Fig. 2

Dynamics and energy barrier of the observed hopping mechanism. a–d AFM images of type I DATP molecule taken at 5.2 K, 7.0 K, 9.3 K, and 14.0 K. The tip-heights Δz are a −90 pm, b −85 pm, c −100 pm, d −90 pm, relative to a STM set point of 100 mV, 10 pA on bare Cu surfaces. e–h Current vs time traces recorded at the fuzzy end at 5.2 K, 7.0 K, 9.3 K, and 14.0 K. Parameters: V_bias = 60 mV, tip height Δz = −10 pm with respect to the STM tunneling set point of 100 mV, 10 pA on bare Cu(111) surface. i Arrhenius type plot of natural logarithm of the jumping rate ln(k) vs 1/T. Displayed are the jumping rates for all jumping events (total jumps, red circles) and the rates for jumps into the upper (blue circles) and into the lower current states (black circles), respectively. An energy barrier (E_α) of 5.15 ± 0.13 meV and a pre-exponential factor (A) of e^9.40±0.22 s⁻¹ for the total jumps is determined by fitting the Arrhenius’ equation (\(\ln (k) = \left( { - E_{\mathrm{\alpha }}/k_{\mathrm{B}}} \right)\left( {1/T} \right) + \ln (A)\), where k_B is the Boltzmann constant) to the right three points (at 5.2 K, 7.0 K, and 9.3 K). Jumps into lower state: E_α = 5.00 ± 0.09 meV, A = e^(8.69±0.16) s⁻¹. Jumps into upper state: E_α = 4.98 ± 0.04 meV, A = e^(9.46±0.08) s⁻¹. Please note, the data point at 14.0 K has not been taken into account since the observed jumping rate for this temperature exceeds the bandwidth of our tunneling amplifier setup. The vertical error bars in i are derived from the standard deviation of k (see Supplementary Fig. 3). Scale bar: 0.5 nm