Figure 1.Comparison between (a) labeled and (b) label-free detection methods.One advantage of using labeled detection methods is that the secondary antibody provides dual-confirmation of the presence of the protein, reducing false-positive readings. However, since the secondary antibody introduces an additional time-consuming step, labeled detection methods are not suitable for rapid and real-time sensing applications.1.2. Sensor Overview and Performance MetricsThere are many types of integrated sensors and various approaches for categorizing them. One method is to use the physical transduction mechanism to create classes of integrated sensors. If this method is used, three distinctly different types of sensors are quickly apparent: electrical, mechanical, and optical [1�C3,5�C8].

An overview of the detection mechanisms and specific examples are shown in Table 1, respectively. However, it is important to note that this table is not meant to be comprehensive, but simply gives the reader a sense of the breadth of research which has been performed in the field. Each sensor was originally demonstrated off-chip, and gradually migrated to an integrated format, also referred to as a Lab-on-Chip. For example, one of the first optical sensors was based on an optical fiber, in which the change between the input power and output power was used as the detection signal [9]. Later, integrated optical sensors were developed using waveguides, resonators, and other on-chip approaches [10�C16].Table 1.Summary of different sensors, detection mechanism, and examples of detection.

Additional details on each detection modality are found in the subsequent sections.Because of the numerous types of sensors, fundamental metrics were developed for comparing device performance. They are related to the response or behavior of the device. In the present review, we will focus on six of these metrics; however, for the interested Carfilzomib reader there are several articles and textbooks which can provide in-depth discussions on sensor theory [4,138].The key performance metrics include the signal, noise level, signal to noise ratio (SNR), linear range (working range), response time and rate, and false-positive/false-negative rate (selectivity). For clarification, Figure 2 shows an idealized version of a sensor in operation. The signal describes the output signal (S) which is generated with a given input or measurand (Figure 2a).

In the linear range of the sensor, this relation is S = a + bs (a = background noise level, b = sensitivity, s = input). Therefore, while a sensor might be able to operate or detect below or above the linear range, because it is out of the linear working range which can be calibrated, these signals will be difficult to quantify accurately.Figure 2.Overview of the key approaches for characterizing sensor performance.