Dopamine, the catecholamine neurotransmitter famous for its role in reward and motor control, is far more than a laboratory curiosity. Aberrations in dopamine signaling underpin Parkinson’s disease, schizophrenia, and attention-deficit hyperactivity disorder, and the molecule’s concentration in biological fluids carries clinical significance that extends from neurological monitoring to pharmaceutical quality control. Yet dopamine is chemically challenging to measure outside well-equipped laboratories: it is easily oxidized, it coexists in serum and plasma with electroactive interferents such as ascorbic acid and uric acid, and conventional analytical instruments like high-performance liquid chromatography and electrochemical workstations demand trained personnel, bulky equipment, and considerable expense. Colorimetric assays performed on paper promise an alternative in which the analytical result is literally visible to the eye or to a low-cost camera, making them attractive candidates for point-of-care testing in clinics, pharmacies, and resource-limited settings.

Microfluidic paper-based analytical devices, first popularized nearly two decades ago, exploit the natural wicking of fluids through patterned hydrophilic channels defined by hydrophobic barriers printed or stamped into cellulose. Reagents are pre-deposited in discrete detection zones; when a sample is applied, capillary action transports it to the reagent beds, where a chemical reaction produces a color change proportional to the analyte concentration. The appeal is obvious: paper is cheap, lightweight, biodegradable, and compatible with smartphone-based readout. The persistent frustrations, however, have been equally well documented. Standard µPAD designs consume relatively large volumes of both sample and reagent, reaction kinetics can be slow and uneven across zones, and many published designs are historically inherited rather than rationally optimized. These shortcomings have limited throughput, inflated per-test costs, and complicated quantitative reproducibility—precisely the problems the Babol team set out to attack through geometry-driven miniaturization.

The researchers’ experimental strategy began with a comparative architectural survey. Six distinct µPAD geometries were designed, fabricated, and evaluated side by side to determine how device architecture shapes the temporal development of the colorimetric signal. In a paper-based assay, geometry controls far more than footprint: it dictates the path length and cross-sectional area of the capillary channels, the wetting dynamics as fluid penetrates the cellulose fiber network, the local residence time of the analyte at the reagent zone, and the uniformity with which reaction products accumulate where the camera or eye is looking. A channel that delivers sample too quickly can outrun the color-forming chemistry; one that is too restrictive starves the reaction zone and suppresses signal intensity. By holding the chemistry constant and varying only the architecture, the team isolated geometry as the dominant optimization variable, then tracked each design’s color development profile over time as a function of signal intensity to identify the most suitable configuration.

Once the leading geometry was selected, the optimization process turned to the device’s dimensions and reaction conditions. Shrinking a µPAD is not simply a matter of scaling down a drawing. Smaller zones hold smaller reagent inventories, so the concentrations and deposition volumes of the colorimetric reagents must be retuned to ensure the reaction remains reagent-limited by the analyte rather than starved of partners. Miniaturization also sharpens the thermal and temporal profile of the assay: with less mass to heat and dry, the required drying step contracted by two minutes, accelerating the overall workflow without compromising the integrity of the immobilized reagent layer. The end result of this iterative refinement was a device in which the amount of reagent consumed per test fell by more than a factor of six relative to the conventional design, a saving that scales meaningfully when multiplied across population-scale screening campaigns.

Analytical performance was characterized rigorously in buffered solution before the device was challenged with anything as messy as serum. The optimized miniaturized platform achieved a limit of quantification of 0.0517 micromoles per liter in buffer, a sensitivity figure that places the low-cost paper sensor within the range needed for meaningful dopamine quantification in relevant biological contexts. Reproducibility across devices and detection zones was assessed to confirm that miniaturization had not introduced excessive variability—a common failure mode when paper pore structures are cut to dimensions approaching the scale of individual fiber networks. The team then interrogated selectivity, spiking the assay with a panel of potential interferents that a real biological sample might carry. Across the tested interferents, signal deviations remained within plus or minus three percent, a margin tight enough to suggest that the optimized chemistry and architecture jointly suppress cross-reactions that plague less carefully tuned colorimetric dopamine sensors.

The decisive validation came in complex biological matrices. Human serum and plasma represent the far end of analytical difficulty for small-molecule sensing: their protein content, endogenous electroactive species, and variable ionic environments can quench color reactions, alter wetting behavior, and confound calibration. The researchers spiked these matrices with known quantities of dopamine and ran the full miniaturized workflow. Recoveries of 101 to 102.5 percent—meaning the device measured back essentially all of the added dopamine, with only a one to two and a half percent deviation from the true value—indicated that the matrix had negligible effect on quantification. For a platform built from paper and costing cents per device, this level of agreement with theoretical recovery is the crucial evidence that the design is not merely a benchtop demonstration but a plausible candidate for real clinical screening.

The implications extend beyond dopamine itself. The study’s central insight—that systematic geometry screening followed by targeted dimensional optimization can deliver order-of-magnitude reagent savings and meaningful time savings without sacrificing accuracy—is architecture-independent in principle. Any colorimetric µPAD assay, whether aimed at glucose, pathogens, biomarkers of kidney function, or environmental contaminants, could plausibly benefit from the same workflow of comparative geometry evaluation, temporal signal profiling, and reaction-condition retuning. Because paper devices are already among the cheapest diagnostic formats available, cutting reagent consumption six-fold translates directly into reduced per-test cost, less chemical waste, and smaller sample volumes drawn from patients—an especially important consideration in pediatric, geriatric, and field-deployment scenarios where every microliter of blood counts.

The research also arrives at a moment of renewed momentum for paper diagnostics. Recent reviews of the field describe steady progress toward real-world deployment of µPADs, with smartphone-assisted readout, nanoparticle-enhanced colorimetric chemistries, and novel fabrication methods expanding what paper-based platforms can achieve. At the same time, the field’s critics have long noted that too many µPAD publications showcase elaborate chemistries bolted onto unoptimized architectures, leaving the device design itself as the weakest link in the analytical chain. By making geometry the headline variable and demonstrating quantitatively what careful architectural refinement buys—in reagent efficiency, speed, reproducibility, and biological-sample robustness—the Babol group offers a template for how the field can mature from proof-of-concept to practical clinical tool.

The authors, F. Ghorbani Valikchali, M. Rahimnejad, A. Ramiar, Mehdi Ezoji, and M. Ahmadnezhad, drew on expertise spanning chemical engineering and biotechnology, mechanical engineering, and electrical and computer engineering at Babol Noshirvani University of Technology, with support from the university’s Biofuel and Renewable Energy Research Center and the technology company Danesh Gostar Hamgam Ba Sanat. Their published results stop short of deployed clinical claims but explicitly frame the optimized device as a practical basis for further evaluation toward preliminary clinical screening applications. If subsequent validation studies replicate the serum and plasma recoveries at larger scale and across patient populations, the miniature paper sensor could move dopamine testing out of the central laboratory and into the pharmacy, the clinic corridor, or even the home—a small rectangle of cellulose carrying, for a few cents and a couple of minutes of waiting, a measurement that once required a room full of instruments.

Hematocrit is a routine but medically important measurement. It helps assess the blood’s oxygen-carrying capacity and can signal conditions including anemia, dehydration, hemorrhage, chronic kidney disease, leukemia, polycythemia, and certain lung disorders. Standard testing usually relies on centrifugation or automated hematology analyzers, which are accurate but require laboratory equipment. The new system instead combines a small polydimethylsiloxane, or PDMS, channel with two electrodes integrated onto a quartz substrate. A blood-derived sample travels through the channel, while an electrical potential is applied across the fluid. The resulting current changes with the concentration of red blood cells, allowing the device to estimate hematocrit.

The researchers built the electrodes using SU-8, a negative photoresist widely used in microfabrication because it can be patterned into precise structures with ultraviolet light. After exposure and development, the polymer is transformed into carbon by heating it to approximately 900 degrees Celsius in an oxygen-free environment. This process, often called carbon-MEMS fabrication, can produce chemically stable carbon structures with a broad electrochemical potential window, low background current, and relatively low overpotentials. Unlike platinum or gold, SU-8 is inexpensive and can be patterned using established semiconductor-style manufacturing techniques. The challenge is that carbonization causes the photoresist to shrink dramatically, while gases, solvents, and internal stresses can cause the emerging electrode to crack, peel away, or combust.

Conventional two-step pyrolysis has generally worked well for tall, high-aspect-ratio carbon posts, where the polymer has exposed sidewalls that allow gases and volatile compounds to escape. The planar electrodes in this study presented a different mechanical problem. Their broad bases adhered closely to the quartz substrate, while their height was small compared with their width. During heating, residual oxygen and solvent could become trapped, and stress accumulated at the polymer-substrate interface. In repeated trials using the conventional approach, the patterned SU-8 partially burned or separated from the substrate instead of converting cleanly into carbon. Cracking and peeling were especially problematic for the T-shaped electrode design, which combined micrometer-scale conducting strips with millimeter-scale circular contact pads.

To solve the problem, the team added an extended vacuum hard-baking stage before the high-temperature conversion. In the first step, patterned SU-8 was heated to 200 degrees Celsius under a continuous vacuum of approximately 10 millitorr for five to six hours, with the duration adjusted for the polymer thickness. This prolonged treatment was designed to remove residual oxygen and solvent while allowing additional cross-linking and adhesion to develop. The samples were then sealed inside evacuated quartz tubes and heated to 200 degrees Celsius for 30 minutes before being ramped to 900 degrees Celsius for one hour. Natural cooling followed the pyrolysis stage. The sealed-tube configuration also contained ash and carbonaceous byproducts, reducing contamination inside the furnace. Afterward, oxygen plasma was used to remove residue from the quartz substrate without substantially damaging the patterned carbon.

The resulting electrodes were not perfectly flat in the conventional sense. As the SU-8 contracted, surface tension and unequal shrinkage generated a concave cross-sectional profile with raised “horns” along the edges of the carbon strips. These features were more than a visual curiosity: they contributed to a rougher, three-dimensional surface that increased the area available for electrical interaction with the fluid. A 25.4-micrometer SU-8 2025 pattern shrank to approximately 5.88 micrometers after pyrolysis, a 76.8 percent reduction in height, while its final width was about 73.3 percent of the original. A thinner SU-8 2 pattern decreased from roughly 4.70 micrometers to 1.07 micrometers. Despite the extreme dimensional changes, the researchers report that the optimized process produced intact, well-defined electrodes with a 100 percent fabrication yield across the characterized batches.

The carbon electrodes were then integrated into a PDMS microfluidic device containing an 8-millimeter-long channel measuring 180 micrometers wide and 70 micrometers high. Each electrode consisted of a narrow strip connected to a 6-millimeter circular contact pad, allowing copper wires to be attached with silver epoxy. Red blood cells obtained from a volunteer donor were separated, washed, and resuspended in phosphate-buffered saline at concentrations ranging from 10 to 60 percent by volume. The researchers applied 100 volts of direct current for 30 seconds and recorded the current response at 16.13 measurements per second. The short measurement period helped limit errors caused by red blood cells settling under gravity inside the channel and reservoirs.

The electrical signal increased linearly as the red blood cell concentration increased in the phosphate-buffered saline. This behavior differs from measurements made in whole blood or blood plasma, where red blood cells generally act as insulating particles and increasing their concentration can lower the current. In the researchers’ low-conductivity buffer, however, electrical contributions from the cell membranes and surface glycoproteins produced a different overall response. The thicker SU-8 2025-derived carbon electrodes generated approximately 40.65 microamperes with a 50 percent red blood cell suspension, compared with 31.59 microamperes from the thinner SU-8 2 electrodes. The thicker design had a 13 percent larger active surface area in the relevant fluidic regions, helping explain its stronger signal.

For hematocrit determination, the SU-8 2025 system achieved 3.67 percent precision and 3.97 percent accuracy, while the thinner SU-8 2 version reached 5.37 percent precision and 4.21 percent accuracy. The stronger-performing carbon device therefore approached the reported performance of earlier systems based on carbon-coated platinum electrodes, which achieved approximately 2.8 percent precision and 2.6 percent accuracy. The comparison is significant because the new electrode requires no noble-metal coating and can be fabricated through a simpler sequence of photolithography, vacuum baking, pyrolysis, and plasma cleaning. The authors also report that a device integrating carbon electrodes cost approximately 15 U.S. dollars to fabricate in their facility, compared with about 38 dollars for an equivalent platinum-electrode device.

Microscopic surface analysis helps explain why the carbon electrodes performed so well. Atomic force microscopy showed a root-mean-square roughness of approximately 14.2 nanometers for the pyrolyzed carbon, compared with 5.29 nanometers for platinum. The carbon surface was therefore roughly 2.7 times rougher, with more pits, grains, and three-dimensional features. Its measured three-dimensional surface area in the scanned region was about 14.14 square micrometers, compared with 12.82 square micrometers for platinum. When the electrode geometry and thickness in the fluid reservoirs were included, the researchers estimated that the carbon design provided an active surface area of approximately 307,346 square micrometers, about 25 percent greater than the 245,787 square micrometers estimated for platinum. A larger electroactive area reduces the current density required to produce a given total current and can lower the charge-transfer overpotential described by the Butler–Volmer equation. The rough and textured surface may also improve local mass transport by disturbing fluid movement near the electrode.

The study does not present the device as an immediate replacement for clinical hematology analyzers, and several limitations remain. The experiments used red blood cells from a single A-positive donor suspended in prepared phosphate-buffered saline rather than a broad set of untreated patient blood samples. The system also requires a relatively high applied voltage and depends on controlling sedimentation, electrode fouling, electrolysis, and fluidic reproducibility. Even so, the work demonstrates that low-aspect-ratio carbon electrodes, once considered difficult to fabricate reliably, can be made using a practical modification of a well-known process. Because the carbon precursor is inexpensive, abundant, and compatible with precise patterning, the approach could be adapted to other electrochemical sensors, cell-analysis platforms, and disposable microfluidic diagnostic devices. By turning a photoresist into a rugged, high-area electrode, the researchers have created a route toward simpler blood-testing technologies that use less platinum while preserving much of its analytical performance.

Subject of Research: Low-cost pyrolyzed carbon electrodes integrated into a microfluidic hematocrit detection system

Article Title: Fabrication of photoresist SU-8 pyrolyzed carbon electrodes and their integration in a microfluidic hematocrit detection system

Article References: Lee, H. Y., Rogers, J. A., Kendrick, C., Habibi, S., et al. “Fabrication of photoresist SU-8 pyrolyzed carbon electrodes and their integration in a microfluidic hematocrit detection system.” Biomedical Microdevices, volume 28, article 57 (2026).

Image Credits: AI Generated

DOI: 10.1007/s10544-026-00838-6

Keywords: Carbon electrodes, microfluidic devices, hematocrit detection, pyrolysis fabrication, biosensing, SU-8, electrochemical sensors, point-of-care diagnostics