Graphene heterostructures for spin and charge transport
Zomer, Paul Joseph
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Zomer, P. J. (2019). Graphene heterostructures for spin and charge transport. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
8
Chapter 8
Robust spin transport in highly irradiated
few-layer graphene
Abstract
Few layer graphene proves to be an excellent channel for spin transport, even when sus-taining heavy damage to its lattice by H+ ion irradiation. Up to doses of 1x1016cm−2
, at an energy of 40 keV, micrometer scale spin relaxation lengths can still be achieved at room temperature. For the highest dose, we observe a strong enhancement of spin lifetime and suppression of spin diffussion coefficient at low temperature, consistent with the creation of single point defects and their associated localized magnetic moments. The robustness of the transport characteristics in few layer graphene make it therefore an interesting particle-radiation-hard material for practical spintronic applications under extreme con-ditions, such as in particle accelerators and aerospace.
In preparation as: P.J. Zomer, O. Lehtinen, A.V. Krasheninnikov, B.J. van Wees, I.V. Grigorieva and I.J. Vera-Marun, Robust spin transport in highly irradiated few-layer graphene.
8
8.1
Introduction
Graphene has become a state-of-the-art material that is nonetheless widely accessible to reasearchers and has found its way into many applications [1]. The reason for this is a remarkable set of characteristics, such as its strength, high conductance, impermeability, optical properties and single atom thickness [2]. One interesting application of graphene is as the channel in a spintronic device [3]. Here graphene has proven itself as an excellent material, with the capability of transporting spin information over micrometer length scales at room temperature [4], opening the path to practical spintronic applications.
That graphene only comprises a single atomic surface makes it also vulnerable to damage. Adsorbates and defects may adversely affect the properties of graphene de-vices, but also offer functionalization enabling graphene as an interesting candidate for sensor applications. For example, single point defects in graphene induce mag-netic moments that interact with electron spins [5], allowing for their detection via spin transport measurements [6, 7]. The vulnerability of an exposed surface can be prevented by encapsulation in van der Waals heterostructures [8]. This approach has led to graphene devices with very large electron mobility, which nonetheless might take a significant effort to translate into practical applications. Alternatively, this ap-proach may not be required when using few layer graphene. Besides monolayers, few layer graphene (FLG) has also demonstrated to be a capable spin transport chan-nel [9], although the number of studies is limited. The multilayer structure of FLG adds to its robustness, which is particularly relevant for reliable operation under demanding environments.
In this work we focus on spin transport in few layer graphene devices that were subjected to H+ ion irradiation, with doses ranging from 1x1014 cm−2 to 3x1016 cm−2. This approach allows to address the role of single point defects created via irradiation on both charge and spin transport. Whereas measurements on single layer graphene require the irradiation to be done in situ, to avoid exposure to ambi-ent conditions of the highly reactive point defects, in FLG the outer layers shield the point defects in the inner layers [10] enabling their study after exposure to ambient conditions. Charge transport measurements combined with Raman spectroscopy give a clear indication of the presence of defects in FLG. Remarkably, spin transport through the defective FLG is shown to be very robust and is only inhibited at the highest irradiation dose of 3x1016cm−2used in this study. Furthermore, we observe an enhancement of the effective spin lifetime at low temperature, indicative of the presence of localized magnetic moments associated to the defects. This demonstrates that despite taking heavy irradiation damage, FLG based spintronic devices retain their functionality, an important realization towards their practical application.
8
8.2. Fabrication 107
Figure 8.1: a) Optical micrograph of a typical few layer graphene device. The FLG flake has
been irradiated with the highest dose used in this study (3x1016cm−2). The scale bar is 20
µm. b) Schematic top view of a device showing the FLG (black) with contacts (grey) and con-tact wiring for charge transport (left side) and spin transport (right side) measurements. c) Square resistance as function of back gate for all irradiation doses used in his study, measured at room temperature (solid lines) and 4.2 K (dashed lines). d) Square resistance per layer mea-sured at Vbg= 0V as function of number of layers in the FLG. Solid symbols represent room
temperature measurements, open symbols with dashed lines low temperature measurements.
8.2
Fabrication
As a first step in our fabrication process, few layer graphene (FLG) flakes are me-chanically exfoliated from highly oriented pyrolithic graphite onto a highly doped
8
silicon wafer with a 500 nm SiO2layer. Flakes selected by optical contrast and atomic force microscopy, which allow to identify the number of layers in FLG with mono-layer accuracy, were irradiated using protons. The H+ions had an energy of 40 keV, with doses set within the range 1014to 1016cm−2 in order to induce single vacan-cies [11]. In particular, doses of 1x1014 cm−2, 1x1015 cm−2, 1x1016 cm−2 or 3x1016 cm−2were used, and for comparison an additional group was not irradiated. The flakes were subsequently fabricated into spintronic devices, where the FLG is used as a spin transport channel, electrically addressed via magnetic tunnel contacts based on cobalt electrodes with a sub-nanometer aluminum oxide tunnel barrier, and the conducting silicon substrate used as an electrostatic backgate [9].
8.3
Measurements
The electronic measurements were done using a low frequency lock-in technique with currents up to 2 µA. The devices are kept under vacuum in a flow cryostat for measurement. The cryostat is placed between the poles of a rotatable magnet. Two distinct measurement schemes are used, as depicted in Figure 8.1(b). First, a local four terminal geometry, used to extract the square resistance Rsq = VlL/IlW between contacts 2 and 3, where L and W are the respective length and width of this region. Second, a four terminal non-local geometry, used specifically for spintronic measurements [12]. The voltage Vnlis detected between contacts 1 and 2, outside the path of the charge current Iinjbetween contacts 3 and 4. This allows for the detection of spins diffusing across the device, via the non-local resistance Rnl = Vnl/Iinj. This geometry detects pure spin current while excluding spurious charge contributions such as Hall effects.
8.3.1
Charge transport measurements
The effect of the irradiation is readily observed in charge transport measurements. The square resistance as function of applied back-gate bias (Vbg) is determined at room and at liquid helium temperatures. The square resistance is shown in Figure 8.1(c) for a selection of devices subject to different doses. For non-irradiated devices we observe the resistance maximum at 0 Vbg, as expected for intrinsic graphene. After the irradiation, the resistance is affected in two ways. First, the resistance max-imum no longer appears to be at 0 Vbg, indicating extrinsic doping. For the lower irradiation doses of 1x1014cm−2and 1x1015cm−2the resistance maximum shifts to negative Vbgvalues, whereas for higher doses it is observed to shift to positive Vbg values. Second, the resistance continuously increases from the lowest to the highest irradiation dose. This trend is clearly observed in Figure 8.1(d), which shows the square resistance per layer, at 0 Vbg, for a larger number of devices with different
8
8.3. Measurements 109
thicknesses.
Previous work on monolayer graphene irradiated with Ne+ ions also showed a continuous increase in resistance with increasing irradiation dose, while the re-sistance maximum gradually shifted to higher positive Vbg [13]. Importantly, our devices have been exposed to air between the irradiation step and measurements, allowing the damaged top surface to react with the environment, as expected in ap-plications. As a consequence we relate our results to the competing contributions between surface defects or adsorbates on the top layer of the FLG, and lattice defects lying further within the FLG. We expect adsorbates and surface damage to domi-nate at low irradiation doses, whereas higher doses will be domidomi-nated by formation of vacancies in the inner layers. This interpretation is supported by our observed shift in resistance maximum to higher positive Vbg for higher doses, as previously observed in monolayer graphene [13].
8.3.2
Spin transport measurements
Next we turn our attention to spin transport. The current injected into FLG, Iinj (Fig-ure 8.1(b), is spin polarized due to the spin imbalance in the ferromagnet. If this imbalance does not fully relax before reaching the detector contact, we can measure the non-local resistance Rnl. By sweeping the value of an external magnetic field par-allel to the contacts, we can control the relative magnetization orientation between the injector and the detector. The anti-parallel and the parallel states show up as two distinct Rnlvalues, with their difference yielding the spin resistance, ∆Rnl. A set of such spin valve measurements for various irradiation doses is shown in Figure 8.2(a). Even up to a dose of 1x1016cm−2, we clearly observed the switching between parallel and anti-parallel configurations, with a typical response of ∆Rnl≈ 1 Ω. This demonstrates successful spin injection, transport and detection in highly irradiated FLG. For a dose of 3x1016cm−2we could no longer detect spin transport in any of the five measured devices.
In order to gain further insight we performed spin precession (Hanle) measure-ments, by applying a magnetic field perpendicular to the FLG. By varying the mag-netic field, the average spin projection at the detector is modulated, showing the effects of coherent spin precession plus the decay induced by diffusive broadening and spin relaxation. Representative precession curves are shown in Figure 8.2(b), fitted with the one-dimensional Bloch equation [12]. This allows for the extraction of the contact polarization, P , spin diffusion constant, Ds, spin relaxation time, τ , and spin relaxation length, λ =√Dτ [4]. Hence these curves give a detailed insight into two separate processes: the spin injection and detection at the contacts (P ) and the spin transport properties of the FLG channel (Ds, τ, λ). To clarify the role played by the irradiation-induced single point defects, it is key to study the evolution of spin transport with irradiation dose. In Figure 8.3 we summarize all spin transport
8
parameters extracted over several devices.
Figure 8.2: a) Room temperature spin valve measurements for the various irradiation doses
used; starting from 0 dose going up, lengths/width and number of layers are: 6 µm/5.35 µm 7, 7 µm/2.8 µm 7, 3 µm/1.25 µm 4 and 3 µm/1.7 µm 5. Arrows underneath the spin valves for 0 dose indicate the magnetic field sweep direction. b) Hanle precession data and fits at room and low temperature, for irradiation doses of 1x1014cm−2and 1x1016cm−2at 0 Vbg. Plots are
normalized with respect to the maximum of each respective fit. The spacing between injector and detector was 3 µm in both cases. Room temperature data is offset by 2 with respect to LT data.
8
8.4. Discussion 111
8.4
Discussion
Let us start discussing the results obtained at room temperature, presented as the red data points in Figure 8.3(a–d). The spin diffusion constant, Ds, in Figure 8.3(a) exhibits values on the order of 10−2 m2 s−1, from the pristine devices up to the intermediate dose of 1015cm−2, with a strong decrease by one order of magnitude for the largest dose where we were able to observe spin transport, 1013cm−2. These results, in agreement with the charge diffusion constant, Dc, are direct proof that large irradiation dose leads to the formation of single point defects that introduce intervalley scattering [13]. Interestingly, the irradiation damage observed from Ds seems to be limited for low to medium doses. The room temperature spin relaxation time, shown in Figure 8.3(b), covers the range of 100 ps to 1 ns with a significant device to device variation, typical for (few layer) graphene devices supported on SiO2[4, 9], and no definite trend with irradiation dose. The spin relaxation length in Figure 8.3(c) remains in a favorable range of 1–4 µm, with a downward trend primarily determined by Ds. Remarkably, we still observe spin lifetimes close to 1 ns even for high irradiation, where we had observed a reduced Ds. This trend is also seen in the Ds and τ for the Hanle curves in Figure 8.2(b). Therefore, single point defects, while limiting charge diffusion, do not hinder the spin lifetime in FLG at room temperature.
Another technologically relevant observation in Figure 8.3(d) is that the con-tact polarization, P , at room temperature is significantly enhanced (up to 30%) for low and intermediate doses, and remains comparably high (10%) at high irradiation dose. This is also consistent with the spin valve results in Figure 8.2(a), where the spin signals for low and intermediate irradiation doses are one order of magnitude larger. We associate this increase in P with an improved structural quality of our aluminum oxide tunnel barrier, leading to a higher tunnel spin polarization. This is rationalized by the damage created at the surface of FLG, where defects on the sur-face pose as anchoring points to form a more uniform aluminum film, and therefore a pinhole-free oxide barrier [14, 15]. We note that a creation of surface defects for the initial stages of irradiation at low doses, rather than bulk defects, offers a consistent picture to understand the trend observed in both P and Ds. The achievement of a large P , comparable with state of art results in monolayer graphene [14], and the permanence of favorable values of Dsand τ for low and intermediate irradiation, of-fers the promise of FLG as a robust and practical material [15] for room temperature spintronics.
Further insight into device performance and its microscopic origin is gained by looking at the temperature dependence of the spin transport. Here we compare the extracted spin transport parameters obtained at 4.2 K (LT), presented as the blue data points in Figure 8.3(a–d), with the ones at room temperature (RT), in particu-lar we consider the ratio between these measurements as shown in Figure 8.3(e–h).
8
0 0 10-3 10-2 10-1 RT LT Ds (m 2 /s ) 0 1 2 Dc Ds D4K /DR T 102 103 104 � (p s) 0 1 2 3 4 �4K / �RT 0 2 4 � ( � m ) 0 1 2 �4K / �RT 1014 1015 1016 0 20 40 60 P (%) Dose (cm-2) 1014 1015 1016 0 1 2 P4K /PR T Dose (cm-2)a)
b)
c)
d)
e)
f)
g)
h)
Figure 8.3: a), b), c) and d) show an overview of the spin transport data; e), f) g) and h) show ratios of low temperature over room temperature data for respectively the diffusion constants, spin relaxation time, spin relaxation length and contact polarization as function of irradiation dose. The dashed lines pass through the average values, the bars show the standard deviation.
For pristine FLG devices and for low and medium irradiation doses, no significant change was observed when comparing room temperature data to low temperature data, with the ratio for all figures of merit mostly within the range 0.5–1.5. This sce-nario of temperature independent parameters is consistent with previous work on pristine FLG [9]. Interestingly, for high irradiation dose we observe that Dsand τ are significantly and oppositely affected by temperature. Here, Dsdecreases from room
8
8.4. Discussion 113
temperature to low temperature, whereas τ increases at low temperature. The re-sult is a set of inversely related ratios of τLT/τRT∼ Ds,RT/Ds,LT∼ 3, evidenced in the highlighted regions in Figure 8.3(a,b). This opposing behavior leads to a temperature independent λ for all doses (see Figure 8.3(g)).
The distinct behavior of Dsand τ for high irradiation dose may be associated with the role of magnetic moments present at single point defects [5]. Within this picture, each defect contains a paramagnetic moment that is partially oriented by an exter-nal magnetic field only at low temperature, which is reflected in Hanle precession experiments as an effective increase at low temperature of the g-factor [6, 7, 16]. Our data is consistent with this hypothesis. In particular, note that the ratios observed for the spin lifetime (see Figure 8.3(f)) show narrow distributions of τLT/τRT∼ 1 for both pristine and low irradiation devices, plus a gradual increase of this ratio to ∼ 1.4 for medium irradiation up to its value for high irradiation. This consistent trend is sig-nificant when compared to the device-to-device variation observed in Figure 8.3(b). Furthermore, in Figure 8.3(e) we have also included the ratio for the charge diffusion constant, Dc, which shows a value of 0.5–1.5 for the whole range of irradiation doses, contrary to the behavior of Ds. This significant difference between the behavior of Dc and Ds, in particular at high irradiation dose as highlighted in Figure 8.3(e), is used in the literature to account for the enhancement in g-factor [7, 17] and provides further evidence for the presence of magnetic moments in highly irradiated FLG.
Overall, the practical performance of the FLG spintronic devices is independent of temperature and surprisingly robust. We observed sizeable values of λ & 1 µm and P & 5% over the broad range of irradiation doses explored here. This is an im-portant realization, when considering specialized applications under extreme condi-tions where heavy irradiation is present. Ultimately, excessive irradiation prevents the detection of spin signal in our devices, but not until reaching a dose of 3x1016 cm−2. Note that this is hampered by the diffusion process, as a combination of a short relaxation length and a highly resistive (noisy) spin channel. Still, this does not rule out the presence of a significant spin lifetime even at such a high irradiation dose.
To further support the notion that defects were introduced in the FLG, Raman spectra were taken using a 633 nm laser for various devices, after measuring their transport properties. A comparison between the three highest doses is shown in Figure 8.4(a). The ratio between the IDand IG peaks has been used to quantify the defect density in graphene [18, 19]. We observed an increase of ID/IGwith increasing irradiation dose (see Figure 8.4(b)), providing support for the successful creation of defects. Molecular dynamics (MD) simulations were used to quantify the probability of creating a defect for the energy used in this work. The best agreement between the simulation and measured data is obtained when scaling the defect density down by a factor 20. A possible explanation for discrepancy between MD simulations and the Raman results is the recombination of defects. To gain information about the nature
8
Figure 8.4: a) Raman spectra for the 3 highest irradiation doses used. The spectra are scaled
to have identical intensities for the G band and the number of layer was 5 for all flakes. b) Ratio of the intensities of the D and G bands as function of irradiation dose. The inset shows ID/IGanalysis used to estimate the defect density (include fits in main figure).
of the defect (sp3 vs vacancy) with Raman, we use the ratio of peaks D/D0, which decreases for vacancies to values of ∼ 7 [20] and can go down to ∼ 2 [21]. For the highest two doses used in this study we obtain a ratio ∼ 3, indicating the presence of point defects.
8
8.5. Conclusion 115
8.5
Conclusion
In conclusion we have shown that the capability for spin transport in few layer graphene is very robust with respect to irradiation damage to the flake, which is an important find for the development of graphene based spin-logic devices. Like single or bilayer graphene, FLG is an excellent medium for spin transport. How-ever, single and bilayer graphene proved more vulnerable to irradiation, evidencing enhanced signatures of damage and lack of spin signal.The outer layers in FLG ef-fectively shield the inner layers from the environment, allowing for devices that are less sensitive to the environment, and are more practical for applications than the more complex hBN/graphene heterostructures. Even highly defective FLG devices are still capable of transporting spin information over micrometer length scales with nanosecond lifetimes. The latter offers a practical platform to study single point de-fects, and their associated magnetic moments, while shielded by the outer layers in FLG.
8.6
Acknowledgments
We acknowledge B. Wolfs, J. G. Holstein and H. M. de Roosz for their technical assis-tance. This research has received funding from the European Unions Horizon 2020 research and innovation programme under grant agreements No 696656 and 785219 (Graphene Flagship Core 1 and Core 2), the Netherlands Organization for Scientific Research (NWO), the Zernike Institute for Advanced Materials and by the Dutch Foundation for Fundamental Research on Matter (FOM).
8
Bibliography
[1] A. C. Ferrari, F. Bonaccorso, V. Fal’ko, K. S. Novoselov, S. Roche, P. Boggild, S. Borini, F. H. L. Koppens, V. Palermo, N. Pugno, J. A. Garrido, R. Sordan, A. Bianco, L. Ballerini, M. Prato, E. Li-dorikis, J. Kivioja, C. Marinelli, T. Ryhanen, A. Morpurgo, J. N. Coleman, V. Nicolosi, L. Colombo, A. Fert, M. Garcia-Hernandez, A. Bachtold, G. F. Schneider, F. Guinea, C. Dekker, M. Barbone, Z. Sun, C. Galiotis, A. N. Grigorenko, G. Konstantatos, A. Kis, M. Katsnelson, L. Vandersypen, A. Loiseau, V. Morandi, D. Neumaier, E. Treossi, V. Pellegrini, M. Polini, A. Tredicucci, G. M. Williams, B. Hee Hong, J.-H. Ahn, J. Min Kim, H. Zirath, B. J. van Wees, H. van der Zant, L. Occhip-inti, A. Di Matteo, I. A. Kinloch, T. Seyller, E. Quesnel, X. Feng, K. Teo, N. Rupesinghe, P. Hakonen, S. R. T. Neil, Q. Tannock, T. Lofwander, and J. Kinaret, Nanoscale 7, 4598 (2015).
[2] A. K. Geim and K. S. Novoselov, Nature Materials 6, 183 (2007).
[3] W. Han, R. K. Kawakami, M. Gmitra, and J. Fabian, Nature Nanotechnology 9, 794807 (2014). [4] N. Tombros, C. Jozsa, M. Popinciuc, H. T. Jonkman, and B. J. van Wees, Nature 448, 571 (2007). [5] M. B. Lundeberg, R. Yang, J. Renard, and J. A. Folk, Phys. Rev. Lett. 110, 156601 (2013).
[6] K. M. McCreary, A. G. Swartz, W. Han, J. Fabian, and R. K. Kawakami, Phys. Rev. Lett. 109, 186604 (2012).
[7] B. Birkner, D. Pachniowski, A. Sandner, M. Ostler, T. Seyller, J. Fabian, M. Ciorga, D. Weiss, and J. Eroms, Phys. Rev. B 87, 081405 (2013).
[8] A. K. Geim and I. V. Grigorieva, Nature 499, 419 (2013).
[9] T. Maassen, F. K. Dejene, M. H. D. Guimar˜aes, C. J ´ozsa, and B. J. van Wees, Phys. Rev. B 83, 115410 (2011).
[10] O. Lehtinen, I.-L. Tsai, R. Jalil, R. R. Nair, J. Keinonen, U. Kaiser, and I. V. Grigorieva, Nanoscale 6, 6569 (2014).
[11] R. R. Nair, M. Sepioni, I.-L. Tsai, O. Lehtinen, J. Keinonen, A. V. Krasheninnikov, T. Thomson, A. K. Geim, and I. V. Grigorieva, Nat Phys 8, 199 (2012).
[12] F. J. Jedema, H. B. Heersche, A. T. Filip, J. J. A. Baselmans, and B. J. van Wees, Nature 416, 713 (2002). [13] J.-H. Chen, W. G. Cullen, C. Jang, M. S. Fuhrer, and E. D. Williams, Phys. Rev. Lett. 102, 236805
(2009).
[14] W. Han, K. Pi, K. M. McCreary, Y. Li, J. J. I. Wong, A. G. Swartz, and R. K. Kawakami, Phys. Rev. Lett. 105, 167202 (2010).
[15] B. Dlubak, P. Seneor, A. Anane, C. Barraud, C. Deranlot, D. Deneuve, B. Servet, R. Mattana, F. Petroff, and A. Fert, Appl. Phys. Lett. 97, 092502 (2010).
[16] T. Maassen, J. J. van den Berg, E. H. Huisman, H. Dijkstra, F. Fromm, T. Seyller, and B. J. van Wees, Phys. Rev. Lett. 110, 067209 (2013).
[17] M. Wojtaszek, I. J. Vera-Marun, T. Maassen, and B. J. van Wees, Phys. Rev. B 87, 081402 (2013). [18] M. M. Lucchese, F. Stavale, E. H. M. Ferreira, C. Vilani, M. V. O. Moutinho, R. B. Capaz, C. A. Achete,
and A. Jorio, Carbon 48, 1592 (2010).
[19] L. G. Cancado, A. Jorio, E. H. M. Ferreira, F. Stavale, C. A. Achete, R. B. Capaz, M. V. O. Moutinho, A. Lombardo, T. S. Kulmala, and A. C. Ferrari, Nano Lett. 11, 3190 (2011).
[20] A. Eckmann, A. Felten, A. Mishchenko, L. Britnell, R. Krupke, K. S. Novoselov, and C. Casiraghi, Nano Lett. 12, 3925 (2012).
[21] A. Zandiatashbar, G.-H. Lee, S. J. An, S. Lee, N. Mathew, M. Terrones, T. Hayashi, C. R. Picu, J. Hone, and N. Koratkar, Nat. Commun. 5 (2014), 10.1038/ncomms4186.
