Glucose Sensors

Abstract

As we have discussed in Chap. 1 there are four major types of transducers for biosensors: (1) optical, (2) electrochemical, (3) piezoelectric, and (4) thermal. The top two most popular transducers are optical and electrochemical. In reality, these two types are also the ones most commonly used for a glucose sensor, which is one of the most common and the most commercially successful biosensors up to date. In this chapter, we will learn about two different types of glucose sensors, optical and electrochemical, along with other types of enzymatic biosensors.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

As we have discussed in Chap. 1 there are four major types of transducers for biosensors: (1) optical, (2) electrochemical, (3) piezoelectric, and (4) thermal. The top two most popular transducers are optical and electrochemical. In reality, these two types are also the ones most commonly used for a glucose sensor, which is one of the most common and the most commercially successful biosensors up to date. In this chapter, we will learn about two different types of glucose sensors, optical and electrochemical, along with other types of enzymatic biosensors.

12.1 Optical Glucose Sensor

A glucose sensor (or a glucose meter ) is a device that monitors the level of blood glucose in humans. Glucose is a simple sugar that exists in the blood stream and comes from the food humans eat. It is essentially a fuel that energizes the cells in a human’s body. It is important to maintain a proper level of blood glucose, and in fact, the human body regulates it tightly at a range between 65 and 104 mg/dL (this number can rise up to 140 mg/dL after taking food). The blood glucose level can serve as indicators for variety of medical conditions, but most importantly, this is an indicator for diabetes . Blood glucose level higher than 126 mg/dL after 8 h of fasting (typically measured in the morning before breakfast) is diagnosed as diabetes. Most diabetic patients aim to maintain the blood glucose level lower than 120 mg/dL after 8 h of fasting (i.e., in the morning). Diabetic patients need to monitor the blood glucose level several times a day, so that they can properly control the disease. As it is impractical for the diabetic patients to visit the hospital or laboratory on a daily basis, a need has emerged to develop a handheld, easy-to-use, glucose sensor .

The very first glucose sensor, or rather an assay kit and a reflection photometer (reflectometer), was introduced in the 1970s. The assay utilizes test strips, known as Dextrostix , and the accompanying meter was called Ames Reflectance Meter (Fig. 12.1). Although this system came with a really high price tag and was primarily used in hospitals only, it basically opened up an entirely new market for glucose sensors and possibly all biosensor markets as well.

Fig. 12.1

Ames reflectance meter that reads from dextrostix . Reprinted from Newman & Turner. Biosens. Bioelectron. 20: 2435-2453, © Elsevier 2011, with permission from Elsevier

This glucose sensor is essentially an optical biosensor. In a Dextrostic/Ames Reflectance Meter, glucose in a blood sample is first oxidized into gluconolactone , and eventually into gluconic acid (Fig. 12.2), under the existence of an enzyme (biological catalyst) called glucose oxidase (GOx). Oxygen (O₂) dissolved in blood is necessary to carry out this oxidation reaction. The byproduct of the reaction is hydrogen peroxide (H₂O₂).

$$ {\text{glucose}} + {\text{O}}_{ 2} \xrightarrow{{{\text{GOx}}}} {\text{gluconic}}\;{\text{acid}} + {\text{H}}_{ 2} {\text{O}}_{ 2} $$

(12.1)

Fig. 12.2

Glucose (left); Gluconolactone (middle); Gluconic acid (right)

In optical glucose sensing, the byproduct, H₂O₂, is the molecule being detected. For the Ames Reflectance Meter:

$$ {\text{H}}_{ 2} {\text{O}}_{ 2} + {\text{benzidine}}\mathop{\longrightarrow}\limits{{}}{\text{H}}_{ 2} {\text{O}} + {\text{oxidized benzidine (blue)}} $$

(12.2)

This color change (colorless to blue) can be read visually using glucose test strips (Dextrostix ), which can be quantified into glucose concentration in mg/dL unit using Ames Reflectance Meter, using the same concept we learned in spectrophotometry.

Although modern glucose meters no longer use this optical transduction schematic (they use electrochemical transducers—see the next section), most glucose assay kits still utilize similar optical transduction schematics, as they offer better sensitivity and reproducibility.

Choosing a right transducer is a very important aspect for biosensor development. Electrochemical biosensors can be made very small and inexpensive, typically at the cost of poor sensitivity and poor reproducibility. Optical biosensors are generally considered superior in sensitivity and reproducibility than electrochemical biosensors, at the cost of device complexity and considerable cost for fabrication. Recently, significantly improved sensitivity has been demonstrated for electrochemical biosensors, through utilizing next generation electrode materials, as well as newer types of enzymes and cofactors. Specifically, the introduction of nanotechnology toward electrochemical biosensors has provided their sensitivity and reproducibility comparable to those of optical biosensors (see Chap. 15).

12.2 Electrochemical Glucose Sensor

As discussed in the previous section, glucose sensing involves oxidizing glucose with GOx while generating hydrogen peroxide as a byproduct (Eq. 12.1). During this process, GOx itself is reduced. If we add a ferricyanide ion Fe(CN₆)³⁻ (oxidized form) to the reduced GOx, then GOx is oxidized again while ferricyanide is reduced to ferrocyanide Fe(CN₆)⁴⁻ (reduced form). Basically, GOx and Fe(CN₆)³⁻ are swapping electrons.

$$ {\text{GOx}}\left( {\text{reduced}} \right) + {\text{Fe}}\left( {{\text{CN}}_{ 6} } \right)^{ 3- } \to {\text{GOx }}\left( {\text{oxidized}} \right) + {\text{Fe}}\left( {{\text{CN}}_{ 6} } \right)^{ 4- } \left( {\text{reduced}} \right) $$

(12.3)

The extra electron from ferrocyanide Fe(CN₆)⁴⁻ can be given back to the other electrode.

$$ {\text{Fe}}\left( {{\text{CN}}_{ 6} } \right)^{ 4- } \left( {\text{reduced}} \right) \to {\text{Fe}}\left( {{\text{CN}}_{ 6} } \right)^{ 3- } \left( {\text{oxidized}} \right) + {\text{e}}^{ - } \left( {\text{given to the electrode}} \right) $$

(12.4)

This cycle generates an electrical current (with no voltage applied; not practical) or change in electrical current (with voltage applied) (Fig. 12.3).

Fig. 12.3

Electrochemical detection of glucose using GOx and ferricyanide ion

In reality, GOx does require a cofactor to truly function as a catalyst. Flavin adenine dinucleotide (FAD⁺) is a well-known cofactor for GOx. During the oxidation of glucose to gluconolactone by GOx, FAD⁺ works as the electron acceptor and is reduced to FADH ₂ .

$$ {\text{glucose}} + {\text{GOx}}{-}{\text{FAD}}^{ + } \to {\text{gluconolactone}} + {\text{GOx}}{-}{\text{FADH}}_{ 2} $$

(12.5)

The cofactor FADH₂ is oxidized back to FAD⁺ utilizing O₂ found in blood, while forming H₂O₂.

$$ {\text{GOx}}{-}{\text{FADH}}_{ 2} + {\text{O}}_{ 2} \to {\text{GOx}}{-}{\text{FAD}}^{ + } + {\text{H}}_{ 2} {\text{O}}_{ 2} $$

(12.6)

H₂O₂ is oxidized back to H₂O at an electrode, generating electrons, and this whole cycle again generates a change in electrical current with a constant voltage applied (Fig. 12.4).

Fig. 12.4

Electrochemical detection of glucose using GOx and FAD⁺

Other than GOx, different enzymes can also be used in electrochemically detecting glucose. For example, the enzyme glucose dehydrogenase (GDH) and a cofactor nicotinamide adenine dinucleotide (NAD⁺) can function similarly to GOx–FAD⁺ combination. In this schematic, oxygen is not required (major benefit over GOx–FAD⁺), and GDH converts glucose to gluconolactone while converting NAD⁺ into NADH (reduced form):

$$ {\text{glucose}} + {\text{GDH}}{-}{\text{NAD}}^{ + } \to {\text{gluconolactone}} + {\text{GDH}}{-}{\text{NADH}} $$

(12.7)

$$ {\text{GDH}}{-}{\text{NADH}} \to {\text{GDH}}{-}{\text{NAD}}^{ + } + {\text{H}}^{ + } + 2 {\text{e}}^{ - } $$

(12.8)

The reduced NADH will be oxidized back to NAD⁺ at the electrode, while generating H⁺ and two electrons. This process will again result in a change in electrical current with constant voltage applied.

The use of pyrroquinoline quinone (PQQ) as a cofactor to GDH has become increasingly popular in electrochemical glucose sensing, due to its rapid electron transfer rate.

$$ {\text{glucose}} + {\text{GDH}}{-}{\text{PQQ}}\left( {\text{oxidized}} \right) \to {\text{gluconolactone}} + {\text{GDH}}{-}{\text{PQQ}}\left( {\text{reduced}} \right) $$

(12.9)

Again, the reduced PQQ will be oxidized back to its oxidized form at the electrode, similar to NAD⁺, and a change in electrical current will be observed with constant voltage applied (Fig. 12.5).

Fig. 12.5

Electrochemical detection of glucose using GDH and NAD⁺/PQQ

This amperometric electrochemical transduction was very popular in the 1980s and quickly became the mainstream in glucose sensing. Electrodes are patterned into a test strip with the necessary enzymes and chemicals immobilized on it. Electrodes are made out of platinum (Pt), gold (Au), or carbon (C), with a reference electrode, so that a small change in current can be measured, usually in µA to nA scale. As it is somewhat difficult to measure this small current change in a reproducible manner, the sensitivity and reproducibility of electrochemical glucose sensor s are inferior to those of optical glucose sensors. However, as the system can be made very small and inexpensive, it has quickly dominated the glucose meter market. Many different companies currently manufacture glucose meters and test strips, and use a wide variety of reaction schematics.

Figure 12.6 shows how the commercial glucose meter/strip measures the blood glucose level: (1) sampling a small volume of blood from the patient’s finger using a finger-pricking device (referred to as lancet device), (2) inserting a glucose strip to its meter, pre-loaded with enzyme (GOx or GDH) and cofactor (FAD⁺, NAD⁺, or PQQ), (3) contacting the inlet of the strip to the blood droplet on a finger, allowing the blood to be absorbed into the strip, and (4) measuring the current change under a constant voltage applied. Diabetic patients need to monitor their blood glucose level several times a day, while the most important reading is the one measured in the morning before taking any food (i.e., after 8 h of fasting). The assay results can be written down in a log book, or can be stored in the small memory of a meter. Some newer versions of glucose meters allow communication with a smartphone for this data logging.

Fig. 12.6

A lancet device pricks the finger to obtain a blood sample (top left and top right). A glucose meter reads the current change from a glucose strip (pre-loaded with enzyme and cofactor; bottom left and bottom right)

12.3 Other Electrochemical Biosensors

Since successful demonstration and commercialization of electrochemical glucose sensors, other types of electrochemical enzymatic sensors have been developed. One of the early attempts was cholesterol detection from blood, another important indicator for variety of medical conditions. In cholesterol detection, an enzyme called cholesterol oxidase (ChOx ) oxidizes cholesterol into cholesterol-4-ene-3-one utilizing O₂, while the enzyme itself is reduced, generating H₂O₂. This schematic is identical to Eq. 12.1 (Fig. 12.7).

Fig. 12.7

Electrochemical detection of cholesterol using ChOx

Similarly, the level of ethanol in blood (blood alcohol content or BAC) can be detected, for various legal and medical purposes (most notably for identifying drunk driving). An enzyme called alcohol dehydrogenase (ADH) and a cofactor NAD⁺ oxidizes ethanol into acetaldehyde without requiring O₂. This schematic is identical to Eqs. 12.7 and 12.8 (Fig. 12.8).

Fig. 12.8

Electrochemical detection of ethanol using ADH–NAD⁺

Other detections are also possible: Lactic acid can be detected with lactic oxidase; Uric acid can be detected by uricase; Urea can be detected by urease, etc.

12.4 Continuous Glucose Monitoring (CGM)

While the procedure shown in Fig. 12.6 has been the gold standard in self-monitoring blood glucose level, the patient can obtain only a few data points per day. For some cases, there is a need to monitor the blood glucose level throughout the entire day, e.g., to continuously monitor the blood glucose level, to figure out the impact of certain type of food or the effect of certain exercise, etc. Therefore, new technology has emerged: continuous glucose monitoring (CGM) . In CGM, an electrode needs to be inserted into the body (to enable in vivo measurements), and the glucose level is measured every 1–5 min (Fig. 12.9).

Fig. 12.9

Continuous glucose monitoring (CGM) devices, wired (left) and wireless (right)

Initially, attempts were made to quantify the glucose level directly from blood in a continuous manner. To do this, the electrode is inserted into the capillary blood vessel, and measurements are made every 1–5 min, using the same electrode. Unlike the single-use glucose strips, however, the electrodes are easily contaminated with proteins, especially with blood clots (creating the situation called thromboembolism , formation of thrombus  = stationary blood clot, and embolus  = circulating blood clot). To avoid this complication, most recent CGM devices measure the glucose level from interstitial fluid (=tissue fluid). A needle is inserted through the skin, and the electrode is located at its end, to measure the glucose level of the interstitial fluid (=tissue fluid) right underneath the skin (subcutaneous measurement) . The needle-type electrode needs to be replaced before the enzyme and the cofactor become inactive or the electrode becomes contaminated, approximately every 3–7 days.

The earlier models require a physical connection between the needle-based electrode and the meter. The newer models have an emitter fixed to the skin, which transmits the information in a wireless mode to the meter located up a few meters away. In addition, CGM device can be integrated with an insulin pump . Based on the measured glucose level, the meter can trigger the insulin pump to inject insulin to the patient, to lower the blood glucose level as needed. Note that diabetes is caused by lack of insulin production (type 1) or insulin resistance (type 2), and injecting insulin to patients is considered as later-stage treatment option.

12.5 Laboratory Task 1: Glucose Assay Kit with a Spectrophotometer

In the first two tasks, you will use a glucose assay kit that utilizes a spectrophotometer. This particular kit does not utilize an enzyme GOx, but directly binds glucose to a chemical o-toluidine under heating and acid conditions. The resulting glucose–toluidine complex has blue-green color. The visible spectrum consists of, from short to long wavelengths, blue-green–yellow-red color. Therefore, coloration of blue-green indicates that absorption occurs yellow-red color from a liquid container (cuvette). The maximum absorption peak occurs at 630 nm.

This reaction is noticeably simpler and easier to use, at a cost of possibility for cross reaction. For example, o-toluidine can also bind to other aldosugars such as glyceraldehyde, ribose, or galactose. However, as the concentrations of these chemicals in blood are much lower than that of glucose, this cross reactivity can be often neglected. You can obviously use other glucose assay kits, provided that you follow the manufacturer’s protocol line-by-line.

In this task, you will need the following:

• Glucose

• Electronic balance, weighing paper, laboratory spatula

• Deionized or distilled water

• Centrifuge tubes (1.5 mL)

• Pipettes and pipet tips (1000 μL)

• A vortex mixer

• Glucose assay kit (QuantiChrom™ DIGL-100 from BioAssay Systems)

• A heating block (temperature sensor, if needed)

• A spectrophotometer (CHEMUSB4 or FLAME-CHEM from Ocean Optics) and appropriate software (OceanView™ from Ocean Optics)

• Disposable plastic cuvettes

• Latex gloves, delicate task wipers (KimWipes®)

Preparation of Solutions

• Dissolve 30 mg of glucose in 1 mL of deionized or distilled water in a 1.5-mL centrifuge tube. Use a vortex mixer to dissolve. This will make a 3 g/dL glucose solution.

• Take 100 μL of 3 g/dL glucose solution and add to 900 μL distilled and/or deionized water. This will make a 300 mg/dL standard glucose solution (STD).

• Dilute the standard in water using 1.5-mL centrifuge tubes as follows.

  No.	STD + water	Vol (μL)	Glucose (mg/dL)
  1	150 μL + 0 μL	150	300
  2	100 μL + 50 μL	150	200
  3	50 μL + 100 μL	150	100
  4	25 μL + 125 μL	150	50
  5	0 μL + 150 μL	150	0

• If you are trained and qualified to use/handle biological specimen and the laboratory is approved for its use, you may obtain blood sample from human subjects, using a lancet device (see Fig. 12.6).

• As this is not a common and easy process, we recommend to take one of the above solutions (#1–#5) from the other team, without knowing its concentration, as your unknown sample.

Colorimetric Detection of Glucose

• Transfer 12 μL of the above to the other centrifuge tubes. Transfer 1200 μL QuantiChrom™ Reagent to each tube. Close the tubes tightly and mix with a vortex mixer.

• Place the tubes in a tube holder and heat (100 °C) in a heating block for 8 min (Fig. 12.10). Cool down in cold water for 4 min.

  Fig. 12.10

  

  A heating block warms the centrifuge tubes close to 100 °C

• Transfer an appropriate amount of reaction mixture into a cuvette. Read the absorbance at 620–650 nm (peak absorbance at 630 nm) against the blank (Figs. 12.11 and 12.12). The signal will stay stable for 60 min.

• Subtract blank absorbance (water, #5) from the standard absorbance values (#1–#4), and plot the absorbance against standard concentrations. Determine the slope using linear regression fitting. The y-intercept should be set to zero. This is your “standard curve” (Fig. 12.11).

• Using this standard curve, determine the glucose concentration of the “model” urine specimen.

• Refer to Chap. 8 for definitions of absorbance A = log₁₀ (I ₀/I), where I ₀ is the light intensity before entering the material and I after passing through the material, and Beer–Lambert law A = εlc, where ε is the molar absorptivity, l is the path length, and c is the concentration of substance.

• Theoretically, A at 0 mg/dL concentration should be zero, making the standard curve to pass through origin. In reality, however, the cuvette, water, and the QuantiChrom™ Reagent absorb light, leading to positive y-intercept in the standard curve. Although it is okay to use this standard curve as is, one may find it more useful to zero-adjust the entire curve by subtracting with A at 0 mg/dL concentration (Fig. 12.11).

  Fig. 12.11

  

  Experimental data from Task 1: absorbance—glucose concentration plot (filled diamond symbols). The glucose concentration of an unknown sample can be calculated from the regression equation: 0.68 = 0.0036x + 0.281, x = 110 mg/dL. The bottom curve (open square symbols) is zero-adjusted by subtracting with the absorbance of 0 mg/dL solution

  Fig. 12.12

  

  Absorbance spectrum for glucose assay kit

12.6 Laboratory Task 2: Glucose Assay Kit with LED /PD Circuit

In this task, the UV/Vis spectrophotometer will be replaced with a red LED and a photodiode, along with accompanying circuit, as described in Chap. 8 (Figs. 12.13 and 12.14). The same glucose solutions from Task 1 will be used.

In this task, you will need the following:

• A breadboard, wires, wire cutter/stripper, a power supply and a DMM

• 10 or 20 kΩ pot, and a screw driver

• Three 100 Ω and two 1 MΩ resistors

• Red LED

• PIN-040A photodiode

• Op-amp LM741 (or LM324)

• Five samples in cuvettes from Task 1 (glucose + QuantiChrom™ reagent + heat-incubated)

  Fig. 12.13

  

  Circuit diagram of Task 2

  Fig. 12.14

  

  LED–PD circuit is measuring the glucose concentration in a cuvette. Left 300 mg/dL. Right 0 mg/dL

As V _out from a PD circuit will be divided by that of 0 mg/dL glucose solution, not I ₀, the A at 0 mg/dL automatically becomes 0. This will make the standard curve pass through origin, or zero-adjusted (Fig. 12.15).

As shown in Fig. 12.14, the cuvettes are aligned to have 0.5 cm path length, and compared to 1 cm path length in Task 1, the absorbance should be roughly a half of those in Task 1 (consider the Beer–Lambert law, A = εlc, where l is the path length).

Fig. 12.15

Experimental data of Task 2: log (V _out,0/V _out)—glucose concentration plot. V _out,0 is the voltage output with 0 mg/dL glucose solution

Question 12.1

Compare absorbance readings (zero-adjusted) of Task 1 and Task 2. Probably, the absorbance readings of Task 2 are slightly smaller than a half of those of Task 1. What factors are responsible for this deviation?

12.7 Laboratory Task 3: Commercial Electrochemical Glucose Sensor

In the following task, commercial glucose test strips will be used, which are based on electrochemical detection. Recent glucose test strips are mostly based on GDH–PQQ, although the exact reaction schematics may vary by manufacturer.

In this task, you will need the following:

• Commercial glucose meter and test strips (Accu-Chek from Roche)

• Glucose stock solution (3 g/dL) from Task 1

• Deionized or distilled water

• Centrifuge tubes (1.5 mL)

• Pipettes and pipet tips (10 or 100 μL)

• A vortex mixer

• Latex gloves, delicate task wipers (KimWipes®)

• You still have leftover glucose stock solution (3 g/dL). Prepare five different glucose solutions, following the Task 1 protocol. Do not add QuantiChrom™ reagent.

• Take a test strip from the bottle of new test strips and insert it into a glucose meter to activate.

• Measure glucose concentrations for the five glucose solutions, by applying a small drop of each to the end of a test strip.

• Repeat the measurement twice for each solution. Calculate the averages and standard deviations (Figs. 12.16 and 12.17).

• If possible, try to measure the glucose level of your own blood. Use a lancet device included in the commercial glucose meter.

  Fig. 12.16

  

  Experimental data of Task 3: commercial sensor reading—glucose concentration plot

  Fig. 12.17

  

  A commercial glucose meter is reading the glucose level

Question 12.2

Is there discrepancy between the real concentrations and your meter readings? If so, can you explain why? Hint: many commercial glucose meters show “plasma equivalent” concentrations rather than the actual glucose concentrations.

Question 12.3

When comparing the colorimetric assay kit and the commercial glucose meter, which one shows a larger standard deviation? Can you explain why?