3.1 Dynamic Temperature Response and Reversibility

To characterize the dynamic performance of the temperature control system, we monitored the time-dependent response of both temperature and nanopore ionic current during a full heating-cooling cycle between − 10 °C and 50 °C (Fig. 3). Unless otherwise stated, nanopore measurements in this study were performed in a 4 M LiCl electrolyte buffered with 10 mM Tris-HCl and 1 mM EDTA (pH 9.0) under an applied voltage of 600 mV, with a bandwidth at 50 kHz and a sampling rate at 1 MHz. Pore diameters were estimated from the open-pore current measured at 600 mV [35].

After stabilization at − 10 °C, the temperature setpoint was changed in a single step to 50 °C and held until a new steady state was reached (Fig. 3a), after which the setpoint was returned to − 10 °C to induce controlled cooling (Fig. 3b). In both directions, the measured temperature exhibited a smooth, monotonic transient response without overshoot, indicating stable closed-loop operation across the entire temperature range.

Both the heating and cooling trajectories followed single-exponential relaxations, allowing the extraction of characteristic thermal time constants (τ) to predict the time required to reach the steady state. Based on the 5τ criterion [10, 36], which corresponds to ~ 99.3% of the total temperature change, the resulting time constants were 34.6 s for heating and 45.0 s for cooling, with settling times of approximately 173 s and 225 s, respectively. The longer cooling time is consistent with the asymmetric thermal boundary conditions of the thermoelectric stage [37]. During cooling, residual heat must be continuously extracted from the sample side and dissipated through the TEC hot side into the environment, making the late-stage approach to the setpoint more sensitive to heat-sink efficiency and ambient heat exchange. In contrast, during heating, the sample temperature is raised by direct heat injection from the TEC, which is less constrained by external heat dissipation and therefore leads to a faster effective response.

The nanopore ionic current closely followed the temperature variation throughout the cycle, but with a measurable delay relative to the temperature sensor response, particularly during heating. Specifically, the current reached its steady-state value approximately 32 s after the temperature sensor stabilized, whereas during cooling, the current approached its final value with no appreciable delay observed within the experimental time resolution. This asymmetry arises from the spatial separation between the embedded temperature sensor and the nanopore sensing region and from the differences in heat transfer pathways between the solid substrate and the electrolyte solution. Heat diffusion into the solution during heating is limited by the thermal mass and low thermal conductivity of the fluid, whereas cooling benefits from more efficient heat extraction through the solid substrate. After completion of the full heating-cooling cycle, both temperature and baseline ionic current returned to their initial values, demonstrating excellent reversibility and the absence of long-term thermal drift or hysteresis.

Fig. 3The alternative text for this image may have been generated using AI.

Dynamic temperature response and reversibility of the TEC-controlled nanopore platform. Heating and cooling are induced by step changes of the temperature setpoint, showing time-dependent evolution of temperature (red) and open-pore ionic current (blue) recorded over a full cycle between − 10 °C and 50 °C. A glass nanopore with a diameter of approximately 9 nm was used. a Heating response following a step increase of the temperature setpoint, exhibiting single-exponential dynamics with a time constant τ ≈ 35 s, corresponding to a settling time of approximately 173 s based on the 5τ (~ 99.3%) criterion. The ionic current reached its steady state with an additional delay of 32 s relative to the temperature sensor. b Cooling response after switching the temperature setpoint back to − 10 °C. The temperature transient yielded a time constant τ ≈ 45 s and a settling time of approximately 225 s, while the current approached its final value with no appreciable delay relative to temperature stabilization. In both cases, the system exhibited smooth, monotonic responses without overshoot, demonstrating stable and reversible temperature control

3.2 Temperature Stability, Response Time, and Baseline Noise Across the Operating Range

We next quantified the steady-state temperature stability and dynamic thermal response of the system across a series of discrete target temperatures spanning the full operational range (Fig. 4). A glass nanopore with a diameter of approximately 9 nm was used. Starting from 20 °C, each temperature-transition experiment was performed independently by applying step changes in the setpoint, with upward steps to 30 °C, 40 °C, and 50 °C and downward steps to 10 °C, 0 °C, and − 10 °C. After each temperature transition, the setpoint was held constant to allow the system to reach steady state, during which residual temperature fluctuations were recorded over extended time windows. The system was then returned to 20 °C and allowed to re-stabilize before initiating the next measurement.

Notably, activation of the Peltier system at room temperature (20 °C) did not introduce measurable electrical interference in the ionic current recordings (Fig. 4a, b). At all stabilized setpoints, the residual temperature fluctuations remained below 0.1 °C (Fig. 4c), demonstrating excellent temperature stability provided by the feedback controller across the entire operating range. The transient temperature responses during both heating and cooling steps were well described by single-exponential dynamics, enabling the extraction of thermal time constants for each temperature transition (Fig. 4d). The extracted time constants increased with the magnitude of the temperature change, indicating that the thermal response is primarily governed by the overall heat load associated with larger temperature excursions. Excluding the 20 °C to 50 °C transition, cooling steps generally exhibited longer time constants than heating steps for comparable temperature changes, consistent with the additional requirement for heat extraction and dissipation at the TEC hot side during cooling. The transition from 20 °C to 50 °C showed a notably longer response time than expected from the linear trend, which we attribute to the larger absolute temperature gradient and reduced heat dissipation efficiency at elevated temperatures, leading to slower convergence toward the final setpoint.

In parallel, the open-pore ionic current and current noise were analyzed at each stabilized temperature under an applied voltage of 600 mV. The baseline current increased monotonically with temperature and followed an approximately linear relationship (Fig. 4e), reflecting the temperature dependence of electrolyte conductivity and ion mobility [9]. The root-mean-square (RMS) current noise also increased with temperature, exhibiting an approximately linear trend (Fig. 4f), which can be attributed to the combined effects of increased thermal noise and enhanced ionic fluctuations at higher conductivities [4, 9, 38]. Despite this increase, the noise levels remained sufficiently low (4.5–6.1 pA RMS at 50 kHz bandwidth) to preserve high signal quality for single-molecule nanopore measurements, confirming that precise temperature control can be achieved without compromising electrical performance.

Fig. 4The alternative text for this image may have been generated using AI.

Temperature stability, thermal response, and electrical characteristics across the operating range. Measurements were performed using a glass nanopore with a diameter of approximately 9 nm. a Current-voltage (I-V) characteristics and b root-mean-square (RMS) current noise measured before and after activation of the Peltier-based temperature controller (including the fans), demonstrating no measurable change in baseline current or electrical noise. c Steady-state temperature fluctuations measured at different temperature setpoints, demonstrating sub-0.1 °C stability across the full operating range. d Thermal time constants extracted from single-exponential fits of temperature transients following step changes from 20 °C to different target temperatures, illustrating the dependence of response time on the magnitude and direction of the temperature change. e Open-pore ionic current and f RMS current noise measured at an applied voltage of 600 mV with a bandwidth of 50 kHz, as a function of temperature, showing approximately linear increase with temperature

3.3 Temperature-Dependent Nanopore Transport of M13mp18 DNA Carriers

Having established the stability of the temperature control system under open-pore conditions, we next evaluated its performance during single-molecule experiments using M13mp18 DNA carrier molecules (Fig. 5a). The carriers consisted of simple double-stranded structures formed by hybridizing linearized M13mp18 single-stranded DNA with complementary oligonucleotides. Nanopore recordings were carried out at a series of controlled temperatures ranging from − 10 °C to 50 °C in 10 °C increments. For each temperature, measurements were performed at a DNA concentration of 1.5 nM, with a recording duration of 10 min, to allow reliable extraction of event statistics. Additionally, all the translocation events were classified into a linear group and a folded group, corresponding to carrier molecules entering the nanopore in a linear or folded configuration (Fig. 5a). Typically, only linear events are considered valid for further analysis, as they do not obscure any structural information of interest through molecular folding.

The event capture frequency increased systematically with temperature (Fig. 5b, Supplementary Fig. 3), consistent with enhanced molecular diffusion and reduced solution viscosity at elevated temperatures [9, 39]. In contrast, the fraction of linear translocation events remained largely constant across the investigated temperature range, centering around approximately 0.24 (Fig. 5b and Supplementary Fig. 3). This behavior indicates that, for the double-stranded carrier structures used here, temperature primarily modulates transport kinetics rather than DNA structural behavior prior to capture. At the nanopore entrance, strong local electric fields and tight geometric confinement dominate DNA conformations, enforcing an extended, field-driven configuration that is largely insensitive to thermal fluctuations over the investigated temperature range. Importantly, the nanopore remained stable across all tested temperatures, with no evidence of increased clogging or baseline instability at either low or high temperatures.

Fig. 5The alternative text for this image may have been generated using AI.

Temperature-dependent single-molecule transport of M13mp18 DNA carriers through a glass nanopore. Measurements were performed using a glass nanopore with a diameter of approximately 8.5 nm. a A schematic illustration of a double-stranded M13mp18 DNA carrier translocating through a glass nanopore. Translocation events were classified into linear and folded groups based on their molecular conformation upon entry into the nanopore. b Capture efficiency and fraction of linear translocation events as a function of temperature, with capture efficiency increasing and the linear fraction remaining approximately constant, staying close to 0.24 for the nanopore used here. Measurements were performed at a DNA concentration of 1.5 nM with a recording duration of 10 min at each temperature

Current-voltage characteristics were also measured at each temperature setpoint of the experiment to assess the effect of thermal cycling on the nanopore itself (Fig. 6). Across all temperatures, the I-V curves remained linear, indicating ohmic behavior over the test range, and the temperature dependence of the slope was consistent with the corresponding changes in open-pore conductance. Notably, I-V curves recorded at room temperature (20 °C) before temperature modulation, during the temperature series when the system passed through 20 °C, and after completion of the full temperature cycle upon returning to 20 °C were found to overlap closely. This reproducibility confirms that temperature modulation does not induce irreversible measurable changes in nanopore geometry or surface properties.

Fig. 6The alternative text for this image may have been generated using AI.

Temperature-dependent current-voltage characteristics of the glass nanopore measured during thermal cycling. Measurements were performed using a glass nanopore with a diameter of approximately 8.5 nm. The I-V curves recorded at 20 °C before temperature modulation, during the temperature series when passing through 20 °C, and after returning to 20 °C at the end of the approximate 2-hour measurement overlap closely, demonstrating that temperature cycling did not induce measurable changes in nanopore properties

3.4 Temperature Dependence of Event Duration, Signal Amplitude, and Signal-to-Noise Ratio

Finally, we analyzed the characteristics of structured carrier translocation events as a function of temperature (Fig. 7, Supplementary Fig. 4). As illustrated schematically in Fig. 7a, the designed carrier contains four reference structures, each composed of six dumbbell motifs, distributed along the carrier backbone. During nanopore translocation, the carrier backbone produces a current blockade plateau, on which the four reference structures generate four distinct current spikes. The spatial separations between the two central reference structures and between the two outer reference structures are denoted as L1 and L2, respectively. The corresponding temporal separations between the associated current spikes are defined as t1 and t2. The carrier signal and structure signal are quantified as the mean current drop of the backbone plateau and the average amplitude of the four reference peaks.

The normalized transit times, t1/L1 and t2/L2, are shown as a function of temperature in Fig. 7b. Data points were averaged over all linear translocation events recorded within 10 min at a carrier concentration of 1.5 nM at each temperature. Over the investigated temperature range from − 10 °C to 50 °C, both quantities exhibited nearly identical values at each temperature and decreased monotonically with increasing temperature. This trend is consistent with faster electrophoretic transport at elevated temperatures, driven by reduced solution viscosity and lower hydrodynamic drag acting on the DNA molecule [9, 40]. Notably, the normalized transit time at − 10 °C was approximately six times larger than that at 50 °C.

The temperature dependence of signal amplitude is shown in Fig. 7c. In each translocation event, the structure signal amplitude was obtained by averaging the amplitudes of the four reference structures. Both the carrier signal and the structure signal increased with increasing temperature, with the carrier signal exhibiting a slightly steeper increase. This behavior reflects the temperature-induced increase in open-pore conductance arising from enhanced ionic mobility, which amplifies the absolute current modulation associated with ion exclusion when DNA occupies the nanopore. Figure 7d presents the average signal-to-noise ratio for both carrier and structure signals as a function of temperature. Although baseline current noise increased with temperature, signal amplitude increased more rapidly, leading to an overall improvement in signal-to-noise ratio at elevated temperatures (Fig. 7d).

Fig. 7The alternative text for this image may have been generated using AI.

Temperature-dependent characteristics of structured DNA carrier translocation events. Error bars represent the standard error of the mean calculated from analyzed translocation events. Measurements were performed using a glass nanopore with a diameter of approximately 8 nm. a A schematic illustration of the M13mp18 carrier containing four reference structures, each composed of six dumbbell motifs. During nanopore translocation, the carrier backbone produces a current blockade plateau (carrier signal), while the four reference structures generate four distinct current spikes (structure signals). The spatial separations between the two central and the two outer reference structures are denoted as L1 and L2, with the corresponding temporal separations defined as t1 and t2. b Normalized transit times t1/L1 and t2/L2 extracted from all linear translocation events recorded over a 10 min acquisition period at a carrier concentration of 1.5 nM for each temperature, showing a monotonic decrease with increasing temperature and nearly identical values across the investigated temperature range. The numbers of recorded events were 33, 58, 74, 92, 109, 125, and 160 at − 10, 0, 10, 20, 30, 40, and 50 °C, respectively. c Temperature dependence of the carrier signal and structure signal amplitudes, both of which increase with temperature, with a slightly stronger increase observed for the carrier signal. For each translocation event, the structure signal amplitude was defined as the average of the amplitudes produced by the four reference structures. d Average signal-to-noise ratio of the carrier and structure signals as a function of temperature, showing an overall increase at elevated temperatures despite increased baseline RMS noise

3.5 Resolving Closely Spaced Structural Features Through Temperature-Controlled Translocation

To further illustrate the practical utility of temperature as a tuning parameter for nanopore sensing, we designed another M13mp18 DNA carrier featuring 60 bp overhang structures placed at defined spacings of 50 bp, 100 bp, 200 bp, and 300 bp along the carrier backbone (Fig. 8). These structures serve as model targets for evaluating the ability of the nanopore platform to resolve closely spaced features under different thermal conditions.

At 50 °C, translocation events exhibited strong current signatures with a high signal-to-noise ratio, consistent with the enhanced ionic conductance and signal amplitude observed at elevated temperatures (Fig. 7c, d). However, due to the correspondingly fast translocation speed, only the overhang pairs separated by 200 bp and 300 bp could be clearly resolved as distinct current spikes, while the 50 bp and 100 bp spacings remained unresolved within the temporal resolution of the measurement. At 20 °C, the reduced translocation velocity provided sufficient temporal separation to additionally resolve the 100 bp-spaced overhang pair, although the 50 bp spacing remained indistinguishable from a single broadened feature. Upon further cooling to − 10 °C, the carrier translocation speed decreased to approximately one-fifth to one-sixth of the value observed at 50 °C. Under this condition, even the most closely spaced 50 bp overhang pair became resolvable as two distinct current spikes. However, the reduced temperature was accompanied by a decrease in signal amplitude and signal-to-noise ratio, consistent with the temperature-dependent trends characterized in Figs. 7c and d.

These results demonstrate a fundamental trade-off between signal amplitude and temporal resolution that can be systematically navigated through temperature control. Elevated temperatures favor a high signal-to-noise ratio and robust signal detection, making them well-suited for applications where signal clarity is the primary concern. Conversely, reduced temperatures extend translocation durations and enhance temporal separation between closely spaced structural features, enabling higher spatial resolution along the molecular contour at the expense of signal strength. The ability to selectively tune the operating temperature thus provides a straightforward experimental strategy for optimizing nanopore measurements according to the specific requirements of interest, whether prioritizing signal-to-noise performance or fine-grained structural discrimination.

Fig. 8The alternative text for this image may have been generated using AI.

Temperature-dependent resolution of closely spaced structural features on M13mp18 DNA carrier. Measurements were performed using a glass nanopore with a diameter of approximately 8 nm. The carrier design incorporates 60 bp overhang structures with inter-structure spacings of 50 bp, 100 bp, 200 bp, and 300 bp. Representative translocation events recorded at 50 °C, 20 °C, and − 10 °C illustrate the trade-off between signal amplitude and temporal resolution. At 50 °C, a high signal-to-noise ratio is achieved, but only the 200 bp and 300 bp spacings are resolved (red circles). At 20 °C, the 100 bp spacing becomes additionally resolvable. At − 10 °C, translocation is slowed by approximately five- to six-fold relative to 50 °C, enabling resolution of the 50 bp spacing, albeit with reduced signal amplitude and signal-to-noise ratio