Imaging quantum chemistry in action with quantum attomicroscopy Authors: Nikolay V. Golubev, Mohammed Th. Hassan Published: 2026 ABSTRACT Understanding how quantum electron and nuclei motions drive biomolecular reactions is a foundational challenge in quantum chemistry and biology. While ultrafast charge migration in DNA is theorized to influence genome stability and signaling, real-time observation remains elusive. Here, we report a theoretical study employing high-level ab initio simulations to reveal attosecond charge dynamics and electronic coherences within canonical DNA base pairs. Experimentally, we propose the “quantum attomicroscope” (Q-attomicroscope)—a conceptual instrument providing attosecond temporal and sub-angstrom spatial resolution. By bridging theory and instrumentation, this work outlines a pathway for laser-mediated DNA manipulation with transformative implications for chemical reactivity and personalized medicine. Introduction For much of the twentieth century, chemical processes were assumed to occur on timescales far too short to be directly observed1. This perception changed dramatically with the advent of femtosecond time-resolved spectroscopy. By effectively “freezing time,” this technique enabled experimental access to the structural dynamics of chemical reactions, providing unprecedented insight into transition states, coherent control, and reaction outcomes. These developments gave rise to the field of ultrafast science and femtochemistry, in which atomic and molecular motion can be monitored in real time2. At their most fundamental level, chemical reactions are governed by the breaking and formation of chemical bonds. These processes are driven by electron motion, which naturally unfolds on femtosecond to attosecond timescales. As a result, the direct observation—and ultimately the control—of the electron dynamics in molecules and chemical reactions as they occur has long remained beyond experimental reach. Understanding the electron dynamics in biomolecules is of fundamental importance because electronic motion constitutes the earliest and most decisive step in virtually all chemical and biological processes3–5. In systems such as deoxyribonucleic acid (DNA), proteins, and enzyme complexes, the redistribution of electronic charge unfolds on femtosecond to attosecond timescales—well before significant changes in molecular structure can occur—and establishes the initial conditions that govern bond breaking, chemical reactivity, and biological function3,6. In DNA in particular, charge transfer has been implicated in photoinduced damage, oxidative stress response, mutagenesis pathways, and long-range biomolecular signaling, where the correlated non-adiabatic dynamics of electrons and nuclei can mediate charge transport across multiple base pairs prior to structural relaxation4,7–14. Recent theoretical and experimental studies have demonstrated that electron correlation alone can drive ultrafast charge redistribution, even in the absence of nuclear motion, highlighting the intrinsically quantum-mechanical origin of the process termed charge migration3,10,15,16. However, most existing experimental approaches rely on ensemble-averaged or time-integrated observables, which obscure the primary electronic mechanisms that initiate downstream chemistry and biological response12. Direct access to real-time electron dynamics in biomolecules would therefore provide a missing link between quantum chemistry and molecular biology, enabling a deeper understanding of how biological systems exploit electronic motion to achieve efficiency, selectivity, and functional robustness. Beyond fundamental insight, elucidating charge-migration pathways in biomolecules could inform strategies for mitigating photochemical damage, designing bio-inspired molecular electronics, and developing quantum-enabled approaches to biotechnology and medicine17,18. In this perspective, we propose such a tool, which we term the quantum attomicroscope (Q-attomicroscope). The concept is based on extending scanning tunneling microscopy (STM) by inducing the tunneling current with a half-cycle laser pulse confined to the sub-femtosecond regime. This approach aims to achieve simultaneous attosecond temporal and angstrom-scale spatial resolution, thereby enabling direct imaging of electron motion during chemical reactions. The underlying principles and proposed implementation of the Q-attomicroscope are discussed in detail in Section 3. The ability to image electron dynamics in real space and time would open transformative opportunities across chemistry, biology, and materials science. In particular, visualizing charge migration in molecules and biomolecules could provide critical insight into fundamental processes such as charge transfer and electron localization. For example, imaging charge migration in amino acids—the building blocks of proteins—would inform our understanding of electron dynamics in more complex biological systems, including DNA and proteins. Such capabilities could address longstanding questions in molecular biology, such as whether electrons tunnel between the two strands of DNA, along which trajectories, and on what timescales. Resolving these questions would deepen our understanding of molecular interactions, including DNA–protein interactions during transcription and translation processes, with direct implications for carcinogenesis, mutagenesis, and DNA repair mechanisms19. Beyond biology, these insights could also guide the development of DNA-based molecular electronics and bioinformatics applications17. More broadly, the principle that control follows observation suggests that Q-attomicroscopy could ultimately enable the manipulation of photochemical reactions and molecular conformations in real space. Such capabilities would represent a paradigm shift, with potential long-term implications for personalized medicine and targeted molecular control. As an initial step toward realizing this vision, we present a theoretical investigation of charge migration in the two fundamental DNA nucleobase pairs: thymine–adenine (T–A) and cytosine–guanine (C–G). We discuss the working principle and potential usage of Q-attomicroscope for observing the ultrafast electron dynamics in these systems. The presented combined theoretical and computational studies provide guidance on optimal experimental parameters and anticipated observables, laying the groundwork for future experiments using the Q-attomicroscope. Experimental approaches to study charge transfer and charge migration dynamics in biomolecular systems: advances and limitations Over the past several decades, a variety of experimental methodologies have been developed to investigate ultrafast dynamics in molecular and biological systems2,20. Time-resolved optical spectroscopies, including pump–probe transient absorption and fluorescence upconversion, have been extensively applied to DNA bases, base pairs, and oligonucleotides to infer charge-transfer dynamics following photoexcitation21–23. These approaches have established characteristic timescales for hole and electron transport along base stacks and have clarified the role of sequence composition, base pairing, and stacking geometry in mediating charge transfer. Photoelectron spectroscopy, particularly time-resolved photoelectron spectroscopy (TRPES), has enabled more direct access to fundamental mechanisms underlying the dynamics and revealed the importance of non-adiabatic couplings in the charge transfer process24–26. More recently, major advances in attosecond science have opened a new chapter in the study of ultrafast molecular dynamics revealing the mechanism of charge migration driven solely by the electron correlation and electron relaxation effects. The generation of attosecond extreme ultraviolet (XUV) pulses via high-harmonic generation has enabled real-time observation of purely electron dynamics in atoms, molecules, and solids12,27–30. In particular, attosecond time-resolved spectroscopy has been used to probe charge migration in molecules4,12,14,29, inaugurating the emerging field of attochemistry3,6,31–36. Despite these breakthroughs, most existing studies rely on high-energy XUV photons and primarily probe charge-transfer dynamics in dissociating molecular fragments rather than in intact molecular systems. Furthermore, while attosecond spectroscopy provides valuable temporal information on transient electronic states, it offers limited insight into the spatial pathways and mechanisms of electron motion. These limitations underscore the need for new experimental tools capable of directly imaging chemical reactions and electron dynamics in space and time. On another note, DNA bases have frequently served as benchmark systems in these studies due to their biological relevance and well-characterized molecular structure37. However, such experiments are typically restricted to gas-phase or weakly bound molecular systems and often involve ionization pathways that lead to fragmentation, limiting their applicability to intact biomolecules4,13,38. Complementary insights have been obtained from scanning probe techniques, including STM and conductive atomic force microscopy, which provide atomic-scale spatial resolution and sensitivity to local electronic structure39–41. These methods have successfully imaged static electronic states, charge localization, and molecular orbitals in individual biomolecules adsorbed on surfaces42,43. Nevertheless, conventional STM operates in a quasi-stationary regime and lacks the temporal resolution required to capture intrinsic electronic motion, which unfolds on femtosecond and attosecond timescales6,31,32,34. A central challenge across all existing approaches is the fundamental trade-off between temporal resolution, spatial resolution, and system integrity. Techniques offering attosecond temporal resolution typically sacrifice spatial information or require conditions incompatible with biological complexity, while methods providing atomic-scale imaging are generally limited to steady-state measurements. Bridging this gap remains one of the outstanding challenges in the field and motivates the development of new experimental infrastructures capable of directly imaging charge migration in biologically relevant systems in both space and time44–46. Theoretical study of the charge migration dynamics in DNA nucleobase pairs Theoretical treatment of ultrafast quantum dynamics in complex molecular systems such as DNA base pairs is highly challenging. A central difficulty in performing accurate numerical simulations lies in the need to account for the strongly correlated motion of a large number of electrons and nuclei forming the pairs. While a fully non-adiabatic description of the coupled electron-nuclear dynamics remains far beyond current computational capabilities for systems of this size, the electronic structure and the ultrafast electron motion can often be treated within well-established quantum-chemical frameworks by assuming fixed or parametrized nuclear configurations. In principle, some thermal vibrations and nuclear motion can occur on timescales comparable to the charge migration dynamics. However, these motions are small-amplitude vibrations of atoms around their equilibrium positions. Together with the dynamics of the associated electron density, these vibrations would appear in the recorded images as a slight blurring of atomic positions which is an unavoidable feature of any sub-atomic microscopy imaging. In a difference, the electron motion that propagates through the molecular backbone or takes place between distant molecular pairs such as those investigated in our study is a large amplitude motion that should be clearly visible in the proposed Q-attomicroscope. A separate aspect of the dynamics that must be mentioned is not the nuclear motion itself, but its influence on the electron dynamics in molecules. It has been shown in many studies47–49 that nuclear vibrations in small and medium sized molecules lead to decoherence of the electron dynamics which in many cases can happen within first few femtoseconds. However, it has been shown recently50 that the large size of a system can actually help to postpone the decoherence and prolong the charge migration. This enhanced charge migration dynamics in larger molecules can be attributed to the reduced sensitivity of the relevant electronic states to the removal of a single electron from a system containing hundreds of electrons. Accordingly, the DNA nucleobase pairs considered in our work are promising systems for sustaining long-lived charge migration. In fact, the unprecedented temporal and spatial resolution of modern experimental techniques, such as the Q-attomicroscope proposed in this perspective, can capture not only the pure electron motion in a system but also its coupling with nuclear rearrangement that occur at later stages of the dynamics. To model the first step of this process, in the follow-up discussion we present fully ab initio numerical simulations of the electronic structure and dynamics in T–A and C–G base pairs studying the motion of electron density triggered by the interaction of these systems with intense ultrashort laser pulses. Our calculations consisted of the following steps. First, preliminary molecular geometries of the nucleobases were constructed using the simplified molecular-input line entry system (SMILES) format using OpenBabel 3.0.051 with the MMFF94 force field52. The ground-state geometries of T, A, C, and G neutral molecules were optimized using the density functional theory at the B3LYP53/aug-cc-pVDZ54 level. The obtained individual molecules were arranged in the corresponding T–A and C–G pairs and the entire structures were re-optimized at the same level of theory. In the next step, we computed the ionic spectra. The noncorrelated reference orbitals were obtained using the Hartree–Fock (HF) method with aug-cc-pVDZ basis set. The cationic Hamiltonian was constructed using the non-Dyson version of the algebraic diagrammatic construction (ADC) scheme55,56 at the third order of perturbation theory for representing the one-particle Green’s function. The active space was chosen to contain all the orbitals except \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1s$$\end{document}1s core states of all heavy atoms. All one-hole (1 h) and two-hole-one-particle (2h1p) electronic configurations were taken into account in the calculations of ionic states and in the follow-up time-dependent density analysis. All simulations have been performed with the Q-Chem software package57. Ionizing an electron from T–A and C–G base pairs, we were searching for conditions under which the correlation between the electrons of neutral systems causes the so-called hole-mixing effect58,59. This effect arises when certain electrons of a molecule are particularly strongly correlated with each other such that localized ionization of a single electron causes the rearrangement of other electrons in the system. Figures 1a and 2a show the calculated ionization spectra of T–A and C–G base pairs, respectively. Vertical bars represent ionic states with different colors depicting the localization of the created hole in the molecular orbitals of the neutral system. It is seen that most of the low-lying ionic states for both systems are composed of a single configuration indicating that ionization of an electron from outer-valence orbitals will not launch any electron dynamics. However, energetically deeper ionic states demonstrate strong evidence of the hole-mixing between the orbitals of the neutral system. Accordingly, ionizing an electron from one of these inner-valence orbitals, one will unavoidably involve the other electron from the corresponding orbital in the mixture thus launching the ultrafast dynamics of the electron density9,10. The time scale of these dynamics is determined by the energy separation between the electronic states participating in the hole mixing. In case of T–A base pair, the energy separation between the states is found to be ~0.4 eV which corresponds to ~10.5 fs oscillation period. In C–G, the states are separated by 0.165 eV which leads to slower dynamics with ~25 fs period of oscillations.Fig. 1Charge migration dynamics in thymine–adenine nucleobase pair.a Ionization spectrum of thymine–adenine nucleobase pair. It is seen that the inner ionic states of the system experience strong hole-mixing effect between the molecular orbitals HOMO–5 and HOMO–6 depicted in (b). c The evolution of the electron difference density triggered by the sudden removal of an electron from HOMO–5 orbital of the system. It is seen that the localized ionization launches the electronic oscillations between molecules in the pair.Fig. 2Charge migration dynamics in cytosine–guanine nucleobase pair.a Ionization spectrum of cytosine–guanine nucleobase pair. Similarly to the case of thymine–adenine depicted in Fig. 1, the ionic states of the system experience strong hole-mixing effect between the molecular orbitals HOMO–1 and HOMO–2 depicted in (b). c The evolution of the electron difference density triggered by the sudden removal of an electron from HOMO–1 orbital of the system. In contrast to Fig. 1, the initiated electron dynamics primarily takes place in a single molecule in the pair. Charge migration dynamics in thymine–adenine nucleobase pair. a Ionization spectrum of thymine–adenine nucleobase pair. It is seen that the inner ionic states of the system experience strong hole-mixing effect between the molecular orbitals HOMO–5 and HOMO–6 depicted in (b). c The evolution of the electron difference density triggered by the sudden removal of an electron from HOMO–5 orbital of the system. It is seen that the localized ionization launches the electronic oscillations between molecules in the pair. Charge migration dynamics in cytosine–guanine nucleobase pair. a Ionization spectrum of cytosine–guanine nucleobase pair. Similarly to the case of thymine–adenine depicted in Fig. 1, the ionic states of the system experience strong hole-mixing effect between the molecular orbitals HOMO–1 and HOMO–2 depicted in (b). c The evolution of the electron difference density triggered by the sudden removal of an electron from HOMO–1 orbital of the system. In contrast to Fig. 1, the initiated electron dynamics primarily takes place in a single molecule in the pair. Investigating the nature of the orbitals involved in hole mixing in T–A and C–G base pairs, we found that in both cases these orbitals are delocalized on the molecules forming the pairs, as shown in Figs. 1b and 2b. While this effect has not been reported for DNA nucleobases so far, orbital delocalization in hydrogen-bonded systems has been previously observed in water dimers60 and clusters61. This delocalization suggests that hydrogen bonds are not merely simple electrostatic interactions, but can also involve electron correlation between distant molecules. The presence of the strong correlation between inner electrons of non-covalently bound system gives rise to many important phenomena in nature such as the interatomic and intermolecular Coulombic decay62,63 and electron transfer mediated decay64, to name just a few. In DNA nucleobase pairs, electron correlation and relaxation effects lead to hole mixing between delocalized molecular orbitals shared by the molecules in a pair. As a result, the sudden removal of an electron from one of these orbitals can trigger the oscillations of the electron density between the individual molecules in the pair. To the best of our knowledge, the existence of the hole-mixing and the associated charge migration dynamics in non-covalently bound parts of the system has not been reported to date. We simulated the dynamics of the electron density in T–A and C–G base pairs following the sudden removal of an electron from the highest occupied molecular orbital (HOMO)–5 (i.e., the orbital five levels below the HOMO) and HOMO–1 orbitals, respectively. Figures 1c and 2c demonstrate the evolution of the electron difference density:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta Q\left({{\bf{r}}},t\right)=Q\left({{\bf{r}}},t\right)-Q\left({{\bf{r}}},0\right),$$\end{document}ΔQr,t=Qr,t−Qr,0,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\bf{r}}}$$\end{document}r denotes the coordinate of an electron and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t$$\end{document}t is time variable. The corresponding hole density \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$Q\left({{\bf{r}}},t\right)$$\end{document}Qr,t has been computed as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$Q\left({{\bf{r}}},t\right)={\sum }_{p,q}{\varphi }_{p}^{* }\left({{\bf{r}}}\right){\varphi }_{q}\left({{\bf{r}}}\right){N}_{{pq}}(t),$$\end{document}Qr,t=∑p,qφp*rφqrNpq(t),where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\varphi }_{p}({{\bf{r}}})$$\end{document}φp(r) and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\varphi }_{q}({{\bf{r}}})$$\end{document}φq(r) are the molecular orbitals of the neutral system, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${N}_{{pq}}(t)$$\end{document}Npq(t) is the so-called hole density matrix58,59. Accordingly, the positive values of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta Q\left({{\bf{r}}},t\right)$$\end{document}ΔQr,t quantity indicate that the electron density increases over time in certain regions of space, while the negative values denote the decrease of the electron density with respect to the initial one. In case of T–A base pair, it is very well seen (see Fig. 1c) that the electronic oscillations take place between T and A molecules. In case of C–G pair, shown in Fig. 2c, the contrast of intra-molecular oscillations is found to be rather weak, which originates from weaker hole-mixing between the corresponding molecular orbitals (see Fig. 2a). Our theoretical predictions provide us with strong evidence that the electron correlation in nucleobase pairs can lead to ultrafast intra-molecular dynamics of the electron density between the molecules in the pairs. However, the complexity of the nucleobase pairs, and DNA and biomolecules in general, prohibits the complete fully quantum analysis and simulations of such systems. To overcome this obstacle, we describe in section 5 the perspective to utilize ultrafast scanning tunning microscope technique which can help to unravel the quantum dynamics in large biologically relevant systems. The development of ultrafast scanning electron microscope Recently, substantial progress has been made toward extending microscopy techniques into the ultrafast domain by integrating them with femtosecond and terahertz light sources. Early efforts focused on ultrafast electron diffraction (UED) and ultrafast transmission electron microscopy (UTEM), which enabled femtosecond-resolved visualization of atomic and lattice motion in molecules, solids, and nanostructures, providing transformative insight into nonequilibrium structural dynamics and phase transitions65–67. These techniques established a new paradigm of “four-dimensional” microscopy, combining nanometer spatial resolution with femtosecond temporal resolution. However, UED and UTEM intrinsically probe nuclear motion through elastic or inelastic scattering and are therefore primarily sensitive to atomic rearrangements rather than the underlying electronic dynamics that initiate chemical processes3. Since charge migration unfolds on sub-femtosecond to few-femtosecond timescales, preceding nuclear motion, the temporal resolution of electron-pulse–based microscopies remain insufficient to directly resolve purely electronic dynamics in molecules. In parallel, STM has undergone a major transformation through the development of time-resolved and light-driven variants. Pump–probe STM schemes based on electronic pulses initially achieved nanosecond to picosecond resolution and were successfully applied to spin and charge relaxation phenomena at surfaces40,68. A major advance came with the introduction of terahertz-driven STM (THz-STM), in which single-cycle or sub-cycle THz pulses are coupled into the tunneling junction, enabling sub-picosecond temporal resolution while preserving atomic spatial sensitivity69,70. THz-STM has provided direct access to ultrafast carrier capture, orbital dynamics, and electron–phonon coupling at the nanoscale, representing a significant milestone for ultrafast scanning probe microscopy68,69,71. Nevertheless, the achievable temporal resolution in THz-STM is ultimately constrained by the optical cycle of the THz field, typically limiting the effective time resolution to hundreds of femtoseconds—still orders of magnitude slower than the intrinsic timescale of charge migration. To overcome this limitation, visible and near-infrared few-cycle laser pulses have been integrated with STM, giving rise to lightwave-driven and photon-induced tunneling microscopy. These approaches demonstrated that femtosecond and even sub-femtosecond tunneling currents can be generated and coherently controlled by tailoring the carrier-envelope phase of ultrashort optical pulses72–74. Landmark experiments have achieved angstrom-scale spatial resolution combined with few-femtosecond or sub-femtosecond temporal sensitivity, enabling the visualization of orbital-specific electronic coherences in single molecules adsorbed on surfaces74. More recently, lightwave-driven scanning tunneling spectroscopy has further extended these concepts to ultrafast energy-resolved measurements at atomic defects and quantum materials75. Despite these remarkable advances, imaging charge migration in complex molecular and biomolecular systems remains a fundamental challenge in the field. First, most ultrafast STM techniques rely on repeated pump–probe averaging and indirect observables, which complicates the reconstruction of transient electron density in real time. Second, thermal effects, signal-to-noise limitations, and junction stability impose severe constraints when attempting to isolate attosecond tunneling events, particularly for weakly conducting or fragile molecular systems73,74. Third, even the most advanced lightwave-driven STM implementations to date have primarily focused on coherent tunneling or localized electronic states, rather than tracking spatially extended charge migration across molecular backbones. As a result, while ultrafast scanning tunneling and electron microscopy have reached an extraordinary level of maturity, they remain fundamentally limited in their ability to directly image charge migration in molecules simultaneously in real space and real time. Bridging this gap requires a new experimental regime capable of generating isolated attosecond tunneling probes with sufficient stability and sensitivity to map electron motion itself—rather than its nuclear or spectroscopic consequences—thereby motivating the development of next-generation quantum attosecond scanning electron microscopy platforms. The development of quantum attosecond electron microscope (Q-Attomicroscope) Earlier, we developed attosecond electron microscopy, which we coined the attomicroscope46, based on the transformation of a transmission electron microscope (TEM). This new tool enabled attomicroscopy electron imaging of graphene via electron diffraction and electron motion in biological systems76. However, capturing charge migration at the subatomic level, such as in DNA base pairs (Figs. 1 and 2), requires the simultaneous combination of attosecond temporal and angstrom spatial resolution. To this end, we are developing the quantum version of the attomicroscope (Q-attomicroscope), aimed at achieving attosecond temporal resolution within a scanning tunneling microscope (STM). The Q-attomicroscope (illustrated in Fig. 3a) is based on generating a light-induced attosecond tunneling current in a modified STM. This can be achieved either by illuminating the STM tip with optical attosecond pulses (OAPs)77–79 or by using a polarization-gated half-cycle pulses (PGPs), similar to the approach employed in the development of Attomicroscope 1.046.Fig. 3Quantum attomicroscope (Q-attomicroscope).a Schematic illustration of the operating principle of the Q-attomicroscope, which extends a scanning tunneling microscope (STM) to the attosecond regime. A half-cycle laser pulse is used to generate an attosecond electron tunneling current, enabling the imaging of photoinduced electronic and chemical dynamics initiated by a synchronized ultrashort deep-ultraviolet (DUV) pump pulse. Time-resolved STM images are obtained by scanning the sample with controlled pump–probe delay. b Principle of generating a half-cycle, polarization-gated laser pulse using ultrafast laser techniques, which is employed to drive attosecond tunneling currents in the STM junction. c Complete Q-attomicroscope experimental setup, from the laser system to the STM end station, designed to image charge migration dynamics in molecular systems. d Simulated STM image of a DNA on a ice/graphite platform which serves as a native-like environment for future charge-migration imaging. Quantum attomicroscope (Q-attomicroscope). a Schematic illustration of the operating principle of the Q-attomicroscope, which extends a scanning tunneling microscope (STM) to the attosecond regime. A half-cycle laser pulse is used to generate an attosecond electron tunneling current, enabling the imaging of photoinduced electronic and chemical dynamics initiated by a synchronized ultrashort deep-ultraviolet (DUV) pump pulse. Time-resolved STM images are obtained by scanning the sample with controlled pump–probe delay. b Principle of generating a half-cycle, polarization-gated laser pulse using ultrafast laser techniques, which is employed to drive attosecond tunneling currents in the STM junction. c Complete Q-attomicroscope experimental setup, from the laser system to the STM end station, designed to image charge migration dynamics in molecular systems. d Simulated STM image of a DNA on a ice/graphite platform which serves as a native-like environment for future charge-migration imaging. The PGP is generated by sending a 1.5-cycle laser pulse through a combination of zero- and multiple-order waveplates, producing circular polarization at the pulse edges and a linearly polarized half-cycle portion at the pulse center (illustrated in Fig. 3b). On the other side, the OAP approach enables the generation of a half-cycle pulse with temporal resolution down to approximately 400 as, based on light-field synthesis77,78. Achieving optimal contrast between the central half-cycle and adjacent field crests typically requires several optimization iterations to ensure that tunneling is strictly confined to the 400 as window and that no residual current is generated from neighboring half-cycles. For this reason, the use of few-cycle pulses alone does not yield a truly isolated attosecond current signal, but rather a train of attosecond probe current events, limiting realistic attosecond imaging. In contrast, the PGP approach is more promising, as it originates from a single laser source rather than multiple spectral channels, as required in OAP synthesis. The PGP feature is characterized by specific polarization dynamics throughout its duration: the leading and trailing edges exhibit circular polarization, whereas the central half-cycle is linearly polarized. Because electron tunneling across the STM junction is a highly nonlinear process that strictly necessitates the alignment of the electric field vector with the tip apex, the circularly polarized portions of the pulse are geometrically incapable of efficiently driving the current. Consequently, the tunneling process occurs exclusively during the peak of the linearly polarized central half-cycle enabling us to achieve sub-femtosecond-resolution tunneling current. In the light-induced current on the tip, while the plasmonic enhancement of ultra-broadband optical attosecond pulses is inherently limited for conventional Ag and Au nanotips, alternative materials such as Al exhibit a broadband optical response that encompasses the entire spectral band of the excitation pulse; additionally, Pt and Ni serve as viable tip candidates. Crucially, the methodology employed herein relies on the extreme temporal confinement of the driving field rather than necessitating a uniformly flat, ultra-broadband spectral enhancement from the tip. The system utilizes a 5 fs compressed supercontinuum beam directed through a polarization-gating setup to generate a single, linearly polarized half-cycle gating pulse (PGP), which is then focused onto the STM tip. Because the electron tunneling probability is exponentially sensitive to the instantaneous peak electric field, the tunneling current is driven almost exclusively by the intense, aligned peak of this isolated half-cycle. This mechanism ensures sub-femtosecond temporal confinement of the tunneling event, effectively decoupling the spatial resolution from the requirement of a perfectly uniform ultra-broadband plasmonic tip response. It is worth noting that optically induced tunneling currents suffer from substantial signal noise, since laser pulse amplitude noise is multiplicatively amplified in the tunneling process. However, recent developments in ultrafast quantum optics within Hassan group80,81 enabled the use of amplitude-squeezed light, which promises a significant reduction in light-induced current noise80. The Q-attomicroscope employs a time-resolved stroboscopic approach to effectively freeze time and capture attosecond electron motion in action. In this scheme, electron dynamics are initiated by a separate pump laser pulse focused on the sample to excite the targeted processes. During Q-attomicroscopy imaging experiments, images are recorded as a function of the time delay between the tunneling current probe and the pump laser pulse, controlled via a delay stage in the pump beam path. By systematically varying this delay, a sequence of snapshots is obtained and subsequently stacked to reconstruct attosecond electron movies. Next, we describe the use of polarization-gated pulses (PGP) to develop the Q-attomicroscope for imaging charge migration in DNA base pairs (T–A and C–G), which were theoretically studied and presented earlier in this article and imaged with static STM elsewhere82. The Q-attomicroscope experimental setup is shown in Fig. 3c. The system is driven by a high-power, high-repetition-rate (100 kHz) laser combined with optical parametric chirped-pulse amplification (OPCPA) to generate 10 fs pulses with an average power of 10 W. The output beam is focused into a neon-filled hollow-core fiber (HCF) to generate a broadband supercontinuum spanning 250–1000 nm. This supercontinuum is subsequently divided into two beams. The first beam is compressed using a chirped-mirror compressor to 5 fs. This beam is then reflected from a rectangular reflector mounted on a precision delay stage (Fig. 3c). This delay stage has a nanometer resolution and provides a delay steps in the order of ten attoseconds. This beam will be sent through a polarization-gating setup to generate a linearly polarized half-cycle gating pulse. The resulting PGP is focused onto the STM tip to generate a light-induced attosecond tunneling current, which serves as the probe for imaging and recording the charge-migration dynamics in DNA (Fig. 3d). The second beam is directed to the second generation of the Attosecond Light-Field Synthesizer (ALFS 2.0), which is being developed to generate a short, intense deep-ultraviolet (DUV) pump pulse. ALFS 2.0 will incorporate two DUV spectral channels. The first channel (DUV1) spans 500–350 nm, similar to the channel demonstrated in ALFS 1.078,83and can deliver up to 200 mW with a pulse duration of 5 fs. The second channel spans 350–250 nm, providing lower power but a comparable pulse duration. The two DUV channels will be coherently combined to form a 50 µJ, 3 fs DUV pulse with a central photon energy of 3.54 eV. This pulse will be tightly focused to a 5 µm spot size, yielding a peak intensity on the order of a few to tens of 1013 W/cm2. Such intensity is sufficient to initiate charge-migration dynamics in DNA base pairs via three-photon absorption, reaching excitation energies of approximately 10 eV (see Figs. 1b and 2b). Direct removal of an electron from isolated DNA base pairs (T–A or C–G) would lead to irreversible sample damage. To mitigate this, the DNA bases will be supported on a monolayer of water molecules, providing a stabilizing environment. This H2O monolayer can be realized by cooling a graphite substrate to 80 K, enabling the formation of an ultrathin water sheet. This approach provides a realistic aqueous-like environment, allowing charge migration in DNA base pairs to be studied under near-native conditions84. Notably, the Q-attomicroscope utilizes a synchronized deep-ultraviolet (DUV) pump pulse to initiate the charge migration, while the precisely delayed half-cycle PGP induces the sub-femtosecond electron tunneling current to probe the transient state. Rather than measuring a static topography—where large biomolecules typically appear as featureless structures in conventional STM—the Q-attomicroscope measures differential transient dynamics. The primary observable is the real-space and real-time evolution of the electron difference density. By systematically varying the pump-probe delay and applying a stroboscopic stacking methodology, the static topological background of the molecule is effectively subtracted out. This procedure isolates the pure, attosecond-scale redistribution of electron density moving across the molecular interface with sub-angstrom spatial precision. The ultimate goal of this proposed experimental study is to record experimentally the attosecond charge migration movies similar to the simulated dynamics presented in animations as supplementary material videos 1 and 2. It is worth noting that the development of the Q-attomicroscope presents several technical challenges. Among these, the most critical is the noise associated with the photoinduced tunneling current, which adds to the intrinsic dark noise of the STM, such as thermal noise. To mitigate laser-induced thermal effects and junction instability, the DNA base pairs will be deposited onto a monolayer of water supported by a graphite substrate cooled to 80 K. This cryogenic, aqueous-like environment efficiently dissipates heat, preventing thermal artifacts from obfuscating the tunneling current84. Furthermore, the thermal load on the tip is substantially minimized through the specific excitation parameters employed. The use of ultrashort (5 fs) laser pulses significantly increases the tip’s optical damage threshold. Additionally, by utilizing visible-range pulses at a moderate repetition rate of 100 kHz—in contrast to conventional methodologies that often employ terahertz radiation at megahertz repetition rates71—the overall heat load on the tip is further reduced. Moreover, we aim to overcome this limitation by employing half-cycle, amplitude-squeezed light, as we have recently demonstrated80,81. In quantum-light–induced tunneling, this approach enables a substantial reduction of current noise, enhancing the signal-to-noise ratio of the generated tunneling current by nearly an order of magnitude80. Hence, Q-attomicroscope introduces several critical innovations to overcome the inherent limitations of current ultrafast scanning probe techniques. First, rather than relying on repeated pump-probe averaging of indirect observables, it employs a polarization-gated pulse (PGP) from a single laser source. This configuration drives the tunneling current exclusively during a linearly polarized half-cycle, isolating a single attosecond tunneling event and enabling a time-resolved stroboscopic approach to map real-time electron density. Second, to address severe signal-to-noise constraints and prevent the multiplicative amplification of laser amplitude noise within the tunneling junction, the setup utilizes half-cycle, amplitude-squeezed light. This quantum-optics technique significantly reduces current noise, enhancing the signal-to-noise ratio by nearly an order of magnitude. Finally, by combining simultaneous attosecond temporal and angstrom-scale spatial resolution, the Q-attomicroscope is uniquely positioned to track spatially extended charge migration across molecular backbones, advancing beyond the mere observation of localized coherent electronic states. Outlook In summary, we propose the development of a Q-attomicroscope by adopting a time-resolved STM approach combined with half-cycle laser pulses to generate sub-femtosecond tunneling current signals. This capability would provide simultaneous attosecond temporal and angstrom spatial resolution, thereby opening the door to directly imaging electron motion and the initiation of chemical reactions in real time. Beyond observation, we envision that this capability will enable the use of tailored light fields45,77,78 to actively control chemical reactions—a long-anticipated goal in the scientific community. Such control over chemistry would have far-reaching implications for applications in biology and materials science. As a proof of principle for the Q-attomicroscope, we propose to image attosecond-resolved movies of charge migration in DNA base-pair molecules. To support this effort, we have performed theoretical calculations to identify the optimal experimental parameters (e.g., photon energy) and to predict the expected electron dynamics. Together, these results establish a clear pathway toward realizing attosecond-resolved, real-space imaging and control of fundamental chemical processes. Supplementary information Transparent Peer Review file Description of Additional Supplementary Files Supplementary Video 1 Supplementary Video 2 Transparent Peer Review file Description of Additional Supplementary Files Supplementary Video 1 Supplementary Video 2