Machine Learning-Based Rapid Detection of Volatile Organic Compounds in a Graphene Electronic Nose.

Rapid detection of volatile organic compounds (VOCs) is growing in importance in many sectors. Noninvasive medical diagnoses may be based upon particular combinations of VOCs in human breath; detecting VOCs emitted from environmental hazards such as fungal growth could prevent illness; and waste could be reduced through monitoring of gases produced during food storage. Electronic noses have been applied to such problems, however, a common limitation is in improving selectivity. Graphene is an adaptable material that can be functionalized with many chemical receptors. Here, we use this versatility to demonstrate selective and rapid detection of multiple VOCs at varying concentrations with graphene-based variable capacitor (varactor) arrays. Each array contains 108 sensors functionalized with 36 chemical receptors for cross-selectivity. Multiplexer data acquisition from 108 sensors is accomplished in tens of seconds. While this rapid measurement reduces the signal magnitude, classification using supervised machine learning (Bootstrap Aggregated Random Forest) shows excellent results of 98% accuracy between 5 analytes (ethanol, hexanal, methyl ethyl ketone, toluene, and octane) at 4 concentrations each. With the addition of 1-octene, an analyte highly similar in structure to octane, an accuracy of 89% is achieved. These results demonstrate the important role of the choice of analysis method, particularly in the presence of noisy data. This is an important step toward fully utilizing graphene-based sensor arrays for rapid gas sensing applications from environmental monitoring to disease detection in human breath.

Redox Buffer Capacity of Ion-Selective Electrode Solid Contacts Doped with Organometallic Complexes.

While ion-selective electrodes (ISEs) with inner filling solutions are used widely, solid-contact ISEs are better suited for miniaturization and mass manufacturing. Calibration-free measurements with such electrodes require the reproducible control of the phase boundary potential between the ion-selective membrane and the underlying electron conductor. The most promising approach to achieve this goal is based on redox buffers incorporated into the ion-selective membrane. Here we introduce the theory and present experimental data for Co(III), Co(II), Ru(II), Fe(II), and Os(II) compounds that show quantitatively how the phase boundary potential at a solid contact doped with redox-active compounds is affected by weighing errors, reagent impurities, and redox-active interferents. Perhaps surprisingly, theory predicts that there is only a minimal dependence of the phase boundary potential on the ratio of the concentrations of a pure oxidized and a pure reduced compounds if those two compounds are not a redox couple. However, theory predicts that even small redox-active impurities of those compounds shift the phase boundary potential drastically. Experimentally, a surprisingly good in-batch reproducibility was observed by us and others for solid contacts prepared to contain either only the reduced or only the oxidized species of a redox couple. This can be explained by redox-active impurities and is unlikely to be repeatable when different suppliers of reagents are used or long-term experiments are performed. This work confirms that the preferred approach to calibration-free sensing is based on redox buffers that comprise the reduced and oxidized species of a redox couple in well-controlled concentrations.

Capacitive Sensing of Glucose in Electrolytes Using Graphene Quantum Capacitance Varactors.

A novel graphene-based variable capacitor (varactor) that senses glucose based on the quantum capacitance effect was successfully developed. The sensor utilizes a metal-oxide-graphene varactor device structure that is inherently compatible with passive wireless sensing, a key advantage for in vivo glucose sensing. The graphene varactors were functionalized with pyrene-1-boronic acid (PBA) by self-assembly driven by π-π interactions. Successful surface functionalization was confirmed by both Raman spectroscopy and capacitance-voltage characterization of the devices. Through glucose binding to the PBA, the glucose concentration in the buffer solutions modulates the level of electrostatic doping of the graphene surface to different degrees, which leads to capacitance changes and Dirac voltage shifts. These responses to the glucose concentration were shown to be reproducible and reversible over multiple measurement cycles, suggesting promise for eventual use in wireless glucose monitoring.

Calibration-free ionophore-based ion-selective electrodes with a Co(II)/Co(III) redox couple-based solid contact.

A high electrode-to-electrode reproducibility of the emf response of solid contact ion-selective electrodes (SC-ISEs) requires a precise control of the phase boundary potential between the ion-selective membrane (ISM) and the underlying electron conductor. To achieve this, we introduced previously ionophore-free ion exchanger membranes doped with a well controlled ratio of oxidized and reduced species of a redox couple as redox buffer and used them to make SC-ISEs that exhibited highly reproducible electrode-to-electrode potentials. Unfortunately, ionophores were found to promote the loss of insufficiently lipophilic species from the ionophore-doped ISMs into aqueous samples. Here we report on an improved redox buffer platform based on equimolar amounts of the much less hydrophilic Co(III) and Co(II) complexes of 4,4'-dinonyl-2,2'-bipyridyl, which makes it possible to extend the redox buffer approach to ionophore-based ISEs. For example, K(+)-selective electrodes based on the ionophore valinomycin exhibit electrode-to-electrode standard deviations as low as 0.7 mV after exposure of freshly prepared electrodes for 1 h to aqueous solutions. Exposure of freshly prepared ISE membranes to humidity prior to their first contact to electrolyte solution minimizes the initial (reproducible) emf drift. This redox buffer has also been successfully applied to sodium, potassium, calcium, hydrogen, and carbonate ion-selective electrodes, which all exhibit the high selectivity over interfering ions as expected for ionophore-doped ISE membranes.

Hypothermia attenuates beta1 integrin expression on extravasated neutrophils in an animal model of meningitis.

Brain injury in meningitis occurs in part as a consequence of leukocyte migration and activation. Leukocyte integrins are pivotal in the inflammatory response by mediating adhesion to vascular endothelium and extracellular matrix proteins. We have demonstrated that moderate hypothermia early in the course of meningitis decreases leukocyte sequestration within the brain parenchyma. This study examines whether hypothermia alters neutrophil integrin expression in a rabbit model of bacterial meningitis. Prior to the induction of meningitis, peripheral blood samples were obtained and the neutrophils isolated. Sixteen hours after inducing group B streptococcal meningitis, animals were treated with antibiotics, i.v. fluids, and mechanically ventilated. Animals were randomized to hypothermia (32-33 degrees C) or normothermia conditions. After 10 hours of hypothermia or normothermia, neutrophils were isolated from the blood and cerebral spinal fluid (CSF), stained for beta1 and beta2 integrins, and analyzed using flow cytometry. Cerebral spinal fluid neutrophil beta1 integrin expression was significantly decreased in hypothermic animals. Beta-1 integrins can assume a higher affinity or "activated" state following inflammatory stimulation. Expression of "activated" beta1 integrins was also significantly decreased in hypothermic animals. Beta2 CSF neutrophil integrin expression was decreased in hypothermic animals, but failed to reach significance. These data suggest hypothermia may attenuate extravasated leukocyte expression of both total and "activated" beta1 integrins.

Sialidase treatment exposes the beta1-integrin active ligand binding site on HL60 cells and increases binding to fibronectin.

The migration of neutrophils from the circulation to areas of inflammation is the result of the sequential activation of multiple cellular adhesion molecules. beta1-Integrins are cell surface glycoproteins and the class of adhesion molecules responsible for binding to the extracellular matrix. The goal of this study was to determine the contribution of glycosylation, specifically the presence of sialic acid, to beta1-integrin adhesion in a neutrophil model. beta1-Integrins on differentiated HL60 cells were remodeled by treatment with the exoglycosidases, sialidase and beta-galactosidase. beta1-Integrin activity was determined by measuring adherence to the extracellular matrix protein fibronectin. The expression of beta1-integrins, beta2-integrins and activated beta1-integrins was determined by flow cytometry. Remodeling of beta1-integrins by treatment with sialidase increased adhesion by greater than 1,000%. Flow cytometric analysis of remodeled beta1-integrins demonstrated an increased expression of the activated beta1-integrin, but only minor increases in the expression of total beta1- and beta2-integrins. We postulate that glycosidase treatment increases adhesion and expression of activated beta1-integrins by exposure of the normally hidden ligand-binding site. The glycosylation of beta1-integrins on neutrophils may act to hide the ligand-binding site in unstimulated cells thereby contributing to the affinity modulation observed in neutrophil beta1-integrin function.

Integrin expression on neutrophils in a rabbit model of Group B Streptococcal meningitis.

Products released by polymorphonuclear cells (PMNs) during an acute inflammatory response can result in diffuse tissue injury. Integrins are cell surface adhesion proteins that play a pivotal role in inflammation by allowing PMNs to adhere to the endothelium and migrate through the extracellular matrix. We examined the expression of beta1 and beta2 integrins on neutrophils from blood and cerebrospinal fluid (CSF) in an animal model of Group B Streptococcal meningitis. We further evaluated whether integrin expression correlates with pathophysiologic markers of central nervous system inflammation. Our data demonstrate that beta3 and beta2 integrin expression on circulating neutrophils does not significantly increase as a consequence of meningitis. In extravesated CSF neutrophils, a significant increase in expression of both beta1 and beta2 integrins is noted. Furthermore, a majority of the beta1 integrins on extravesated neutrophils have undergone affinity modulation. Using regression analysis, we demonstrated that increasing beta1 integrin expression correlates with decreasing CSF glucose concentration and serum/CSF glucose ratio. Regression analysis approached significance when CSF protein was compared to PMN beta1 integrin expression. Polymorphonuclear leukocytes beta1 integrin expression also showed a direct correlation to myeloperoxidase activity in brain tissue. Beta2 expression on CSF PMNs did not correlate with these markers of inflammation/sequestration. These data demonstrate integrin expression on extravesated neutrophils markedly increases during meningitis and support a role for beta1 integrins on neutrophils in the pathophysiologic consequences of meningitis.