Atomic-scale control of substrate-spin coupling via vertical manipulation of a 2D metal-organic framework

TL;DR

This work demonstrates atomic-scale control of substrate–spin coupling in a 2D kagome MOF (DCA3Cu2) on Ag(111) by vertically manipulating MOF adsorption height with an STM tip. The authors link Kondo temperature $T_ ext{K}$ to MOF–substrate hybridization $|V_ ext{hyb}|$ using an Anderson impurity model, identifying a ~4% difference in $|V_ ext{hyb}|$ between CuA and CuB sites that accounts for distinct $T_ ext{K}$ values. They achieve reversible, site- and region-wide switching of Kondo coupling through controlled height changes, illustrating mechanical control of exchange coupling without magnetic fields. These findings have implications for atomic-scale design of spintronic functionalities in 2D magnetic materials and open avenues for STM-based transport measurements in MOFs.

Abstract

Two-dimensional (2D) materials with frustrated crystal geometries can host strongly correlated electrons, potentially leading to a range of exotic many-body quantum phases such as Mott insulators, quantum spin-liquids, and Kondo lattices. The ability to control exchange-coupling within these systems is therefore highly desirable. Here, we use an atomically sharp scanning tunneling microscope probe to vertically manipulate a 2D Mott insulating kagome metal-organic framework (MOF) featuring Kondo-screened local magnetic moments on Ag(111). We show that by controlling the adsorption height of the MOF, we can also controllably and reversibly change the strength of Kondo coupling between the MOF's local spins and the substrate's conduction electrons. This mechanical control of Kondo coupling could be extended to other forms of interlayer exchange coupling, potentially allowing for atomic-scale design or control of spintronics technologies.

Figures (5)

• Figure 1: Adsorption heights and Kondo temperatures of Cu sites.a STM image of DCA3Cu2 MOF on Ag(111) ($V_\mathrm{b} = -20$ mV, $I_\mathrm{t} = 25$ pA). DCA molecule balls-and-stick model superimposed (grey: carbon; white: hydrogen; blue: nitrogen). Dashed red (blue) circles indicate CuA (CuB) sites, white diamond indicates MOF unit cell. b NcAFM image of DCA3Cu2 on Ag(111) (same region as a; tip functionalized with CO molecule). c NcAFM frequency shift, $\Delta f$, as a function of tip-sample distance $\Delta z$, acquired with a CO-tip at CuA (red) and CuB (blue) sites: difference in $\Delta z_{\mathrm{min}}$ -- corresponding to $\Delta f$ minimum in $\Delta f(\Delta z)$ curves -- illustrates difference of $\sim$0.2 Å in adsorption height. Solid white lines: cubic fits used to determine $\Delta z_{\mathrm{min}}$. d,e Temperature-dependent d$I$/d$V$ spectra at CuA and CuB sites, respectively, illustrating the evolution of the zero-bias peak with temperature. Examples of fitted curves shown in 4.4 K data as solid black lines. A Fano function was used to capture the zero-bias peak. f Evolution of CuA (red) and CuB (blue) zero-bias peak half-width at half-maximum $\Gamma$ as a function of temperature, from which $T_\mathrm{K}$ was determined for each site. Error bars and uncertainties represent one standard deviation in fitting parameters.
• Figure 2: Effect of MOF-substrate hybridization on zero-bias Kondo resonance and Kondo temperature.a Schematic illustration of stronger MOF-substrate hybridization ($|V_\mathrm{hyb}|$) at Cu$_\mathrm{B}$ sites than at Cu$_\mathrm{A}$ sites due to MOF adsorption height difference (DCA adsorption angle within MOF exaggerated for emphasis). b$T_\mathrm{K}$ as a function of $|V_\mathrm{hyb}|$ according to Eq. (\ref{['Kondo_hybridization']}). Red (blue) marker indicates point on curve corresponding to experimental value of $T_\mathrm{K}$ for Cu$_\mathrm{A}$ site (Cu$_\mathrm{B}$ site). c Theoretical Fano resonance (see Methods) for different values of $|V_\mathrm{hyb}|$, using Eqs. (\ref{['Kondo_temp']}) and (\ref{['Kondo_hybridization']}) for $\Gamma$ and $T_{\mathrm{K}}$, at $T=4.4$ K (to match experiments). Highlighted red (blue) curve associated with value of $|V_\mathrm{hyb}|$ which most closely reproduces experimental spectra at Cu$_\mathrm{A}$ (Cu$_\mathrm{B}$) sites.
• Figure 3: Switching of Cu site adsorption height.a,c STM images of DCA3Cu2 on Ag(111), before and after vertical manipulation, respectively ($V_\textsubscript{b} = -20$ mV, $I_\textsubscript{t} = 50$ pA). DCA balls-and-sticks model superimposed. b Conductance, $G$, as a function of tip-sample distance difference, $\Delta z$, during vertical manipulation at site marked by cross in a,c ($V_\textsubscript{b} = -20$ mV). Blue (orange) curve corresponds to tip approaching (retracting from) sample. Black circle indicates initial tip position at tip-sample distance defined by setpoint $V_\textsubscript{b} = -20$ mV, $I_\mathrm{t} = 100$ pA. Red arrow shows increased conductance upon returning to $\Delta z = 0$. Black single-headed arrows indicate rapid changes in conductance. Black double-headed arrow indicates difference between initial and final conductance, showing switching of Cu site adsorption height. d,e Schematic illustration of DCA3Cu2 on Ag(111) before and after manipulation with STM tip, respectively (DCA adsorption angle exaggerated for illustrative purposes). Cross and white arrow indicate corresponding positions in STM images in a,c.
• Figure 4: Switching of Cu site Kondo coupling strengths.a,b STM images of DCA3Cu2/Ag(111) before and after vertical manipulation, respectively, at initial Cu$_\mathrm{A}$ site marked by cross ($V_\textsubscript{b} = -20$ mV, $I_\textsubscript{t} = 50$ pA). Manipulation parameters as in Fig. \ref{['3']}. c d$I/$d$V$ spectra acquired at positions indicated in a, b, before (solid curves) and after (dashed curves) manipulation. Setpoints: $V_\mathrm{b} = -100$ mV, $I_\mathrm{t} = 100$ pA. Black curves offset for clarity. d Same spectra as in c, overlaid for comparison (the two pairs of curves are offset for clarity). Local structural and d$I$/d$V$ signatures corroborate switching from Cu$_\mathrm{A}$ to Cu$_\mathrm{B}$ site at manipulation location (black cross and spectra), and from Cu$_\mathrm{B}$ to Cu$_\mathrm{A}$ at adjacent location (orange circle and spectra).
• Figure 5: Extended and reversible structural and Kondo coupling switching.a,c,e STM images of DCA3Cu2/Ag(111), before and after vertical manipulation ($V_\textsubscript{b} = -20$ mV, $I_\textsubscript{t} = 50$ pA). Blue ovals indicate pairs of Cu sites which switch upon manipulation. Dashed red (blue) circles: Cu$_\mathrm{A}$ (Cu$_\mathrm{B}$ site) sites. b,d Conductance $G$ as a function of tip-sample distance change $\Delta z$, during vertical manipulation using $V_\textsubscript{b} = -20$ mV, at site of black cross in a, c, e. Blue (orange) curves indicate tip approaching (retracting). Black circle indicates initial tip position defined by setpoint $V_\textsubscript{b} = -20$ mV, $I_\mathrm{t} = 100$ pA. Red arrow shows increased conductance upon returning to $\Delta z = 0$. Black single-headed arrows indicate rapid changes in conductance. Black double-headed arrow indicates difference between initial and final conductance.