Quantum-coherent Control of Individual Spins on Surfaces

Figure.1 ESR.STM
The combination of ESR with STM merges some of the most powerful advantages of both techniques: The atomic-scale spatial resolution of STM with the much higher energy resolution of ESR. This integration of two techniques allows us to achieve the single spin sensitivity at atomic precision. Ensemble ESR uses oscillating magnetic fields at RF frequencies (10 - 100 GHz) to drive and sense the spin resonance. This is usually achieved by using a special cavity to enhance those fields. However, incorporating such an approach into an STM is very difficult and has not been achieved to date. On the other hand, an STM tunnel junction always contains strong electric fields. This led to the key idea, conceived in 2008 by QNS’s director and his colleague, Arzhang Ardavan from Oxford University, to realize that RF frequency oscillating electric fields in the tunnel junction could be used to drive ESR in an STM, illustrated in Figure 1.

*Fig.1 ESR STM. A quantum spin (yellow) is put on a thin insulating film (purple) that is supported on a metal substrate (grey). The STM tip contains a magnetic apex (grey arrow) which results in a spin-polarized tunnel current. We add a GHz frequency RF voltage to the normal DC tunnel voltage that is applied between tip and sample.
Figure.2 ESR spectra of Ti spin on MgO
There are several key ingredients highlighted in Figure 2. First, the spin needs to be decoupled from a metallic substrate via a thin insulating film in order to maintain its quantum-ness. Second, the tip needs to be spin-polarized in order to measure the states of the spin on the surface. Third, an external magnetic field is used to adjust the Zeeman energy of the quantum spin on the surface to be in resonance with the RF electric field. And fourth, a high-frequency oscillating electric field is applied between tip and sample. With this set of ingredients, we are able to electrically drive the spin resonance of atoms and molecules on a surface. A typical ESR-STM spectrum is obtained by measuring the change of the tunnel current while sweeping the frequency of the RF voltage at a fixed external magnetic field, see Figure 2. We see the ESR signals as current changes of several hundreds of femto-ampere (10-12 A), when the RF voltage is resonant with the Zeeman energy. One can also clearly see that the Zeeman energy, and hence the frequency of the ESR resonance, changes in proportion to the magnitude of magnetic field. The exquisite energy resolution of ESR STM can be appreciated here by pointing out that magnetic fields of 0.65 T and 0.70 T result in well-separated ESR spectra. For the experts, the minimum linewidth of ESR-STM is on order of MHz which corresponds to about 10 neV.

*Fig.2 ESR spectra of Ti spin on MgO. Here the x-axis is the frequency of the RF voltage and the y-axis is the tunnel current. According to the basic ESR equation, a higher field requires a higher frequency. Measured with our “Bob” system at 0.9 K.

Longer-term Goals

• Measure and coherently control the quantum states of atoms on surfaces
• Use precise atom manipulation to build engineered nanostructures
• Use atoms as qubits for quantum information protocols
• Demonstrate quantum sensing with improved sensitivity
• Extend ESR-STM beyond 3d transition metal atoms on MgO

Near-term Goals

• Implement pulsed spin resonance
• Broaden the investigation in double-resonance spectroscopy
• Demonstrate switchable entanglement between two qubits
• Demonstrate pulsed electron-nuclear spin double resonance

Research results to date

Research results to date 1
Figure.3 Multi-spin ESR STM

QNS has been a leader in ESR STM from the very beginning. Some of the QNS highlights include: In 2021, QNS researchers used their “Alice” and “Bob” tools to extend ESR STM from the control of single spins to multiple spins. This breakthrough work gives a glimpse at the potential power of controlling multiple qubits on a surface in an atomically engineered nanostructure. Here, we used one Ti spin in the tunnel junction as an ESR sensor to measure the spin states of a second Ti spin, called the remote spin, which was weakly coupled to the first. The schematic is shown in Figure 3. The remote spin (Ti-2) is driven by the combination of the resonant electric field from the tip and its interaction with the neighboring iron spin, which acts like a permanent single-atom magnet. This breakthrough enables the coherent control of multiple spins coupled to each other using one STM tip. We believe that computational protocols that rely on the coherent control of multiple spins on a surface are now within reach. Submitted for publication.

*Fig.3 Multi-spin ESR STM. Titanium spin 1 is driven by the magnetic STM tip while Ti spin 2 is driven by its neighboring iron atom. Two RF voltages are used, one for each Ti spin resulting in a novel double-resonance spectroscopy.
Research results to date 2
Figure.4 “Eve” ESR STM scan head.

In 2021, QNS finished the construction of a totally home-built ESR STM system, nicknamed “Eve”. Eve relies on the continuous circulation of Helium gas to cool to low temperatures. With normal Helium (He-4 isotope) this system achieves 1.6 K in continuous circulation and 1.0 K in single-shot mode with a hold time of 10 hours. Eve is equipped with two RF lines to bring high-frequency signals into the tunnel junction. One is connected directly to the tip (our standard method) and a second one acts as an antenna, see Figure 4. We found that our antenna works slightly better and can transmit RF powers at higher frequencies. Eve incorporates a fairly open design, which will enable us to incorporate further improvements in the coming years. The entire process of the Eve construction took about two years with a dedicated team of students, researchers, and engineers. Eve measured its first ESR spectrum of Ti on MgO in August. Welcome Eve!

*Fig.4 Multi-spin ESR STM. You can see tip and sample (on sample holder) as well as the RF antenna.
Research results to date 3
Figure.5 Spin-flips of Dysprosium on MgO.

In 2020, QNS measured the anisotropic magnetic properties of titanium atoms on the bridge binding site of MgO. This work was measured with our “Bob” ESR-STM. Bob is a home-built STM that is attached to a commercial dilution refrigerator made by Janis Company. The entire room-temperature sample preparation chamber and in-situ sample transfer are also home-built. Bob’s lowest temperature is 10 mK. Bob was finished in March 2020 and this work is the maiden voyage of Bob. Interesting facts: Bob has been cold since its initial cool down with STM in March 2020 and has 3 STM samples in it that were dropped accidentally …. Yet, Bob is happy to keep taking data! Titanium on this binding site has very low symmetry and consequently we measured three different strengths of the Zeeman energy in the x,y, and z directions. This work also investigated how the ESR amplitude changes with the angle of the applied magnetic field, shining light on the ESR detection and driving mechanisms. Work is almost accepted for publication in Phys. Rev. B (2021).

In 2020, QNS joined forces with researchers from DICP in San Sebastian, Spain and used our “Bob” STM to measure the properties of Cr spins on the superconductor bismuth-palladium (Bi2Pd). We used the low-temperature of Bob and its atom manipulation capability to show that chains of Cr atoms along certain directions might host Majorana fermions. Work published in Phys. Rev. B (2021).

In 2020, we studied the properties of Dysprosium (Dy) atoms on thin MgO films using our “Alice” system. We found that Dy is an excellent single-atom magnet, even better than the holmium atoms we studied previously. Dy is a rare earth atom with the spin contained in an inner electronic shell. This keeps the spin more protected from the environment

In 2019, QNS researchers collaborated with the director’s former group at IBM to perform two key experiments. First, we extended ESR STM from the continuous wave to the pulsed mode. This is an important step towards quantum-coherent control of spins on surfaces. The work was published in Science (2019). We also measured the interaction between the spin-polarized STM tip and Fe and Ti spins on MgO with high spatial resolution in ESR-STM. This work resulted in a nanoscale magnetic resonance imaging and was published in Nature Physics (2019).

In 2018, we used our “Alice” system to measure ESR STM of iron (Fe) on MgO in a vector magnetic field and showed that most Fe tips require an in-plane component of the magnetic field to show ESR STM contrast. We also showed ESR STM can be observed even in the absence of an external magnetic field. This work was published in Nano Lett. (2019). In 2018, QNS received its first ESR STM, a commercial system from the Unisoku Company, nicknamed “Alice”. Alice was installed in a temporary lab space for less than 1 year before moving to our new research building in the summer of 2019. Alice has been exceedingly successful for QNS and allowed us to be one of the first teams (at the same time with colleagues in Switzerland) to demonstrate ESR-STM outside of the director’s former IBM lab.

*Fig.5 Spin-flips of Dysprosium on MgO. Spin-polarized tunnel current as a function of time when excited at 150mV. Measured with “Alice”.
Figure.6 Hyperfine Interaction.

In 2018, QNS collaborated with IBM and used the IBM machine to measure several key results in ESR-STM. First, we measured the interaction between the nuclear spin and the electron spin of Fe and Ti atoms, see Figure 6. This hyperfine interaction was sensitive to the local environment. Results were published in Science (2018). We also found a rather strong hyperfine interaction in copper spins with S=1/2 and showed the hyperpolarization of the nuclear spin by exchanging angular momentum with the spin-polarized tunnel current. Results were published in Nature Nanotechnology (2018). We found a way to create a decoherence-free subspace by utilizing a clock-transition in an engineered Ti dimer. This led to longer quantum coherence times. Results were published in Science Advances (2018). Finally, we studied the primary sources of quantum decoherence for ESR-STM of Fe on MgO and found the tunneling current to be the primary source. Results were published in Science Advances (2018).

In 2017, all experiments had to be performed at outside facilities since QNS had no research space yet. We collaborated with IBM and achieved two key results. First, we found that Ho on MgO is a single-atom magnet that is stable for many days at low temperature but can be switched at elevated tunnel voltages. The results were published in Nature (2017). In a second work, we focused on the ability of STM to move atoms and built tailored spin structures. We used that ability in ESR-STM to measure the weak dipolar coupling between Fe spins. This allowed us to determine the magnetic moment of nearby atoms. We further suggested a use as a nano-GPS. The work was published in Nature Nanotechnology (2017).

*Fig.6 Hyperfine Interaction. ESR spectra of three Ti atoms that look identical in the STM images on the right. Blue spectrum has a single ESR line due to the electron spin. Orange and red spectra have multiple lines due to the nuclear spin of Ti isotopes. Measured with IBM machine.

Research Equipment

Research Equipment
Figure.7 Typical sample for ESR STM.

QNS is in the lucky position to operate three low-temperature ESR-STM systems that combine a room temperature sample preparation chamber with in-situ sample transfer into the STM. All three systems are equipped with high-frequency cabling to perform ESR-STM studies and are equipped with two-axis vector magnets. This is quite important since for many studies it is crucial to control an out-of-plane magnetic field independently from an in-plane magnetic field. The three room-temperature preparation chambers operate in ultra-high vacuum with a base pressure of 1.0x10-9mbar or better and are optimized for the growth of thin oxide films. They contain a residual gas analyzer to make sure that the vacuum chamber remains clean.

*Fig.7 Typical sample for ESR STM. Grey: clean surface of silver metal substrate. Red and blue: oxygen and magnesium atoms of MgO film (here one layer thick). Yellow: magnetic atom such as titanium or iron.

Figure.8 Gold surface.

Samples that are mounted on sample holders can be brought into the vacuum chambers through a load-lock. Such a sample will have a rough surface that is covered by a thin layer of water and dirt. As a first step, we use sputter-cleaning, which acts like a sand-blasting tool, to remove the dirt as well as some of the metal. We prefer to use ionized argon as our sand, which can be accelerated with a high voltage before it hits the surface of our sample. After such a sputter process, our samples are cleaner but also rough, containing many impact craters. As a second step, we need to heat the sample to typically 500 degrees Celsius. At such temperatures, the surface atoms are mobile and rearrange to make a flat sample surface. Many times, this process has to be repeated over and over to get a clean surface, see Figure 8 for a typical example. In order to preserve the quantum states of our qubit atoms it is crucial to decouple them from the metal. A metal has too many free electrons that would destroy the quantum states. We achieve this by growing a thin film of an insulator on top of the clean metal, in most of our experiments that is magnesium oxide (MgO) on silver, see Figure 7 for a schematic and Figure 8 for a topographic image. We use metal beam epitaxy in an oxygen environment for this growth, which is surprisingly difficult and often makes us scratch our heads. Once such a sample is cooled down close to room-temperature in the vacuum chamber, we pick it up with a grabber in vacuum and transfer it into the cold STM.

*Fig.8 Gold surface. STM image of the famous herringbone reconstruction of Au(111) surface. Different stacking gives rise to a stripe contrast. A single atomic step is also visible at the bottom. Image measured at 1.6K with our “Eve” system.
Alice ESR STM system
Commercial system
Received: December 2018 First Results: February 2019
Key results

• First ever ESR-STM on a molecule
• Pulsed ESR-STM
• ESR on two spins (double resonance)
Temp: 0.4 K in single shot mode
2K in continuous mode
Magnet: 9 T perpendicular to sample
4 T in plane
ESR: 15-18 GHZ best range
Figure.9 “Alice” STM system.

Our first ESR-STM system is nicknamed “Alice” and is a commercial system, which we purchased from Unisoku Company. Alice is a dewar-type STM where the STM is about 1.5 m deep inside a dewar that is filled with liquid helium and contains two superconducting magnets. The liquid helium cools Alice’s STM down to about 5 K. When we want to go to lower temperature, we fill a small pot, called a 1 K pot, with liquid helium and pump on the evaporating helium gas. This lowers the temperature of the pot and the STM and we can continuously operate at about 2 K. When we want to go to even lower temperature, we liquefy a small amount of helium-3 gas (a super expensive isotope of normal helium). After we have liquefied all our helium-3 gas, we pump on it, which lowers our temperature to 0.4 K with a hold-time of 1.5 days. When all the liquid helium-3 is evaporated and recovered, we repeat this cycle. Alice was installed in December 2018 in our temporary lab space and moved to our new building in the summer of 2019. Alice is a real work horse that has been permanently cold and operational – of course except for the move. The installation and the moving process both took less than one month. We are very happy with Alice and we have already achieved many great results as outlined above.

*Fig.9 Our “Alice” STM system in our old temporary lab space. The room-temperature sample preparation chamber is wrapped in foil. The STM sits inside the large green dewar near floor level.
Bob ESR STM system
Partially home-built system
Start of assembly: August 2019 First Results: March 2020
Key results

• ESR as a function of magnetic field angle
• Spectroscopy of spins on superconductor
• ESR on two spins (double resonance)
Temp: 0.01 K in continuous mode
Higher temperatures are possible
Magnet: 6 T perpendicular to sample
9 T in plane
3 T vector field at any angle
ESR: 12-23 GHZ best range
Figure.10 STM and dilution fridge of Bob.

Our second ESR STM system is nicknamed “Bob” and its claim to fame is its ultra-low temperature. For Bob, we purchased a commercial dilution refrigerator from Janis Company. In contrast to Alice, Bob always runs in continuous circulation mode. However, the insides of such a dilution fridge are significantly more complex than Alice’s cooling system. Bob has reached the lowest temperature at the base plate of the STM of 7 mK. Figure 10 shows a photograph of Bob taken from below ground level (with the camera pointed up towards the ceiling). At the very bottom is the home-built STM, which is very similar to the one shown in Figure 4. Above the STM is a copper section which is bolted to the lowest and coldest part of the dilution fridge. It transfers the cold temperature to the STM. When inserted into the dewar and cooled, the baffle section (with the white Styrofoam cylinders) marks the neck of the dewar where the temperature transitions from 6 K at the bottom to room temperature. We started designing and making parts for Bob around January 2019 and then started construction in August 2019 as one of the first acts after moving into our new building. We finished the construction in March 2020 when we cooled down the STM part for the first time. Bob has been cold ever since that first cool down! When we first tested the sample exchange in Bob, we lost 3 samples in the low-temperature section where they are still lying around without causing any real troubles, amazing! Bob has been used for several research topics as described above.

*Fig.10 STM and dilution fridge of Bob. Photograph taken with the dewar and heat shields removed. Gold-colored section is the dilution fridge where the temperature changes from 6K to 10mK. Copper-colored lower section is our STM stage with tip and sample near the bottom.
Eve ESR-STM system
FULLY home-built system
Start of assembly: September 2019
First Results: August 2021
Key results

• Brand new system
Temp: 1.6 K in continuous mode with He-4
1.0 K in single-shot mode with He-4
Lower temperature with He-3 planned
Magnet: 4 T perpendicular to sample
6 T in plane
4T vector field at any angle
ESR: 15-30 GHZ best range
ESR drive on tip OR on antenna
Figure.11 Photograph of Eve.

Our third ESR-STM system is nicknamed “Eve” and was completed in August 2021. Eve is completely home-designed and home-built and is based on a continuous circulation of helium gas for cooling below 5 K. We are currently operating Eve with normal helium gas for the circulation and we reach 1.6 K in continuous circulation and 1.0 K in single-shot mode. Switching between these modes simply means opening and closing one valve and can be done while tunneling. We expect to reach 1.0 K in continuous circulation when we change to the expensive helium-3 isotope for circulation. The most important aspect of Eve is the fact that is entirely home-built, which allows us to easily modify every aspect of the system for future improvements. One such improvement that we have in mind is to incorporate a low-temperature current amplifier to be able to measure at lower tunnel currents and achieve a better signal-noise ratio. In the current day and age it is rather rare that people build an STM system from scratch, especially an ESR-STM with all its special requirements. However, through a concerted effort of our students, staff members and engineering team, we achieved this task in a little under two years. We are looking forward to our first scientific explorations with Eve.

*Fig.11 Photograph of Eve. Middle part shows the room-temperature preparation chamber where clean samples are produced. Top part: lower section of the sample transfer arm, which shuttles samples between STM and room-temperature chamber. Lower section: upper part of the low-temperature section. The STM is about 1.5 meters below this image.