et al.10 both attest that SP carriers are indeed injected in FM–OSEC junctions, thus validating previous experiments performed on organic spin valves using MR(B0) dependence6. It should be said, however, that the two techniques described are rather complicated, require sophisticated analysis to obtain meaningful results and, most importantly, are not ‘table-top’ type techniques. To obtain a technique comparable in accessibility to those used for inorganic materials11,12 — based, for example, on the detection of circularly polarized electroluminescence — it is essential to find an OSEC material with a sufficiently strong spin–orbit interaction

that allows implementation of such circular polarization spectroscopic techniques. At present, however, this would mean that such an organic material, if found, would have a limited λS owing to the unavoidable enhanced spin–orbit coupling. A solution would be to find a material that also has high carrier mobility, which by compensating the spin relaxation due to the spin–orbit coupling would yield λS on the order of tens of nanometres at room temperature.

A final point that should be mentioned is that the two works raise important questions about the temperature dependence of λS in OSEC. Namely, what is the origin of the spin–flip processes that were found? And what is the temperature dependence of these processes? These questions are crucial, and may determine the ability to use OSEC spintronics devices at room temperature5,14. ❐

Z. Valy Vardeny is in the Department of Physics, University of Utah, Salt Lake City, Utah 84112, USA. e-mail: val@physics.utah.edu

References1. Wolf, S. A. et al. Science 294, 1488–1495 (2001).2. Zutic, I., Fabian, J. & Das Sarma, S. Rev. Mod. Phys. 76, 323-410 (2004).3. Dediu, V. et al. Solid State Commun. 122, 181–184 (2002).4. Pramanik, S. et al. Nature Nanotech. 2, 216–219 (2007).5. Sanvito, S. Nature Mater. 6, 803–804 (2007).6. Xiong, Z. H., Wu, D., Vardeny, Z. V. & Shi, J. Nature

427, 821–824 (2004).7. Xu, W. et al. Appl. Phys. Lett. 90, 072506 (2007).8. Jiang, J. S., Pearson, J. E. & Bader, S. D. Phys. Rev. B

77, 035303 (2008).9. Drew, A. J. et al. Nature Mater. 8, 109–114 (2009).10. Cinchetti, M. et al. Nature Mater. 8, 115–119 (2009).11. Kikkawa, J. M. & Awschalom D. D. Nature

397, 139–141 (1999).12. Crooker, S. A. et al. Science 309, 2191–2196 (2005).13. Bakule, P. & Morenzoni, E. Contemp. Phys. 45, 203–225 (2004).14. Yunus, M., Ruden, P. P. & Smith D. L. Appl. Phys. Lett.

93, 123312 (2008).

© 2009 Macmillan Publishers Limited. All rights reserved© 2009 Macmillan Publishers Limited. All rights reserved.