The aim

The main aim is to obtain the map of spin distribution on the surface and to characterize the spin properties of selected nanoscale features of surface.

The ideas

Now it is possible to formulate three ideas how (in principle) to measure the spin orientation of atoms on surface.

• Using a magnetic (ferromagnetic or antiferromagnetic) tip for magnetic material surfaces. By comparison of STM images obtained with magnetic and nonmagnetic tips to make a conclusion about spin orientation on the surface. Another possibility to study the surface magnetic features variable in space (domain walls?) by only the magnetic tips.
• Using the III-V semiconductor tips (GaAs, InAs, etc.) excited by circularly polarized light for magnetic material surfaces. In this case it is able to change the spin orientation of electrons in the tip by change of helicity of circular polarization of excited light. The tunnel current is dependent on orientation of electron spin and has to show the dependence on helicity of light polarization. In this sense no need to compare the STM images obtained by different tips (magnetic and nonmagnetic) to make a conclusion on surface spin effect.
  Using the circular polarization of recombination radiation from the semiconductor tip appeared as a result of tunneling of spin-oriented electrons from magnetic surface.
• The third possibility is based on nonmagnetic scanning tunneling spectroscopy of magnetic material surfaces. It is reduced one as a magnetic sensitive but can be useful to spatially resolve the spectroscopic features of magnetic materials (as rear earth compounds, for example) predicted from a theory or already obtained by other (non spatially resolved) methods.

Magnetic tip/magnetic surfaces.

• R. Wiesendanger , H.-J. Guntherodt, G. Guntherodt, R.J. Gambio and R.Ruf. "Observation of vacuum tunneling of spin-polarized electrons with the Scanning Tunneling Microscope". PRL 65 (1990) 247.

In this paper they obtain the difference in atomic steps height on antiferromagnetic ordered planes (100) of Cr(100) measured by ferromagnetic CrO2 -tip and nonmagnetic (W) one: 1.2 and 1.6 A by CrO2- tip, and 1.4 A by W-tip. ( From my conversation with Dr. I. Shvets from Dublin Trinity Colledge in 1996: "Nobody can reproduce this result")

• R. Wiesendanger , I.V.Shvets, D.Burgler, G.Tarrach, H.-J. Guntherodt, and J.M.D. Coey. "Evidence of selective imaging of different magnetic ions on the atomic scale by using a scanning tunneling microdcope with ferromagnetic probe tip". Europhysic.Lett. 19 (1992) 141-146.

They observed the difference in STM images of Fe3O4 obtained with Fe- and W-tips. For W-tip only atomic rows were in images, no corrugation along the rows; for Fe-tip a strong (comparable with inter-row corrugation) atomic corrugation was obtained along the row. They interpreted this phenomenon as interaction of spin moments of Fe2+ and Fe3+ ions in Fe3O4 and Fe-tip.

My request to Dr. I. Shvets on 19 March 1998 to comment the situation in this field of research was answered that they not published anything in this field since 1995 and he is not aware of any recent breakthroughs by other groups either. Limitation Since 1990 there have been published only two papers in this direction. The main limitation of this approach is the need to compare the STM images recorded both by magnetic and nonmagnetic tips. In this sense the mentioned papers has a demonstrative character only. It seams as not convenient approach for real use in SP-STM measurements. Perhaps it could be applied for surface objects with spatially variable magnetic moments like domain walls or something similar without comparison magnetic -nonmagnetic tip images.

III-V semiconductor tips (GaAs, InAs, etc.) excited by circularly polarized light/magnetic surfaces

Semiconductor tips (GaAs, InAs, etc.) excited by circularly polarized light/magnetic surfaces (Direct scheme A).

• G.Nunes and N.M. Amer. "Atomic resolution scanning tunneling microscopy with a gallium arsenide tip" Appl.Phys.Lett. 63 (1993) 1851.

Atomic resolution STM images recorded with cleaved GaAs tips were demonstrated for Si(111)-(7x7) surface in ultrahigh vacuum. The tips were cleaved from single-crystal (100) oriented 450 mm wafers along (110) direction in air. The tip size was 2mm. The tip apex radius was smaller than 100 nm (SEM resolution limitation) . In vacuum tips were prepared by heating at 630 oC. The sample was biased at +2 V, tunnel current was 1 nA.

• M.W.Prince, M.C.M.van der Wielen, R.Jansen, D.L.Abraham and H.van Kempen. Photoamperic probes in scanning tunneling microscopy. APL 64 (1994) 1207.
• M.W.Prince, R.H.M. Groeneveld, D.L.Abraham, H.van Kempen and H.W. van Kesteren. APL 66(1995) 1141.
• M.W.Prince, R.Jansen, and H.van Kempen. "Spin polarized tunneling with GaAS tips in scanning tunneling microscopy" PR B53 (1996) 8105-8113. Ph.D.Thesis of M.Prince "Magnetic imaging with photoexcited semiconductor tips in scanning tunneling microdcopy" , 1995.

They used GaAs tip and Pt/Co multi-layer film deposited on a glass. The tip was made by cleavage (100) wafers in (110 and 1-10) directions. The light excitation was directed from the tip side with angle of 45a°. 1.96 eV He-Ne laser. The polarization of photocurrent from the tip was estimated as 10%. The amplitude of modulated current was of 4 pA in comparison of total tunnel current 1 nA or 0.4% polarized.

• M.W.Prince, R.Jansen, and H.van Kempen. "Theory of spin polarized transport in photoexcited semiconductor/ferromagnet tunnel junctions" PR B57 (1998) 4033-4047.

The analysis of possible phenomena in tunnel junction and in semiconductor tip itself under polarized light illumination has been done. The most convenient regime for tunnel current polarization could be obtained at the following conditions:

• forward spin polarized electron flow from semiconductor to ferromagnetic, Vm<0
• at smaller tunnel barrier height (thickness); it means at larger tunnel current (Vt = constant)
• at higher light excitation
• at smaller amount of surface states in tip (clean and without defects tip) or larger surface spin relaxation time Tss
• for pure semiconductor tip, large bulk spin relaxation time Ts
• at lower temperatures, larger Tss, Ts

Part of these conclusions has been already confirmed in experimental works of Prins et al.(1995, 96)

• Y. Suzuki, W. Nabhan, K. Tanaka. "Magnetic domains of cobalt ultrathin films observed with a scanning tunneling microscope using optically pumped GaAs tips" APL, 71 (1997) 3153.

They performed experiments with GaAs tips under illumination by circularly polarized light (5 mW He-Ne laser , 633 nm/1.96 eV, 10% polarized) in UHV at room temperature. Ferromagnetic ultrathin Co layers with perpendicular magnetization were employed as a sample. The light gone from the side of transparency sample. The measured modulated current shown (-2 pA) and +2pA for the locations with different direction of magnetization at tunnel parameters (-1 V, 1 nA). The current polarization was 0.2% for selected spin orientation. The size of magnetization location is 200-500 nm typically. The MFM measurements of the same sample showed the similar result.

III-V semiconductor sample (GaAs, InAs, etc.) excited by circularly polarized light/magnetic tip (Direct scheme B).

• K.Sueoka, K. Mukasa and K.Hayakawa. Possibility of observing spin-polarized tunneling current using scanning tunneling microscope with optically pumped GaAs. Jpn.J.Appl.Phys. 32 (1993) 2989-2993.

GaAs thin film sample pumped by circularly polarized light (from back side of the sample) and a ferromagnetic polycrystalline Ni tip were used. The tunnel current was perturbed by modulating the power and polarization of the pumping light. The perturbation arise due to three dominant effects: the thermal expansion of tip - surface junction, the variation in excited carriers concentration in GaAs and spin-polarized tunneling effect. The spin polarized effect was distinguished by observing the current dependence on the bias. The experiment was performed in air. Wavelength was 830 nm or 1.49 eV(single-mode LD, 30 mW), the tip was magnetized in H=1T ex situ. Tip bias was -1.2 V, It=1 nA. Current polarization was 10% in average. The observed tunneling current modulation depends upon both helicity of circular polarization and tip magnetization. The estimated spin polarization of the tip was negative.(?)

Measure of circular polarization of recombination radiation from the semiconductor tip appeared as a result of tunneling of spin-oriented electrons from magnetic surface. (Reverse scheme A)

No papers.

Measure of circular polarization of recombination radiation from the semiconductor sample appeared as a result of tunneling of spin-oriented electrons from magnetic tip. (Reverse scheme B)

• S.F.Alvarado and P.Renaud. "Observation of spin-polarized tunneling from ferromagnetic into GaAs. PRL 68 (1992) 1387.

They used ferromagnetic Ni-tip for GaAs(110) surface tunneling study in UHV. Magnetization of tip was made in situ by means of small electromagnet . As a measuring value they analyzed the circular polarization of the recombination luminescence excited by electrons tunneling from a ferromagnetic Ni tip. The degree of negative spin polarization of electron extracted from tip was large, Pm= -31%. The degree of polarization decrease with increase of kinetic energy of injected electrons.

• S.F. Alvarado. "Tunneling barrier dependence of electron spin polarization" . PRL 75 (1995) 513.

He studied the dependence of degree of circularly polarized light emitted from Al0.06Ga0.94As sample as a result of tunneling of electron from Ni tip versus the tunnel barrier height. There was a competition in tunneling from 3d (polarized) and 4sp (unpolarized) states. For narrower and lower barrier the Pm increase and reach 50 %. It means that coupling between 3D-like states and the semiconductor increases faster than for the 4sp-like states. Important point: maybe the UHV conditions permit to keep low density of surface states in semiconductor with large value of constant of spin relaxation.

• S.F. Alvarado. In "New trends in magnetism magnetic materials and theiir application " eds. J.L.Moro n and J.M.Sanches, Plenum, NY, London 1994, p.175.

The system Fe tip/GaAS surface is studied. The 43% of circularly polarized recombinative radiation from tip/surface tunnel junction was obtained. The 95% electron magnetization in Fe was proposed. Limitations In principle there is no limitation.

Comparison of direct and reverse schemes

A very small amount of polarized tunnel current is observed in direct scheme A for GaAs tip/magnetic sample combination : 0.2-0.4% (Prins et al., Suzuki et al.). The original degree of spin polarization of tip electrons was estimated as 10% in both cases. For the Co/Pt sample (Prins et al) the magnetization of sample was not equal to 100%.

In direct scheme B a larger degree of tunnel current polarization was observed, 6-10 % (Sueoki et al). Taking into account that original degree of polarization of semiconductor sample carriers is about 50%, we can compare the results of schemes A and B. A : 0.4%x5=2% is less than in B (6%) but of the same order of magnitude.

Opposite in reverse scheme B (Alvarado et al) the degree of circular polarization of recombination light can reach 50%. It means that spin relaxation time in semiconductor surface was rather large so the spin-oriented electrons injected from magnetic tip were able to recombine with the holes before the loss of spin orientation. For 95% magnetization in Fe they obtained 43% in STM measurement with Fe tip.

They obtain the optimal conditions upon STM measurements for very narrow tunnel barrier. For Ut=1.8 V and change of It from 20pA to 500 pA the polarization in Ni tip/AlGaAs system was changed from 25% to 50 %. The change of tunnel barrier thickness was estimated as 1.8 A.

Only papers of Alvarado et al and Suzuki et al were made in UHV conditions. Alvarado et al pay strong attention for study the processes of radiative recombination in GaAs samples (they made special publications in this field) and prepared the MBE grown sample with optimized parameters. Also clean atomically flat sample was produced by cleaving in situ in their works.

In Suzuki paper the authors used the oxide GaAs tips prepared by cleaving in air for UHV STM. No information about semiconductor surface states, spin relaxation time, etc. was presented in the paper. We think that spin relaxation time in oxidized GaAs tip was short enough and degree of spin polarization of semiconductor electrons was very low. The similar problems probably were in the papers of Prins et al and Sueoka et al.

The main conclusion from data presented is that this field of STM activity is very difficult in experimental realization. The experimental papers presented here are only the first steps. Not all possible combinations of tip/sample/light system are yet realized. The key problem lies in the spin relaxation processes in semiconductor surface.

• The study must be carried out in UHV only.
• The semiconductor tips prepared in UHV have to studied specially from the point of view of spin relaxation process.
• The reverse scheme A has to be checked too. Perhaps it is a compromised and convenient measurement scheme for SP-STM.

Special note about the teams involved in this activity in present time.

van Kempen team from Holland is broken because M. Prins is working in another field (Philips Lab) and R. Jansen continues this activity in MTI, USA.

Two teams from Japan have been involved in this field recently. It is Suzuki et al. (APL 1997) and Musaka et al.

So, only three teams (USA and Japan, no from Europe) continue activity in this field.

The basic idea here is to study STS dependence for selected nanosize objects (up to atoms) in a magnetic surface. With the data from other methods measurements and from theory to try to understand the electronic structure (including its spin orientation) of magnetic surfaces. The first papers in this field have been published recently.

• A.Davies, J.A. Stroscio, D.T.Pierce, and R.J. Celotta. "Atomic scale observations of alloying at the Cr-Fe(001) interface". PRL 76 (1996) 4175.

Using the STS information obtained for Fe(001) and Cr(001) samples separately (PRL 1995) the authors could distinguish the Cr and Fe atoms in a alloy. There were no atomic resolution in STM images but bright places in STM images were identified by STS as Cr atoms.

• A. Biedermann, O.Genser, W.Hebenstreit, M.Schmid, J.Redinger, R.Podloucky, and P.Varga. "Scanning tunneling spectroscopy of one-dimensional surface states on a metal surface" PRL 76 (1996) 4179.

In Fe-Si alloy they could identified the boundaries between Fe-Si domains enriched by Fe atoms by means of analysis of STS data and comparing it with STS on FE(001) previously obtained. Moreover they determined the one-dimensional character of this Fe state.

• M.Bode, R.Pascal, M.Dreyer and R.Wiesendanger. "Nanostructural and local electronic properties of Fe/W(110) correlated by scanning tunneling spectroscopy" PR B54 (1996) 8385-8388.

For Fe film of different thickness (1, 2, or 3 ML) the different STS curves have been observed. They explained the STS results by strain-induced change of electronic structure of heteroepitaxially grown film.

• M.Bode, Ch.Zarnits, R.Pascal, M.Dreyer and R.Wiesendanger. "Atomic and local electronic structure of Gd thin films studied by STM and STS" PR B56 (1997) 3636-3639.

By STM a set of different superstructures was observed in Gd film. The own STS curves were observed for each superstructure. They observed peaks in STS both for empty and occupied states.

Conclusion

This direction seems to be interesting for the materials with magnetism like rear earth elements and transition metals. In combination with XPS, UPS and theoretical predictions it could be very useful. The most interesting application for the systems with low dimensions: rear earth/transition metals quantum dots and wires.

• There is no large activity in SP-STM field as a whole.
• The first attempts at the beginning of nineties did not give the fast successful result for magnetic tip/magnetic surface system. There is no new published papers in this direction.
• In comparison with magnetic tip measurements the semiconductor tip/magnetic surface system has additional complexity like circularly polarized light. The slow moving in this direction is observed. Only recently in 1997 the first map of spin distribution on surface with very poor resolution was obtained for Co ultrathin film. Nevertheless the use of semiconductor tips seems rather promising direction of search. But the experiments must be done only in UHV with carefully prepared and tested tips.
• Alternative approach to get the information of electronic structure of magnetic surface and some suggestion about electron spin state is Scanning Tunneling Spectroscopy. No many publications in this field too. Prof. R.Wiesendanger who was a "father" of SP-STM with magnetic tip prefers this STS way now.

One of our STM (Room temperature STM) has a very open scanner construction which permits to put the light excitation in tunnel junction in parallel with tip axis. It can be used for experiments with polarized light excitation of tunnel junction semiconductor tip/magnetic surface.

Another STM (Variable temperature STM, 120-800 K) is combined with optical spectrometer (400-800 nm). It permits to realize the experiments with measure the polarized recombinative radiation from tunnel junction semiconductor tip/magnetic surface.

Taking into account that the spin relaxation in III-V semiconductor tips seems to be central problem in SP-STM the experience of Moscow team in chemical surface reaction is useful for tip preparation in situ. Optical education of its members (as a rule) and experience in laser, nonlinier optics, light polarization are also important.

Semiconductor (GaAs, InAs, etc) tips testing and characterization.

• Atomic resolution images on HOPG and Cu(100)
• TEM measurements
• Recombination radiation measurements. To use a matrix of tips on corresponding substrate. UHV preparation and measurements in optical spectrometer/LAS600 setup.
• Spin relaxation time measurements on tip matrix in optical spectrometer/LAS600 setup. Study of preparation conditions (ion etching, heating, gas adsorption) influence on spin relaxation time.

Spin polarized measurements with semiconductor tip

Reverse scheme A.

• A study of spin polarized properties of tip. Observation of circularly polarized light from tunnel junction. Point mode.
• A study of spin polarized properties of tip. Observation of circularly polarized light from tunnel junction. Frame mode.
• Surface temperature spin dependence measurements

All measurements are carry out in VT STM/optical spectrometer setup.The experiment conditions:

• Surface - 100 % perpendicular oriented spins
• Collection angle - 32°
• Temperature of surface - 120-800 K

Direct scheme A.

• Observation of modulated tunnel in phase with circularly polarized light excitation. Point mode.
• Observation of modulated tunnel in phase with circularly polarized light excitation. Frame mode.

All measurements are carry out in RT STM setup. The experiment conditions:

• Surface - 100 % perpendicular oriented spins
• Excitation - from tip side, along tip, angle 10a°
• Temperature of surface - 300 K Laser - emitted laser diode, 1.43 eV, 1-10 mW

Magnetic tips fabrication and testing.

The requirements:

• Sharp
• Have to have a magnetic moment on an apex
• Magnetic moment orientation ?

The candidates:

• Si needle crystal /Ni or Fe deposition
• W/ Ni or Fe deposition
• Si needle crystal/Ni cap.

Testing:

• TEM measuremants
• Atomic resolution images on HOPG and Cu(100)
• Squid measurements on matrix of tips
• XPS measurements on matrix of tips

Magnetic tip measurements:

• STS on Fe(100) with tips of different magnetization
• Domain walls. The size of domain must be not larger than 0.5
• Magnetic quantum dots

Nonmagnetic tip STS measurements:

• Repeating of Fe(100) STS results
• Optimization of STS scheme measurements (lock-in , etc)
• Determination of object of research on the base of XPS,UPS, etc data.

Dr. Konstantin N. Eltsov