A Compact Instrument for Temperature-Programming-Assisted Capillary–Nanoliquid Chromatography

Abstract

The miniaturization of liquid chromatography (LC) columns to capillary and nanoscales allows temperature programming to be an effective alternative to solvent gradients for modulating eluotropic strength. This approach simplifies instrument design and operation, as a single pump can suffice to achieve efficient separations. This study presents the development and application of a compact, lab-built high-pressure system for temperature-programmed capillary and nanoLC separations. The instrument includes a high-pressure capillary–nanoflow syringe pump, a time-based nanoliter injection system, a programmable capillary column oven for controlled temperature gradients, and a UV-Vis detection system with a custom nanoliter-scale detection cell. Each system component was designed and built in-house, with rigorous calibration to ensure accuracy and operational reliability. Experimental data confirm the system’s capability to deliver precise, reproducible temperature, and flow rates. Functionality was validated through temperature-programmed separations on packed and open tubular capillary columns. The results demonstrated that the developed instrument offers enhanced separation efficiency and reduced analysis time compared to isothermal methods, underscoring its potential for advanced applications in miniaturized liquid chromatography.

1. Introduction

Liquid chromatography (LC) is an essential analytical technique extensively employed across various fields, including pharmaceuticals, environmental science, food safety, and omics research. The conventional high-performance liquid chromatography (HPLC) system, typically utilizing columns with inner diameters (i.d.) ranging from 2.1 to 4.6 mm, has been the workhorse of analytical laboratories for decades [1]. However, the miniaturization of the technique—capillary LC (i.d. 0.1–0.5 μm) and nanoLC (i.d. <0.1 μm)—offers several significant advantages over traditional HPLC, particularly in terms of chromatographic efficiency, sensitivity, and solvent consumption [2,3].

Compared to conventional HPLC, miniaturized LC techniques, such as capillary and nanoLC, offer enhanced analytical performance, particularly regarding chromatographic efficiency and detection sensitivity [4]. While particle size remains a key factor influencing column efficiency, reducing the column’s inner diameter is critical in improving chromatographic performance. A smaller inner diameter enhances chromatographic efficiency in several ways. Firstly, it reduces the radial dispersion of the chromatographic band, resulting in narrower eluted peaks. Additionally, in particle-packed columns, the interactions between the tube wall and the packed bed become more significant at smaller diameters, leading to a more ordered chromatographic medium and, consequently, more uniform chromatographic bands [5]. Moreover, monolithic nanoLC columns can be prepared with smaller domain sizes and uniform structures, making them more efficient than their conventional-scale equivalents [6]. Finally, nanoLC can also exploit open tubular columns, which do not exhibit Eddy dispersion due to multiple path effects, potentially outperforming their packed counterparts [7].

Capillary LC and nanoLC also operate at significantly lower flow rates, resulting in less on-column dilution and more concentrated eluted bands. This improves detectability with spectrophotometric and mass spectrometry detectors, even from minuscule sample amounts [8]. Miniaturized LC is an invaluable tool in molecular omics strategies because it handles complex samples and detects minute analyte quantities from small sample volumes. This capability is particularly critical for achieving rapid, sensitive, high-throughput analyses, which are essential in such applications [9]. However, the applicability of miniaturized LC extends beyond omics sciences; any molecule analyzable by conventional HPLC can be analyzed using capillary or nanoLC [10].

Environmental sustainability is another critical advantage of miniaturized LC. With growing concerns over the environmental impact of laboratory practices, the ability of capillary and nanoLC to drastically reduce organic solvent consumption—up to 1000-fold for capillary LC and 10,000-fold for nanoLC. This is achieved by transitioning from traditional flow rates of 0.5–1.0 mL/min to much lower rates of 100 nL/min to 5 μL/min. Such reductions result in substantial cost savings in solvent procurement and waste management and support industry-wide initiatives to reduce ecological footprints [9].

Although all these features were well established early on, making capillary and nanoLC highly desirable in many research areas, the development of this technique has been slow and strongly limited by the challenges associated with managing reduced extra-column volumes and very low flow rates [11]. Extra-column band broadening becomes much more critical at the reduced scale, and micro/nanoflow rates require precise, accurate, and pulseless solvent delivery instrumentation [2,3]. The pumps must produce low flow rates (nL–µL/min) with precision and accuracy. The connecting tubes must have low extra-column volume, and the sample introduction devices and detectors must be compatible with capillary columns to avoid dispersion of the chromatographic band.

Despite these challenges, ongoing efforts to address the limitations of miniaturized LC are promising. Advances in instrument design and user training are making the technology more accessible. Over the past 20 years, significant progress has been made in developing instruments designed explicitly for miniaturized LC, resulting in various commercially available instruments, columns, and consumables [2,3]. Additionally, capillary and nanoLC facilitate the miniaturization of instruments, leading to the emergence of diverse small-footprint and portable LC systems [12]. Portable miniaturized LC instruments, characterized by their simplicity and cost-effectiveness, are gaining traction in various fields. These portable systems offer rapid, on-site analysis capabilities, which are handy for environmental monitoring and field-based studies [13]. In such applications, portable miniaturized LC instruments help reduce issues related to sample degradation and contamination during transport, ensuring more reliable and timely data.

An additional, less explored advantage of miniaturized liquid chromatography is using temperature programming as a selectivity modulation factor to enhance efficiency and resolution. Selectivity can be fine-tuned in isocratic elution mode by applying temperature gradients. Isocratic elutions eliminate the need for column reconditioning between analyses and utilize straightforward instrumentation, as they can be performed with a single pump. This simplicity facilitates the design and construction of more accessible LC instruments.

A change in the analysis temperature causes pronounced effects on several parameters, such as viscosity, solubility, diffusivity, and vapor pressure, influencing column efficiency, resolution, back pressure, selectivity, retention time, peak shape, and the properties of stationary phases [14]. In conventional LC, heat flow through the column’s cross-section is inefficient, making temperature programming unfeasible. However, heat flow is fast and uniform through the chromatographic bed at the miniaturized scale, particularly with capillary–nanoLC columns (i.d. <300 µm). Consequently, the mobile phase rapidly responds to temperature changes, making temperature programming a viable alternative [15].

Temperature programming in nanoLC has been previously demonstrated multiple times [16]. For instance, a 5 °C increase in column temperature can be comparable to a 1% acetonitrile change in the mobile phase, and a 3.75 °C rise is similar to a 1% methanol variation [17]. Therefore, the effects of temperature programming in miniaturized LC are akin to those observed when applying a solvent gradient. This advantage is particularly useful in designing and constructing novel miniaturized LC instruments. By using temperature programming, separations can be conducted in isocratic mode, thus simplifying the instrument by eliminating the need for complex gradient pumping systems and significantly reducing costs.

This paper describes the construction and performance of a lab-made, compact instrument designed for temperature-programmed nanoLC. This small-footprint platform comprises a high-pressure syringe-type pump, a miniaturized oven capable of performing isothermal and temperature programming, and a capillary UV-Vis detector compatible with capillary and nanoLC. By detailing this instrument’s design, construction, and application, we aim to demonstrate its potential as a viable alternative to conventional HPLC for routine analytical purposes. The successful implementation of this technology could pave the way for broader adoption of miniaturized LC in various industries, promoting more sustainable and efficient analytical practices.

2. Materials and Methods

2.1. Reagents

The PAH standards—naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, chrysene, and dibenzo(a,h)anthracene—were supplied by Supelco, while the alkylbenzene standards—toluene, ethylbenzene, butylbenzene, pentylbenzene, and heptylbenzene—were purchased from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions of each standard were prepared at a concentration of 500 mg/mL and stored at −18 °C. Intermediate solutions were prepared through dilution of these stocks, with concentrations adjusted based on the response of the analytical instruments.

HPLC-grade solvents, including methanol (MeOH) and acetonitrile (ACN), were provided by Tedia. Formic acid (analytical grade) was sourced from Sigma-Aldrich, while acetic acid (analytical grade) and tetrahydrofuran (THF, 99%) were obtained from J.T. Baker. Merck supplied isopropyl alcohol (IPA), and high-purity water was produced using a Milli-Q purification system from Millipore.

2.2. Chromatographic Columns

Laboratory-prepared C18-packed capillary columns were manufactured in polyether ketone (PEEK) tubes of 0.25 µm i.d., using a particle suspension technique. First, a suspension containing 16 mg of 3.5 μm C18 particles (Cromatorex, Nagoya-shi, Japan) and 0.7 mL of a THF/IPA solution (1:6 v/v) was stirred and transferred into a packing column connected to a Shimadzu pumping system. Methanol was pumped at 200 bar for one hour to ensure adequate packing. After this, the pump was turned off, and the column was left in place for an additional hour to ensure complete depressurization. The column was then washed with MeCN for 30 min.

A wall-coated open tubular column (WCOT) was prepared using fused silica tubing (50 μm i.d., 375 μm o.d., 2.30 m length) from Polymicro Technologies (Phoenix, AZ, USA). The tubing was pretreated by flushing it with a 2% v/v hydrochloric acid (HCl) solution (Tedia, Fairfield, CT, USA) under a nitrogen flow at 80 bar for 5 min. It was then placed in an oven at 220 °C for 3 h. Following this, the tubing was washed with water and dichloromethane, 1.5 mL each, and dried with a nitrogen stream—a dehydration step involved heating the column at 200 °C for 2 h under a continuous nitrogen flow. Silanization was achieved using hexamethyldisilazane (HMDS) from Sigma-Aldrich (St. Louis, MO, USA), with the tubing heated at 290 °C for 6 h. A stationary phase coating (50% methyl and 50% phenyl polysiloxane) was applied by percolating a 2% w/v solution of the stationary phase through the column for 30 min. The column was then heated at 40 °C for 15 min. Film immobilization (0.25 μm thickness) was achieved by percolating azo-tert-butane (ATB) through the column for 5 min and heating at 220 °C for 1 h. This immobilization process was repeated three times to ensure stable film adherence.

A porous layer open tubular column (PLOT), a 5 m long tube with a 25 μm internal diameter, was also prepared using a polystyrene divinylbenzene (PS-DVB) stationary phase. Tube pretreatment began by filling it with a 1 mol/L NaOH solution, followed by placement in an HPLC oven at 60 °C for 20 h. The tube was subsequently washed with H₂O and MeCN for 30 min each and dried with N₂ gas for 30 min. For silanization, the tube was treated with a solution of 40 μL γ-MAPS and 10 mL of 6 mol/L acetic acid. The capillaries were filled with this solution, sealed, and heated in an HPLC oven at 60 °C for 20 h. The tubes were rewashed with H₂O and ACN for 30 min each and dried with N₂ gas. The polymerization solution was prepared by mixing 5 mg of AIBN, 200 μL of styrene, 200 μL of divinyl benzene, and 600 μL of ethanol in an Eppendorf tube followed by 5 min of ultrasonic sonication. After filling, the tubes were sealed and placed in a thermal bath at 74 °C for polymerization. Finally, the PS-DVB PLOT column was washed with H₂O and ACN for 30 min each and dried with N₂ gas for at least 1 h.

2.3. Capillary–NanoLC Instrument Development

2.3.1. High-Pressure Syringe Pump

For the construction of the high-pressure syringe pump (Figure 1), a 5-phase PK564AW hybrid stepper motor with 500 steps per revolution (0.72° by step) and a built-in planetary reduction gearbox (50:1) was purchased from Oriental Motors (Tokyo, Japan), as part of the commercial package RK564AA-N50. An R16-FSI-191-237 ball screw with a 16 mm diameter, 2 mm pitch, and equipped with a nut was acquired from HIWIN (Taiwan). Various bearings were sourced from the local market, including a ball bearing model 16003 2Z, an angular contact bearing model 7203 BEP, a tapered roller bearing model 32006 X/Q, an angular contact bearing model 7205 BECBM, and two ball bearings model 619/6 2Z.

A rectified 316 stainless steel piston, 5 mm in diameter and 50 mm in length, was manufactured upon request. A U-R134MB-(I)-(5.00-2.50)-GFP 55-316BRT-SOW-(L 4.42) seal, consisting of a graphite-reinforced Teflon seal with a 316 stainless steel inclined spiral spring, was custom-made by Bal Seal (Foothill Ranch, CA, USA). Other structural components, such as the pump holder, piston chamber, spacer cylinder, and pump cover, were manufactured from stainless steel at the mechanical services of the São Carlos Institute of Chemistry (IQSC/USP).

For automated piston filling, a 4-position valve model EMTCST4UW, with a microelectronic actuator, was purchased from VALCO Instruments (Houston, TX, USA). A strain gauge-type pressure sensor, model EPB-C11-5KP (working range of 0–350 bar and input power at 10 VDC with non-amplified output), was purchased from Measurement Specialties (Hampton, VA, USA). Optical start/end sensors, model C860NP, and mechanical microswitch sensors were sourced locally in São Carlos (SP, Brazil).

An 8-bit PIC16F877A microcontroller was purchased from Microchip (Chandler, AZ, USA) for electronic control. A 16-bit AD7705 analog-to-digital converter (ADC) and an AD9833 waveform generator were acquired from Analog Devices (Norwood, MA, USA). An RKD514L-A stepper motor driver, with two selection switches for adjusting the motor’s electric current supply, was purchased from Oriental Motors (Tokyo, Japan). The communication conversion used by the PIC16F877A microcontroller for USB communication was made through a TATO-USB2 adapter developed by TATO Equipamentos Eletronicos (São Paulo, Brazil).

Finally, the software developed to control the pumping system was programmed in Visual Basic 6 (Microsoft) using Active X components from Measurement Studio 6 (National Instruments). The pump control form was developed as a DLL (Dynamic Link Library), programmed using Visual Basic version 6.

2.3.2. Nanoliter Volume, Stop-Flow Injection System

The injection system featured a four-port valve with an internal loop EPCI4W.06 model from Valco (Houston, TX, USA). This model consists of a valve with a 60 nL internal loop of the “W” type, featuring 1/16” connections, a 0.25 mm column port, and remaining ports with a 0.40 mm diameter. The electronic setup for valve actuation includes a 24 VDC power supply (powered by 110/220 VAC), a medium torque EPCA-CE controller module, a manual controller, and an RS232-USB converter, all connected to the EPCI4W.06 valve via an EHMA model microelectronic actuator and related interconnection cables.

The valve control is managed by a DLL programmed in Visual Basic 6. This DLL accepts three inputs—volume, time, and flow—and one command: injection. Since the injection volume/time in the timed injection technique depends on the flow rate supplied by the pump, the DLL recalculates the injection time whenever a new flow rate is set to maintain the specified volume. Upon receiving the injection command, the DLL directs the valve to execute the injection using the latest configuration.

2.3.3. Capillary–NanoLC Column Oven

The capillary–nanoLC furnace (Figure 2) was designed with six key components: a metal body, thermal insulation, a ceramic tube, a cooling system, a resistive heating element, and a temperature sensor. The furnace body was constructed from a stainless steel box measuring 280 mm in length, 40 mm in height, and 40 mm in depth. Thermal insulation was provided by Fiberfrax^®, a silica and alumina fiber blanket known for its ability to withstand high temperatures, with a maximum tolerance of up to 1430 °C. This material also offers low thermal conductivity and minimal heat storage.

The heating mechanism comprised a ceramic tube through which the column passed. This tube was externally wrapped with a 0.102 mm diameter, 4 m long nickel–chromium wire featuring a resistivity of 1.0 × 10⁻⁶ Ω·m. This configuration ensured uniform heat distribution, with the heater delivering 24 W of thermal power when powered at 110 V.

A rapid cooling system was integrated, utilizing two 25 × 25 mm fans connected to the furnace via ducts. Additionally, the cooling system featured a door mechanism composed of a bronze shaft with two Celeron disc-shaped doors operated by a Tower Pro SG90 micro-servomotor. This motor was responsible for opening and closing the doors, assisting in efficient temperature regulation.

The oven’s electronic system comprised four critical circuits: a heater power controller, a temperature sensing circuit, a power supply with mains voltage monitoring, and a driver circuit for activating the cooling system (including door servos and DC fan motors). The furnace’s overall operation was managed using a PIC16F877A microcontroller, which communicated with the PC via an RS232 interface.

Temperature monitoring was achieved using a DM-503 PT-100 platinum resistance temperature detector (RTD) from Labfacility Ltd. (Bognor Regis, UK), with a Class B tolerance. This sensor could measure temperatures up to 500 °C with an accuracy of around 0.5 °C within the 0–100 °C range. In an RTD sensor, the resistance varies with temperature, and its constant current excitation generates an output voltage proportional to the temperature. This signal was then amplified and digitized by the ADC integrated into the microcontroller.

The heater power was regulated with a control circuit using a MAC8M TRIAC transistor from On Semiconductor. This transistor, mounted on heat sinks, could handle loads of up to 8 A at alternating voltages up to 800 V. It was controlled with a MOC3021 optocoupler from Fairchild, which consisted of a GaAs LED emitter and a light-activated bilateral switch. This ensured reliable and isolated control of the heating element. A synchronization circuit was also implemented to align TRIAC operation with mains cycles, reducing electromagnetic interference, which can be a significant noise source in high-precision laboratory environments.

2.3.4. Miniaturized LC UV-Vis Detection System

A Shimadzu SPD-M10A detector was integrated with a CBM-20AD system controller and Class VP software 1.0 for miniaturized UV detection. A custom detection cell was built to match the dimensions of the conventional cell in the SPD-M10Avp detector, ensuring compatibility while reducing the detection volume (Figure 3).

The cell was constructed using anodized aluminum nesting blocks that supported the optical path, which ran through a central hole in the structure. A stainless steel tube (250 µm i.d) was placed between the blocks to guide the fused silica capillary, serving as the optical conduit. The optical path was formed from a U-shaped fused silica capillary tube with a 75 µm i.d. (225 µm o.d.) and a length of 31 cm. This capillary was carefully molded into two 90-degree bends using a torch, creating an optical path length of approximately 7.92 mm (35 nL).

For optical alignment, hemispherical focal lenses from a commercial capillary cell (LC Packings) were used to ensure accurate light transmission through the capillary. This custom design effectively reduced the detection volume while maintaining high sensitivity for nanoscale liquid chromatography analysis.

3. Results

3.1. Performance of the Developed Capillary–NanoLC Pump

The developed pumping system demonstrated excellent performance for high-pressure capillary and nanoliquid chromatography, operating up to 350 bar with an operational flow ranging from 50 nL/min to 250 µL/min. Flow calibration was initially conducted using the gravimetric method, where the pump was filled with water and set to specific flow rates. Effluent aliquots were collected every five minutes in 0.5 mL Eppendorf flasks and weighed on an AG285 analytical balance (Mettler Toledo). As illustrated in Figure 4, across the nine tested flow rates, flow accuracy ranged between 99.9% and 101.71%, with the highest deviation observed at the lowest flow rate (100 nL/min). This minimal deviation demonstrates the pump’s high precision and ability to maintain set flow rates with minimal fluctuation. The accuracy range indicates the pump performed excellently, with deviations within ±1.71% across most tested flow rates. The slight increase in deviation at the lowest flow rate was expected due to the inherent challenges of controlling such small volumes. However, accuracy significantly improved above 500 nL/min, confirming the pump’s stability and precision at typical operational conditions. This performance eliminates the need for software flow correction, ensuring that the flow outputs closely match the set flow rates.

Performance in actual operational conditions was tested using a 250 µm i.d. ×15 cm capillary column packed with 3.0 µm C18 particles, with mobile-phase flow rates (50/50 ACN/H₂O) set at 1, 3, and 5 µL/min. After system stabilization, indicated by steady column pressure, uracil (a non-retained compound) was injected every 10 min for flow estimation. The injection of uracil, which does not interact with the stationary phase, provided accurate flow estimates, allowing for the direct measurement of the pump’s flow rate without interference from sample retention. The average flow rates obtained were 1.01 µL/min with a relative standard deviation (RSD) of 0.97%, 3.00 µL/min with an RSD of 0.83%, and 4.99 µL/min with an RSD of 0.73%.

3.2. Performance of the Injection Volume System

The injected volume was calibrated by estimating the volumes delivered by the valve across 23 different injection times evaluated based on the peak areas of pravastatin and rosuvastatin from injections of a standard solution (Figure 5A). The injection volume exhibited linear behavior only in the 0 to 300 ms range, corresponding to approximately 50% (30 nL) of the maximum sampling capacity of the valve’s internal loop (60 nL). In the 300 to 5000 ms range, the injected volume per unit of time decreased exponentially, and for times longer than 5000 ms, the injection volume reached the maximum of 60 nL.

These results are consistent with the literature, which reports that up to 50% of the loop volume can be used linearly in partial-loop or time-based injection methods. Beyond this point, dilution within the valve causes the exponential decay observed after 300 ms. A Nelder regression was applied based on the experimental data, and the resulting parameters were incorporated into the valve control software. Figure 5B illustrates the regression results for a 3 μL/min flow rate, while Figure 5C shows the injection system’s performance after calibration.

3.3. Performance of the Temperature-Programmable Capillary–NanoLC Oven

The thermal power provided by the oven is directly proportional to the effective voltage applied to the heating element. However, when the oven is in operation, the thermal power required at any given moment is the sum of the power naturally dissipated by the oven and the energy needed to heat the column at the user-specified rate (heating slope), following the correlation P = P_dissipated + P_{slope+adjusted}. To determine these values, experiments were necessary to measure the dissipated power across a temperature range of 0 to 100 °C and the power required to achieve different heating slopes. These parameters were experimentally determined and incorporated into the control software. When the user programs a temperature or temperature ramp, the system automatically calculates the required power for the heating element.

The oven’s performance was assessed under both isothermal conditions and temperature ramping. Figure 6a shows the results of 25 min isothermal runs at 60 °C to assess temperature control precision. The system maintained a temperature accuracy of 0.001 °C, demonstrating high precision. Similar experiments were conducted across a temperature range from 0 to 140 °C, with standard deviations of recorded temperatures consistently below 0.01 °C. Figure 6b illustrates the linear relationship between the user-programmed temperature and the experimentally measured temperature, confirming the oven’s accuracy.

Additionally, Figure 6c presents the calibration curve for estimating the power required to perform temperature ramps with different slopes, measured in °C/min. The developed system exhibited precise and reproducible temperature control during ramping experiments. Figure 6d shows an example of a run comprising both isothermal and temperature gradient segments, with a standard deviation of less than 0.0026 °C in temperature measurements. This demonstrates the system’s ability to execute complex temperature programming reliably.

3.4. Performance of the Miniaturized Detection Cell

To assess the performance of the developed UV-Vis miniaturized LC detection cell, a Shimadzu SPD-M10Avp diode array detector, controlled by a CBM-20AD system and LCsolution software 1.0, was employed. The developed cell’s performance was compared to the commercial detection cell initially included in the detector. As shown in Figure 7, the retention times of the analytes remained nearly identical when both the commercial microcell and the lab-made cell were used. Although the internal diameter (i.d.) of the tubing in the commercial cell is not provided by the manufacturer, it is likely very similar to the i.d. of the capillary used in the optical path of the lab-made cell. While the adjusted retention times (t_r’) of the analytes did not vary significantly between the two cells, the reduced peak dispersion achieved with the lab-made cell led to higher analytical efficiency and improved chromatographic resolution.

3.5. Comparison of the Developed System with a Commercial Miniaturized LC Instrument

To compare the performance of the developed system with that of a commercial miniaturized LC instrument, the contribution of the extra-column volume (ECV) to band broadening was measured using the linear regression method. Specifically, the extra-column variance (σ²ec) was determined by injecting a solution containing uracil, naphthalene, phenanthrene, anthracene, and pyrene at a 5 μL/min flow rate. The mobile phase, composed of ACN/H2O (70:30 v/v), was optimized to yield retention factors (k) between 1 and 5 for the analytes.

The detection system was configured with the shortest response time (0.2 s) and the highest acquisition rate allowed (12.5 Hz) to minimize external influences on the measurements. Using the data obtained, a plot of column variance (σ²col) versus the retention volumes (VR²) of the PAH peaks was constructed (Figure 8). The intercept of the linear regression line from these plots corresponds to σ²ec. At the same time, the slope is inversely related to column efficiency (1/N).

The calculated σ²ec values were 0.0044 μL² for the homemade system and 0.0362 μL² for the commercial instrument. These values correspond to ECVs of 0.27 μL and 0.76 μL, respectively. Additionally, the intrinsic efficiencies of the columns were determined to be 17,943 and 11,087 theoretical plates, respectively.

The developed system maintained an extra-column variance contribution within the commonly accepted threshold of 10% of the total peak variance, indicating competitive performance in preserving column efficiency compared to a commercial miniaturized LC instrument.

3.6. Chromatographic Evaluation of the Fully Developed System

The performance of the developed instrument was evaluated for the separation of PAH mixtures using packed (250 µm o.d.) and alkylbenzene compounds using WCOT and PLOT open tubular columns (50 and 25 µm i.d., respectively), showcasing its versatility for both conventional miniaturized LC and OTLC separations.

Initial assessments were conducted using a C18-packed capillary column (14 cm × 250 µm × 3.5 µm). Figure 9 presents the chromatographic separation of a mixture of eight PAHs performed in both isothermal mode and under a temperature gradient. In part (a), the analyte was separated at a constant temperature of 30 °C. In part (b), separation was conducted with a temperature gradient, where the temperature was held at 30 °C for the initial 8 min and then linearly increased at a rate of 8 °C/min until reaching 85 °C, which was subsequently maintained for the next 30 min.

For comparative analysis, parameters N, N/m, H, k, h, and E were calculated based on the last retained peak (dibenzo(a,h)anthracene), while Rs and α were derived from the acenaphthene and fluorene peaks (

Table 1. Parameters extracted from the chromatograms in Figure 8, comparing isothermal and temperature-programmed separation modes.

Parameter	Isothermal	Temperature Programming
wb (50%) Phenanthrene (min)	0.817	0.2
N (Theoretical Plates)	14,439	48,554
N/m (Plates per Meter)	103,136	346,813
H (Height Equivalent to a Theoretical Plate, μm)	9.7	2.88
R (Resolution)	1.59	1.56
α (Selectivity)	1.35	1.25
k (Retention Factor)	13.56	5.67
v (Linear Velocity)	1.42	2.82
h (Reduced Plate Height)	2.82	0.82
Ø (Flow Resistance)	1785	1488
E (Separation Impedance)	13,700	1010
ΔP (Pressure Drop, bar)	120	100

). Retention times (t_r) revealed complete separation under both conditions; however, the isothermal method required substantially longer total analysis time (40 min). Under temperature programming, the t_r value for the last eluted peak was reduced by approximately 22 min, halving the analysis time.

Efficiency improvements were notable, with the temperature gradient increasing theoretical plates by about 34,000, equating to approximately 243,000 plates per meter. This represents a threefold increase in column efficiency. Temperature programming also impacted k and α, as both parameters decreased under the gradient. Further, temperature influenced retention time, which decreased as temperature rose, thereby increasing the linear velocity (v). A marked reduction in h, from 2.82 to 0.82, was also observed with TP application.

Temperature additionally affected column pressure, with a backpressure reduction of approximately 20 bar. This influenced both flow resistance (Ø) and separation impedance (E), with Ø reducing to 291 and E decreasing by 12,690—an excellent result for coated capillary columns. Reduced parameter analysis confirmed high column performance, further enhanced by temperature programming.

Temperature programming was also applied to WCOT and PLOT open tubular columns, demonstrating its efficacy in reducing analysis time and enhancing peak intensity and column performance. For this purpose, a WCOT OTLC-4 column (50 μm i.d.) and an OT-PS-DVB polymer column (25 μm i.d.) were used to separate a mixture of five alkylbenzenes and uracil as a dead time marker. For the separations on the WCOT column (Figure 10), a mobile phase of ACN/H₂O (60:40 v/v) was employed at a flow rate of 5 μL/min, with analyte standards at a concentration of 1.4 μL/mL, and detection at 210 nm. The isothermal run was maintained at 30 °C, while for temperature programming, the oven was set to increase from 30 °C to 100 °C over 14 min at a rate of 5 °C/min.

When using the PLOT column, ACN/H₂O (40:60 v/v) was used as the mobile phase at a flow rate of 3 μL/min, with detection at 210 nm. The isothermal run was also conducted at 30 °C, and for temperature programming, the column temperature was held at 30 °C before increasing to 80 °C over 10 min at 5 °C/min. The resulting chromatograms, showing alkylbenzene separations in both isothermal and temperature-programmed modes, are presented in Figure 11.

As illustrated, temperature programming significantly reduced the analysis time for both open tubular columns and generally increased peak intensities. However, the gain in signal was less pronounced than with packed columns. In isothermal mode, the WCOT column could not resolve the last eluted peak; however, temperature programming allowed for well-defined shapes for all six alkylbenzene peaks. Additionally, retention time (t_r) reduction was accompanied by narrower peak widths, increasing the WCOT column efficiency from 972 to 2442 theoretical plates for the heptylbenzene peak—a 2.5-fold enhancement, indicating the value of temperature programming in open tubular liquid chromatography.

The same trend was observed with the PLOT columns, where retention time reduction and peak narrowing improved signal intensity for the later eluting analytes. For example, efficiency for the pentylbenzene peak doubled with temperature programming compared to isothermal conditions.

It is essential to highlight that the detectability observed with the PLOT column was approximately five times higher than that of the WCOT column. This difference can be attributed to variations in on-column dilution during the elution process, which is inversely proportional to the square of the column’s radius and length [2]. In this case, the PLOT column, with nearly half the radius and length of the WCOT column, facilitated the elution of more concentrated analyte bands. This explains the differences in noise levels and detectability observed in Figure 9 and Figure 10.

The success of temperature programming in capillary LC demonstrates its potential as an alternative to solvent gradient methods, reducing analysis time and enhancing column performance while simplifying instrumentation by eliminating the need for multiple pumps and solvent gradient equipment.

4. Conclusions

Developing and validating a temperature-programmable capillary–nanoLC system comprising a high-pressure syringe pump, time-based nanoliter injection system, programmable oven, and custom detection cell demonstrated considerable potential for miniaturized liquid chromatography applications. The results show that the syringe pump achieved precise flow rates from 50 nL/min to 250 µL/min, with flow accuracy surpassing 99.9%, ensuring reliable performance across a broad operational range. The nanoliter injection system exhibited predictable, linear injection behavior up to 50% of the valve loop capacity, while the calibration of the system further improved its accuracy.

The programmable capillary oven achieved consistent and precise temperature control, even under temperature gradients, with deviations below 0.01 °C. Temperature programming proved to be an effective means of modulating eluotropic strength, offering an efficient alternative to solvent gradients by achieving high reproducibility and consistent heating.

Testing the complete system on packed and open tubular capillary columns confirmed that temperature programming considerably enhanced separation efficiency, reducing analysis time by up to 50% while maintaining robust chromatographic resolution. In packed columns, it led to a 3-fold increase in theoretical plates; in open tubular columns, a 2.5-fold efficiency improvement was achieved, demonstrating the suitability of this approach for high-efficiency separations in miniaturized LC.

In conclusion, this work highlights the feasibility and advantages of temperature programming in capillary and nanoLC, offering a simplified, high-performance solution for miniaturized chromatography without the complexity of solvent gradient systems. This instrument holds promise for diverse analytical applications.