Access Official Technical PDF

Share Summary:LinkedInEmail

Glucose FET biosensors use a field-effect transistor (FET) as the electrical transducer: a chemical interaction at the sensing surface changes the channel’s conductance, which can be monitored as current. Carbon nanotubes (CNTs) and monolayer molybdenum disulfide (MoS2) are two candidate channel materials for compact sensing systems. Device development work exploring both materials illustrates an important engineering distinction: demonstrating transistor behavior under dry conditions is not the same as obtaining reliable glucose measurements in a liquid.

The work was conducted in the context of a multimodal wearable-sensor project. CNT-FET and MoS2-FET arrays were fabricated and characterized, with glucose as the intended analyte. The devices produced recognizable transistor characteristics, but glucose-sensing measurements were not successful. Their fabrication and test results offer a practical view of the issues that must be resolved before a FET sensor can be evaluated in its intended operating environment.

How field-effect biosensing works

A FET has source and drain contacts connected by a semiconducting channel, with a gate voltage controlling the channel’s charge carriers. In a biosensor, the channel surface is prepared to interact with the target analyte. Binding or a chemical reaction can alter the carrier population, changing channel conductance and the measured drain-source current.

For selective glucose detection, one approach is to immobilize glucose oxidase (GOx) near the channel. The enzyme catalyzes a reaction involving glucose; the resulting chemical changes can influence the electrical behavior of the sensing system. With suitable calibration and controlled conditions, changes in current can be assessed against analyte concentration. The gate arrangement may vary: some FET biosensors use a back gate, while others use a liquid gate or combine approaches.

Two common transistor measures help describe device behavior. The threshold voltage indicates the gate voltage associated with turning the channel on, while the ON/OFF ratio compares current in conducting and nonconducting states. These characterize transistor operation; by themselves, they do not establish glucose sensitivity, selectivity or performance in a liquid sample.

Choosing between CNT and monolayer MoS2 channels

Carbon nanotube FETs

Single-walled CNTs are graphene sheets rolled into tubular structures. Depending on their structure, individual nanotubes can be metallic or semiconducting. CNT-FETs commonly show p-type behavior, in which holes act as the main carriers under the tested operating conditions. In the reported devices, CNTs were dispersed on silicon dioxide and patterned metal contacts formed source and drain electrodes.

For glucose functionalization, a pyrene-based linker, Pbase, was used to attach GOx to the CNT surface. The pyrene group associates with the nanotube through non-covalent interactions, while the linker’s other end can bond with the enzyme. This approach is intended to attach the recognition chemistry without substantially changing the CNT’s electronic properties.

Microscopy helped assess the deposited material. Dense CNT clusters were considered less suitable for a channel because their electrical behavior can be more metallic; less dense networks were preferred. A representative dry CNT-FET showed p-type transistor behavior and an ON/OFF ratio of about 102. However, devices did not survive the functionalization process, and glucose-sensing data were not obtained.

Monolayer MoS2 FETs

Monolayer MoS2 is a two-dimensional transition-metal dichalcogenide that behaves differently from bulk MoS2. The monolayer form has a bandgap and commonly exhibits n-type FET behavior, with electrons carrying current through the channel. This makes it a distinct candidate from the typically p-type CNT devices.

In the work described, MoS2 flakes were transferred to substrates before photolithography and metal-contact deposition. Some devices were damaged during lift-off, and the transfer process may also have affected device quality. A representative dry device showed n-type behavior, but its measured ON/OFF ratio was below 102 and its electrical curves were noisy.

Why liquid testing proved challenging

The central obstacle was separating current through the semiconductor channel from unwanted ionic current through the liquid. If source and drain electrodes are exposed to an electrolyte, ions in the solution can create a parallel electrical path. That path can obscure or short the channel current—the signal the FET is meant to measure.

Electrode passivation is therefore a key design requirement for liquid-phase FET biosensing. The source and drain need to be isolated from the solution while the sensing region remains accessible. Approaches discussed for this purpose include a dielectric layer over the electrodes, such as a high-κ material, two-step lithographic structures or a fluidic channel. The appropriate method depends on the device design and fabrication process.

In the CNT devices, functionalization was followed by device failure for reasons that were not established. In the MoS2 tests, solution exposure led to ionic-current concerns; the channels also had high resistance, and the electrodes were exposed. Consequently, neither material yielded successful glucose-sensing measurements in this work. These outcomes underline why dry electrical characterization and wet sensor validation must be treated as separate stages.

Practical implications for sensor development

For engineering teams developing chemical FETs, the results point to a staged evaluation process:

  • Verify the transistor first. Measure transfer and output curves in dry conditions and assess leakage, polarity and ON/OFF behavior.
  • Inspect the channel and contacts. Microscopy can reveal material distribution, clustering, transfer damage and contact-patterning issues.
  • Recheck after functionalization. Enzyme and linker processing can affect device integrity, so electrical measurements should be repeated at each relevant step.
  • Control liquid paths. Isolate source and drain electrodes to reduce the risk of ionic current bypassing the semiconductor channel.
  • Separate device operation from sensing claims. A working FET is a prerequisite, not proof of analyte response, selectivity or stable operation.

For wearable applications, further development also has to address reproducibility, stability, fabrication scalability, cost, integration with digital processing, flexible substrates and selectivity among analytes. These are development priorities, not capabilities established by the dry-device results.

Conclusion

CNT and monolayer MoS2 channels each produced the expected p-type and n-type transistor behavior, respectively, but the tested devices did not demonstrate successful glucose sensing. The practical lesson is to design the complete measurement system—not just the semiconductor channel. Functionalization compatibility, device integrity and electrode passivation must all be validated before changes in current can be confidently attributed to glucose. For industrial and wearable sensor developers alike, disciplined separation of dry-device characterization from liquid-phase testing is essential to assessing a FET biosensor’s readiness.

Access Official Technical PDF

About The Author