**3. Types of Porous silicon biosensors**

#### **3.1. Optical p-Si biosensors**

**Figure 3.** SEM images of gold nanostructures (a,b) used to fabricated a porous silicon monolayer. We show the sur‐ face (c) and the cross sectional (d) images. These structures were prepared at CIE-UNAM porous silicon laboratory.

The p-Si material can be prepared either in powder or wafer permitting to elaborate devices that can be dispersed in a given medium or reused [12]. Furthermore p-Si is a material that allows the fabrication of high quality photonic crystals [13] by applying the method descri‐ bed before to obtain multilayers structures. Such characteristics therefore allow several bio‐

**2. Porous silicon biosensors: construction and transduction principles**

of alkenes and alkynes [27, 28], radiation [29], and other chemical approaches [15, 16].

A proper pore-size distribution helps to achieve an efficient biosensor; p-Si fabricate from p+ and n+ -type silicon substrates is mesoporous, and suitable for immobilisation of biomacro‐

The aim of a biosensor is to produce either discrete or continuous signals, which are propor‐ tional to a single analyte or a related group of analytes [14]. Because of its particular proper‐ ties, the p-Si can be used as a transducer to convert this analytes into an optical or electrical signal [1]. Its large surface area enables an effective capture of the biological analytes al‐ though such a large surface area also implies high reactivity with the enviroment. This can cause the degradation of the biosensor and/or possible false positives. For this reason, stabili‐ zation of the p-Si surface via an appropriate surface chemistry is a required step for obtain‐ ing a succesful biosensor [15]. The surface chemistry should be designed in such a way as to obtain the desired effects, and yet still displaying bioactivity [16]. Also the binding affinity with the studied analytes must be taken into account [15, 17]. Some common techniques to function‐ alize p-Si include: oxidation [18, 19, 20], silanization [1, 15, 21, 22, 23, 24, 25], hydrosilylation

sensing approaches usign this porous material [1].

144 State of the Art in Biosensors - General Aspects

Chemical or biomolecule detection can be based on changes in the optical spectral interfer‐ ence pattern [22, 23]. When white light passes through the p-Si an interference pattern is ob‐ served, this effect is called a Fabry–Perot fringe pattern, the binding of molecules induces changes in this pattern which are relate to a change in the refractive index of the p-Si [22]. This change is shown by a shift of the fringe pattern that can be quantified [22]. The effect depends on the refractive index value of the analized solution but also on how it penetrates into the pores [11]. The simplest kind of such p-Si biosensors is made of mono and doublelayer films [1]. Some biological systems studied with these biosensors are: DNA hybridiza‐ tion [22, 32, 33], antibody cascading and the prototypical biotinstreptavidin interaction [22]. Using an analogous optical transduction modality it is possible to build p-Si biosensors with others complex optical structures as multilayer devices [6, 30].

These can be built up by alternating the applied current densities during the electrochemical etching generating a periodic or quasiperiodic combination of refractive indices [6]. This kind of structures offers better reflectance spectra (without side lobes) if compared with a mono or doubled-layered structure [30]. The etching parameters must be chosen to accom‐ modate the analyte of interest whilst maximising the optical response. Some of the p-Si opti‐ cal structures used in biosensing are: 1D photonic crystals [35, 36], rugate filters [37], microcavities [6, 38] and quasicrystals [39]. The use of these optical structures in biosensors allows integrability of all optical components and do not require electric contacts [11].

The photoluminescence properties of p-Si are also useful mechanisms for developing biosen‐ sors. It is possible to associate the amount of analytes studied with the changes in the photo‐ luminescence spectra [1, 40, 41]. For example the quenching in the photoluminescence spectra after DNA deposition was used to study the transduction of DNA hybridization [42]. In this case the behavior was attributed to non-radiative recombination processes [1]. In recent years a successful implementation of this type of biosensor was obtained [6, 32, 42, 41] however until now this kind of biosensor is less accurate than its interferometric counterparts [1].

In a similar way the amount of an analyte of interest can be quantified by measuring the flu‐ orescence signal intensity of a fluorescence molecule used as a marker fixed at a p-Si struc‐ ture before and after an analyte is located into the p-Si [43].

We offer two comprehensive case examples to illustrate how the optical p-Si biosensors work. The first example is a sensor of a fluorescent molecule: fluorescein-5-maleimide (FM), by using a 1D photonic crystal or Bragg mirror [44]. The basic mirror was made by alternat‐ ing layers of high (2.83) and low (1.65) refractive indexes, with a first layer that allows a good penetration of the active molecule into the porous structure. The surface of the first layer was functionalized by silanization with 3-mercaptopropyl)-trimethoxysilane (MPTS) to link the fluorescent molecule. The silicon mirror was fabricated in order to achieve a re‐ flectance spectrum in a range that overlaps the fluorescent excitation of the molecule. The samples were analyzed by fluorescent spectrometry. The emission signal from fluorescent molecules was enhanced because of the p-Si mirror. That is, the p-Si structure provided a platform for high-sensitivity measurements. This biosensor uses two different detection platforms by using reflectance measurements as we show in the figure 4 and by analyzing the fluorescent spectrum as it can be observed in the figure 5.

**Figure 4.** Reflectance spectra for freshly etched (thin line), silanized (normal line), and functionalized (thick line) mir‐ rors. Vertical lines correspond to wavelength of excitation of 491 nm and emission of 521 nm of the FM molecule in a phosphate solution. Uncertanties in the reflectance intensity and wavelength were of ±2% and ±1 nm, respectively [44].

**Figure 5.** Fluorescence emission of FM molecules deposited on the MPTS functionalized surfaces. Monolayers corre‐ spond to sample m45 and m70 (no mirrors). Sample M45 shows the best fluorescence signal. Fluorescence emissions of FM in solution are shown in the inset for comparison; the concentrations for each spectrum from left to right, are 0.37, 0.7, 1.2, 3.77, 5.39, 7.7, and 11 mM. Uncertanties in fluorescence intensity and wavelength were of ±0.1% and ±1 nm, respectively [44].

The second example is a microcavity [45]. This microcavity is formed when a luminescent p-Si layer is inserted between two Bragg reflectors made of p-Si. The broad luminescence band is altered and very narrow peaks are detected. The position of these peaks is extremely sen‐ sitive to a small change in refractive index, such as that obtained when a biological analyte is placed in the large internal surface of p-Si. A DNA biosensor was developed by using such an oxidized microcavity [45]. After successful silanization of the p-Si surface, DNA was im‐ mobilized into the porous surface through a careful diffusion. Finally, the DNA-attached wass exposed to its complementary strand of DNA (cDNA). A red-shift in photolumines‐ cence is observed. Full-length viral DNA molecules were also detected with the microcavity biosensor [45vis]. The advantages of optical sensing are significantly improved when this approach is used within an integrated optics context [46].

#### **3.2. Electric and Electrochemical P-Si biosensors**

after DNA deposition was used to study the transduction of DNA hybridization [42]. In this case the behavior was attributed to non-radiative recombination processes [1]. In recent years a successful implementation of this type of biosensor was obtained [6, 32, 42, 41] however until

In a similar way the amount of an analyte of interest can be quantified by measuring the flu‐ orescence signal intensity of a fluorescence molecule used as a marker fixed at a p-Si struc‐

We offer two comprehensive case examples to illustrate how the optical p-Si biosensors work. The first example is a sensor of a fluorescent molecule: fluorescein-5-maleimide (FM), by using a 1D photonic crystal or Bragg mirror [44]. The basic mirror was made by alternat‐ ing layers of high (2.83) and low (1.65) refractive indexes, with a first layer that allows a good penetration of the active molecule into the porous structure. The surface of the first layer was functionalized by silanization with 3-mercaptopropyl)-trimethoxysilane (MPTS) to link the fluorescent molecule. The silicon mirror was fabricated in order to achieve a re‐ flectance spectrum in a range that overlaps the fluorescent excitation of the molecule. The samples were analyzed by fluorescent spectrometry. The emission signal from fluorescent molecules was enhanced because of the p-Si mirror. That is, the p-Si structure provided a platform for high-sensitivity measurements. This biosensor uses two different detection platforms by using reflectance measurements as we show in the figure 4 and by analyzing

**Figure 4.** Reflectance spectra for freshly etched (thin line), silanized (normal line), and functionalized (thick line) mir‐ rors. Vertical lines correspond to wavelength of excitation of 491 nm and emission of 521 nm of the FM molecule in a phosphate solution. Uncertanties in the reflectance intensity and wavelength were of ±2% and ±1 nm, respectively [44].

now this kind of biosensor is less accurate than its interferometric counterparts [1].

ture before and after an analyte is located into the p-Si [43].

146 State of the Art in Biosensors - General Aspects

the fluorescent spectrum as it can be observed in the figure 5.

The highly sensitive surface of p-Si and the possibility to measure changes in its electrical properties added to its capacity to adsorb an enormous amount of different compounds, can be used for electrical biosensor applications [47, 48]. These approaches consider the use of electrical contacts on the p-Si layer made by metal deposition to measure the changes in the electrical properties such as capacitance and conductance when an analyte is attached to the p-Si layer [49]. An example of this type of biosensor is a macroporous sensor to detect DNA hybridization by characterizing the difference between the dipolar moment in p-Si layers with and without the analyte [47]. Another DNA detector of nanoporous silicon biosensor is described in reference [50]. This biosensor is an electrochemical device that transduces the hybridization of DNA into a chemical oxidation of guanine by Ru (bpy)2+ 3, the reduced form of which is then detected electrochemically.

Another effective platform to develop a p-Si biosensor is by applying electrochemical char‐ acterization. There are two main types of electrochemical transduction in biosensors: poten‐ tiometry and amperometry/voltammetry [12].

In potentiometryc biosensors the main parameter is the potential difference between the cathode and the anode in an electrochemical cell [51, 52]. This difference can be transduced as an electrical signal [12]. Amperometric and voltammetric biosensors consider the redox reaction that takes place in the anodization cell when an analyted of interest is placed. In this case the analyte is immobilised and an analyte oxidation/reduction process produces a flux of electrons measured, in terms of current intensity, cross the electrodes of the electro‐ chemical cell [12]. These biosensors are too sensitive to pH modifications [51].

Examples of these sensors are the potentiometric and amperometric urea sensor based on nanoporous silicon technology described by Joon-Hyung Jin et al [52]. One of the electro‐ chemical devices consists on three thin-film electrodes patterned on p-type silicon wafer by using platinum RF sputtering and silver (Ag) evaporation. The working electrode, on which the urease is inmobilized with a polymeric conductor: polypirrole (PPy) is sensitive to urea dissolved in artificially made electrolyte solution. The reference electrode is p-Si -based Ag/ AgCl thin-film reference electrode (TFRE). The other is a platinum (Pt) thin-film counter electrode. In a potentiometric urea sensor, urea concentration is related to the measured po‐ tential applied between the working and reference electrode according to the Nernst equa‐ tion. The other device is developed under amperometric regime. In this case the ureasecatalyzed hydrolytic reaction of urea causes current flow between the working and counter electrode and the amount of current flow is proportional to the urea concentration that rep‐ resents a change of pH, which is based on the Cottrell equation. In this study [50] it was found that urea sensitive electrodes (PSUE's) and Ag/AgCl TFRE's based on p-Si layers pro‐ vides better adhesive strength between thin-films, and silicon-based electrodes. This reduces the leaching out of TFRE components and enhances the sensitivity of a sensing electrode. The presence of carbon, nitrogen and sulfur, which were attributed to the urease-doped PPy films were confirmed by EDX characterization. The p-Si-based Ag/AgCl TFRE can be recom‐ mended as an ideal non-polarizable reference electrode to determine the electrochemical cell potentials and currents of sensing electrodes. Amperometry for monitoring the urea concen‐ trations caused by urease-catalyzed reactions is superior to a potentiometric method in that the amperometric urea sensors gives a longer linear range, higher sensitivity and shorter re‐ sponse times than the potentiometric urea sensors, especially at low urea concentrations.

Another very interesting application of porous silicon biosensors is for liver diagnosis [53]. Min-Jung Song et al presented a study of a biosensor array system consisting of cholesterol, bilirubin and glutamate sensors. The p-Si electrochemical system consisted of porous silicon layers formed on each working electrode that increased greatly the effective surface area. The electrodes in the sampling wells minimized a cross-interference effect to permit multiple sampling by immobilization of the enzymes using a silanization technique. The biosensor arrays tested used aqueous samples of the enzymes prepared in a 50 mM phosphate buffer solution (pH 8). All measurements were performed at room temperature at amperometric detection regime of each sensor was carried out using at a potential of +0.6 V vs. Ag/AgCl for the biosensors of the hydrogen peroxide generated in the silanized layer where the enzymat‐ ic reactions occur. In general, normal cholesterol concentrations in the human do not exceed 200 mg per 100 ml [53]. Higher cholesterol concentrations are considered abnormal.

In this case, the current detected is linearly proportional to cholesterol concentrations in the range of 1 mM to 50 mM; sensitivity was measured at approximately 0.2656 µA/mM. The bilirubin calibration curve covers a large concentration range between 0.002 mM and 0.020 mM, which includes normal levels (0.2 ~ 1.0 mg/dl), and levels typical of abnormal serum bilirubin. The sensitivity of the calibration curve approximated 0.15354 mA /mM. The detec‐ tion of the ratio of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) that in human serum indicates an abnormal symptom of the liver is also based upon electro‐ chemical oxidation at the Pt electrode surface. Since L-glutamate is a product of both ALT and AST reactions occurring in the buffer solution, the enzyme activities can be determined from the current changes at the L-glutamate sensor. On average, the serum ALT and AST levels measured in healthy people by optimized conventional ALT and AST assays approxi‐ mates 10 U/l at 25 ˚C and any increase in enzyme levels that exceed 100 U/l is taken to indi‐ cate liver disease. The sensitivity determined from the semi-logarithmic plot approximated 0.13698 µA/(U/l) for ALT over the range of 1.3 U/l to 250.0 U/l. For AST, the sensitivity was about 0.45439 µA/(U/l) in the same concentration ranges as for ALT.

This device offers several important advantages which include


hybridization of DNA into a chemical oxidation of guanine by Ru (bpy)2+ 3, the reduced

Another effective platform to develop a p-Si biosensor is by applying electrochemical char‐ acterization. There are two main types of electrochemical transduction in biosensors: poten‐

In potentiometryc biosensors the main parameter is the potential difference between the cathode and the anode in an electrochemical cell [51, 52]. This difference can be transduced as an electrical signal [12]. Amperometric and voltammetric biosensors consider the redox reaction that takes place in the anodization cell when an analyted of interest is placed. In this case the analyte is immobilised and an analyte oxidation/reduction process produces a flux of electrons measured, in terms of current intensity, cross the electrodes of the electro‐

Examples of these sensors are the potentiometric and amperometric urea sensor based on nanoporous silicon technology described by Joon-Hyung Jin et al [52]. One of the electro‐ chemical devices consists on three thin-film electrodes patterned on p-type silicon wafer by using platinum RF sputtering and silver (Ag) evaporation. The working electrode, on which the urease is inmobilized with a polymeric conductor: polypirrole (PPy) is sensitive to urea dissolved in artificially made electrolyte solution. The reference electrode is p-Si -based Ag/ AgCl thin-film reference electrode (TFRE). The other is a platinum (Pt) thin-film counter electrode. In a potentiometric urea sensor, urea concentration is related to the measured po‐ tential applied between the working and reference electrode according to the Nernst equa‐ tion. The other device is developed under amperometric regime. In this case the ureasecatalyzed hydrolytic reaction of urea causes current flow between the working and counter electrode and the amount of current flow is proportional to the urea concentration that rep‐ resents a change of pH, which is based on the Cottrell equation. In this study [50] it was found that urea sensitive electrodes (PSUE's) and Ag/AgCl TFRE's based on p-Si layers pro‐ vides better adhesive strength between thin-films, and silicon-based electrodes. This reduces the leaching out of TFRE components and enhances the sensitivity of a sensing electrode. The presence of carbon, nitrogen and sulfur, which were attributed to the urease-doped PPy films were confirmed by EDX characterization. The p-Si-based Ag/AgCl TFRE can be recom‐ mended as an ideal non-polarizable reference electrode to determine the electrochemical cell potentials and currents of sensing electrodes. Amperometry for monitoring the urea concen‐ trations caused by urease-catalyzed reactions is superior to a potentiometric method in that the amperometric urea sensors gives a longer linear range, higher sensitivity and shorter re‐ sponse times than the potentiometric urea sensors, especially at low urea concentrations.

Another very interesting application of porous silicon biosensors is for liver diagnosis [53]. Min-Jung Song et al presented a study of a biosensor array system consisting of cholesterol, bilirubin and glutamate sensors. The p-Si electrochemical system consisted of porous silicon layers formed on each working electrode that increased greatly the effective surface area. The electrodes in the sampling wells minimized a cross-interference effect to permit multiple sampling by immobilization of the enzymes using a silanization technique. The biosensor arrays tested used aqueous samples of the enzymes prepared in a 50 mM phosphate buffer

chemical cell [12]. These biosensors are too sensitive to pH modifications [51].

form of which is then detected electrochemically.

148 State of the Art in Biosensors - General Aspects

tiometry and amperometry/voltammetry [12].


In the following paragraph we will describe in detail an example of an electrochemical sensor [48].

Porous silicon samples were prepared from p+-type, boron doped silicon wafers with a re‐ sistivity of 0.008-0.012 Ohm cm by standard anodization (electrolyte: 15% of HF) at a current density of 30 mA.cm-2. The porosity measured by the gravimetrical method was approxi‐ mately 62%. The pore size was estimated by TEM and ranged from 50 nm-75 nm in diame‐ ter. These diameters are large enough to allow the sensing molecules to penetrate and attach. For DNA, the diameter of the nucleotides is approximately 5Å, which is small enough to fit into the porous matrix. Stabilization of p-Si is necessaryto passivate its surface and this was done by thermal oxidation. Thermal oxidation of p-Si requires several precau‐ tions and high temperatures (>700ºC). Covering the whole internal surface with a thin SiO2 layer stabilizes the structure, permits water penetration into the pores and facilitates probe and target penetration [54]. Al l p-Si samples were thermally oxidized in oxygen ambient at 900ºC for 10 minutes.

The electrochemical instrumentation used for these experiments included a BAS 100B/W Electrochemical Analyzer and a BAS VC-2 voltammetry cell (model MF-1065). It is well suit‐ ed for small sizes and has a special micro-cell for volumes as small as 50µL. The micro-cell, which included the working electrode, separated a small volume containing the sample from a bulk solution containing the reference and auxiliary electrode with a salt bridge. A platinum wire served as auxiliary electrode and the modified p-Si samples function as working electrodes. It is important to mention that p-Si, especially oxidized p-Si, is not con‐ ducting and it is in fact the p+ doped silicon that was conducting the electrical current. The top area of the exposed PSi samples was 0.8 cm2 and all lateral areas were insulated with a commercial epoxy resin (see figure 6). The epoxy resin was deposited very carefully and dried for one hour. The samples were attached to the electrochemical system as shown in fig. 6. Potentials were measured relative to an aqueous, saturated Ag/AgCl double junction (reference electrode). The voltammetry experiments were carried out at different scan rates in an electrochemical buffer solution composed by 50 mM sodium phosphate (pH 7) with 0.7 M NaCl. A schematic representation of the electrochemical measurement set up and the electrode arrangement is shown in fig. 6.

**Figure 6.** Measurement system: the p-Si electrode is used as working electrode. A platinum wire is the auxiliary elec‐ trode and Ag/AgCl the reference electrode. Inset: cross section showing the different parts of the working electrode [48].

Three different synthetic oligonucleotides were obtained from MWG Biotech, INC, and have the following sequences: (probe): 5'-TAI-CTA-TII-AAT-TCC-TCI-TAI-ICA-3',(target):5'- GCCTAC-GAG-GAA-TTC-CAT-AGC-T-3' and (two-base mismatch target):5'-GCC-TAC-GAG-GAA-TTG-GAT-AGC-T-3. Tris(2,2'-bypyridyl) ruthenium (II) chloride hexahydrate was purchase from Strem Chemicals. All other chemicals were of analytical grade and pur‐ chased from Aldrich and Fluka. Deionized distilled water was obtained from Millipore.

The detection of DNA consists of the following steps: p-Si silanization, probe immobiliza‐ tion, hybridization, and voltammetric detection.

The electrochemical instrumentation used for these experiments included a BAS 100B/W Electrochemical Analyzer and a BAS VC-2 voltammetry cell (model MF-1065). It is well suit‐ ed for small sizes and has a special micro-cell for volumes as small as 50µL. The micro-cell, which included the working electrode, separated a small volume containing the sample from a bulk solution containing the reference and auxiliary electrode with a salt bridge. A platinum wire served as auxiliary electrode and the modified p-Si samples function as working electrodes. It is important to mention that p-Si, especially oxidized p-Si, is not con‐ ducting and it is in fact the p+ doped silicon that was conducting the electrical current. The

commercial epoxy resin (see figure 6). The epoxy resin was deposited very carefully and dried for one hour. The samples were attached to the electrochemical system as shown in fig. 6. Potentials were measured relative to an aqueous, saturated Ag/AgCl double junction (reference electrode). The voltammetry experiments were carried out at different scan rates in an electrochemical buffer solution composed by 50 mM sodium phosphate (pH 7) with 0.7 M NaCl. A schematic representation of the electrochemical measurement set up and the

**Figure 6.** Measurement system: the p-Si electrode is used as working electrode. A platinum wire is the auxiliary elec‐ trode and Ag/AgCl the reference electrode. Inset: cross section showing the different parts of the working electrode [48].

Three different synthetic oligonucleotides were obtained from MWG Biotech, INC, and have the following sequences: (probe): 5'-TAI-CTA-TII-AAT-TCC-TCI-TAI-ICA-3',(target):5'- GCCTAC-GAG-GAA-TTC-CAT-AGC-T-3' and (two-base mismatch target):5'-GCC-TAC-GAG-GAA-TTG-GAT-AGC-T-3. Tris(2,2'-bypyridyl) ruthenium (II) chloride hexahydrate was purchase from Strem Chemicals. All other chemicals were of analytical grade and pur‐ chased from Aldrich and Fluka. Deionized distilled water was obtained from Millipore.

and all lateral areas were insulated with a

top area of the exposed PSi samples was 0.8 cm2

150 State of the Art in Biosensors - General Aspects

electrode arrangement is shown in fig. 6.

Several methods may be employed to bind DNA to different supports [55]. One method commonly used for binding DNA involves silanization of an oxidized surface. The function of silane coupling agents is to provide stable bond between two non-bonding surfaces: for example, an inorganic surface to an organic molecule. 3-glycidoxypropyltrimethoxysilane was used to silanize the oxidized p-Si A 5% aqueous solution of silane was prepared (pH 4.0). This converts silane into a reactive silanol through hydrolysis. The p-Si samples were then immersed into the continuosly stirred solution and left overnight. 3-glycidoxypropyl‐ trimethoxysilane is hydrolyzed to a reactive silanol by using double distilled water (pH 4). p-Si samples were then submerged into silanol solution for approximately 17 hours. Con‐ stant stirring of the solution was necessary to continuously mix the solution.

After successful silanization, DNA was immobilized onto the surface of p-Si through diffu‐ sion. Aqueous solutions of DNA containing 150 µl of DNA (50 µM) were carefully placed directly above de p-Si layer. The DNA molecules covalently bond to the silanized surface, where they become immobilized. The samples were then placed in a steam container where they were heated in an oven at 37ºC for approximately 20 hours. The DNA attached samples were then rinsed in double distilled water and dried with nitrogen.

The DNA attached to p-Si was exposed to its complementary strand DNA (target), the mis‐ match sequence (mismatch probe) and itself (probe). Binding was allowed to proceed for 1 hour at room temperature into hybridization buffer containing 1 M NaCl, 10-20 mM sodium cacodylate, 0.5 mM EDTA, 150 mM KCl and 5 mM MgCl2. Throughout the steps, binding was confirmed using Fourier Transform Infrared Spectroscopy (results not shown here).

Cyclic voltammetry (CV) was carried out having the DNA modified, p-Si electrode as work‐ ing electrode, an Ag/AgCl as the reference electrode, and platinum wire as the counter elec‐ trode. 6 µl of *Ru(bpy) 2+ <sup>3</sup>* (0.1µM) was poured into 150 µl of electrochemical buffer solution. After allowing the solution to diffuse into the samples for 15 minutes, CV was performed. Solutions were deoxygenated via purging with nitrogen for 10 minutes prior to measurements.

p-Si DNA-electrodes and *Ru(bpy) 2+ <sup>3</sup>* were used for specific gene detection. *Ru(bpy) 2+ <sup>3</sup>* ex‐ hibits a reversible redox couple at 1.05 V and oxidizes guanine in DNA at high salt concen‐ tration [56] according to:

$$\frac{1}{2}\operatorname{Ru}(bpy)\_3^{2+} \rightharpoonup \operatorname{Ru}(bpy)\_3^{3+} + e^- \tag{1}$$

$$\text{Ru(bpy)}\_{3}^{3+} + \text{DNA} \rightarrow \text{DNA}\_{vx} + \text{Ru(bpy)}\_{3}^{2+} \tag{2}$$

where DNAox is a DNA molecule in which guanine has been oxidized by *Ru(bpy) 3+ <sup>3</sup>*. If the DNA probe contains guanine then *Ru(bpy) 3+ <sup>3</sup>* will oxidize guanine in DNA, even without the presence of the target DNA. In order to prevent that the DNA probe reacts with *Ru(bpy) 2+ <sup>3</sup>* the guanine has been replaced for another less reactive nucleotide. Some previous results show that the addition of an oligonucleotide that does not contain guanine produces a small enhancement in the oxidation current [56]. Those results have shown that the inosine 5' monophosphate is 3 orders of magnitude less electrochemically reactive than guanosine 5' monophosphate and still recognizes cytidine [56]. This fact is very important to recognize all four bases in the target sequence. Nevertheless there is a drawback that can have consequen‐ ces on the hybridization efficiency. Since the deaminated hypoxanthine in the ionosine can only form two of the three hydrogen bonds in a Watson-Crick base pair, it may be desirable to use a guanine derivative that is redox-inert but capable of forming all three hydrogen bonds. However some studies have shown [56] that the specificity afforded by inosine sub‐ stitution was sufficient but they propose 7-deazaguanine as alternative. For this reason the DNA probe sequence does not contain the guanine base but the target does. Figure 7 (top) shows the CV obtained in solution for the hybrid DNA (probe-target) at different scan rates (target DNA concentration of 0.5 x 10-10M). Figure 7 (bottom) shows that the anodic current of *Ru(bpy) 2+ <sup>3</sup>* is linearly proportional to the scan rate.

**Figure 7.** (top) Cyclic voltammograms of the probe-target DNA sequence in 0.1 μM *Ru(bpy) 2+ <sup>3</sup>* solution at different scan rates (mV.s-1): (1) 20; (2) 50; (3) 80; (4) 100; (5) 200. (bottom) The anodic current changed with the scan rate. The target concentration was 0.5x10-10M [48].

This result is congruent with a process that is controlled by adsorption. Figure 8(top) shows the CV (scan rate of 50 mV.s-1) of varied concentrations of target DNA (probe-target se‐ quence, curves 2 to 5) and different targets (probe-mismatch target sequence, curve 1 and probe-probe sequence, curve 6). In curve 1, the mismatch target sequence contains two more pairs of base G than the target sequence and that is why the current in this case is bigger than the current obtained in the probe-target sequence cases (curves 2 to 5) or the probeprobe sequence (curve 6). Moreover in curve 6 the current intensity decreases as a conse‐ quence of the absence of the

show that the addition of an oligonucleotide that does not contain guanine produces a small enhancement in the oxidation current [56]. Those results have shown that the inosine 5' monophosphate is 3 orders of magnitude less electrochemically reactive than guanosine 5' monophosphate and still recognizes cytidine [56]. This fact is very important to recognize all four bases in the target sequence. Nevertheless there is a drawback that can have consequen‐ ces on the hybridization efficiency. Since the deaminated hypoxanthine in the ionosine can only form two of the three hydrogen bonds in a Watson-Crick base pair, it may be desirable to use a guanine derivative that is redox-inert but capable of forming all three hydrogen bonds. However some studies have shown [56] that the specificity afforded by inosine sub‐ stitution was sufficient but they propose 7-deazaguanine as alternative. For this reason the DNA probe sequence does not contain the guanine base but the target does. Figure 7 (top) shows the CV obtained in solution for the hybrid DNA (probe-target) at different scan rates (target DNA concentration of 0.5 x 10-10M). Figure 7 (bottom) shows that the anodic current

**Figure 7.** (top) Cyclic voltammograms of the probe-target DNA sequence in 0.1 μM *Ru(bpy) 2+ <sup>3</sup>* solution at different scan rates (mV.s-1): (1) 20; (2) 50; (3) 80; (4) 100; (5) 200. (bottom) The anodic current changed with the scan rate. The

This result is congruent with a process that is controlled by adsorption. Figure 8(top) shows the CV (scan rate of 50 mV.s-1) of varied concentrations of target DNA (probe-target se‐ quence, curves 2 to 5) and different targets (probe-mismatch target sequence, curve 1 and

of *Ru(bpy) 2+ <sup>3</sup>* is linearly proportional to the scan rate.

152 State of the Art in Biosensors - General Aspects

target concentration was 0.5x10-10M [48].

**Figure 8.** top) Cycled voltammograms of different concentrations of target DNA: (2) 0.5 x 10-10M, (3) 100 x 10-10M, (4) 200 x 10-10M, (5) 500 x 10-10M. Curve 1 shows the CV for the probe-mismatch target DNA sequence (0.5 x 10-10M) and curve 6 for the probe-probe DNA sequence (0.5 x 10-10M). (bottom) The anodic current changed with the concentra‐ tion of the target DNA. In all cases, 0.1 μM of *Ru(bpy) 2+ <sup>3</sup>* was used [48].

base G in the probe. Nevertheless a significant increase in current was observed for curves 2 to 5 where the target DNA undergoes hybridization to the complementary DNA. This cur‐ rent increase suggests that the hybridization was successful and that the electron transfer from the guanines of the hybridized strand to *Ru(bpy) 2+ <sup>3</sup>* is responsible for the increase in the current. In comparing curves 1, 2 and 6 (same DNA concentration but different target sequence) it is observed that the sensor responds differently to each target and therefore a good selectivity is achieved. Figure 8 (bottom) shows the anodic peak currents of *Ru(bpy) 2+ <sup>3</sup>* at four different concentrations (curves 2 to 5). The peaks are linearly related to the concen‐ tration of the target DNA sequence between 0.5 x 10-10 and 500 x 10-10M. The detection limit of this approach was 5 x 10-11M. The sensitivity achieved in this work is similar to the one obtained in reference [57], where a sensitivity of 9.0 x 10-11M was reported for a sensor that uses gold substrates instead. In summary those results clearly show that the p-Si sensor shown here has a good selectivity and sensitivity to the target compound, two very impor‐ tant characteristics that a sensor ought to have.

An overview of the requirements for a good performance of a p-Si biosensor was presented and the generalities of the fabrication of different kinds of these biological sensors as well. In the next section we will discuss the new materials, uses and future of porous silicon.

## **4. Future of porous silicon biosensors**

Sensors allow our systems and devices to be in relation with the real events that we need to register or control. So, precision (same response to the same stimuli: repeatability) and accu‐ racy (indicating magnitude value as close as possible to the real magnitude of the stimulus to be sensed: minimum absolute error spread) are two main requirements for any sensor when the industry selects a structure type for market use. However, other properties will define the success of a new kind of sensor in the market. These are: technological compati‐ bility with the existing devices, geometric dimension requirements, low noise insertion, ease of adjustment and setup, low power consumption, performance standardization (linear if possible), low thermal or aging characteristic drifts, robustness, reliability, low obsolescence, and very wide field of applications. p-Si is a material that accomplishes all of these require‐ ments with enough margins to think that it will become increasingly popular in the short term. For instance, integrated circuits (IC) are made of crystalline Silicon, which means it is fully compatible for associating a p-Si sensor to any electronic device. The electrochemical technology used to create a p-Si layer does not collide with the IC lithography. The geomet‐ ric dimensions required to create this type of sensor are sufficiently small to be integrated in an IC. The homogeneity of the porous and its radius control (internal surface density con‐ trol) as well as its layer stability is improving very fast.

An important factor to take into account in the implementation of a p-Si biosensor is its chemical stability after sample storage. Due to its high superficial area, p-Si based materials tends to be oxidized when are exposed to air ambient conditions. This oxidation plus the ad‐ dition of other molecules present in the air ambient could modify the biosensor reponse to an analyte after certain amount of time. Pasivation techniques and surface functionalization described before have been proved being successful to prevent or minimize these stability issues. Work has to be done in order to improve the existing methods and assure the repro‐ ducibility of p-Si biosensors reponse over a lapse time of years.

Porous silicon has proven to be a succesful material for biosensing applications [1, 58]. In some cases even femtomolar concentrations in biomolecules was demostrated [1]. The wide range of applications in sensing biological substances include: healt applications [7, 43, 58], virus detection [43], inmunosensors [59, 1], DNA biosensors [58,60], drug delivery [16], bio‐ security on food [61], biological warfare agents [62], implantable biosensor technology [58], among others [12, 15, 63]. Some new trends in the fabrication of p-Si biological sensor devi‐ ces are the use of nanomaterials [54]. The sensitivity and performance of biosensors are be‐ ing improved by using nanomaterials for their construction allowing simple and rapid in vivo analyses [61]. Recently, the plasmonic properties of metal nanoparticles have been used to develop a p-Si biosensor that present Ramman enhacement [64]. Another attractive meth‐ od for monitoring biomolecular interactions in a highly parallel fashion is the use of micro‐ arrays. This p-Si novel porous chip was demonstrated as stable and reproducible, and the fluorescent bioassay reproducibility has been shown [65].

Lately integrated systems of p-Si has been developed [66, 67] e.g a guided mode biosensor based on grating coupled p-Si waveguide [68].

Another type of biosensensing approach that is appearing is an acoustic wave transducer that is coupled with a bioelement e.g an antibody. When the analyte molecules (antigen) get attached to a membrane, the membrane mass changes, resulting in a modification in the res‐ onant frequency of the transducer that can be measured [66]. Rencently p-Si has been pro‐ posed as a good material for this kind of biosensors [64].

Notwithsranding the many advantages of p-Si mentioned during this work, several chal‐ lenges will need to be overcome to be able to make biosensors a viable commercial product. They fall into two main areas: those concerning the fabrication of p-Si for cost effective and robust devices, and those addressing the ability to handle real-world sample matrices such as whole blood [1]. Both are presently the focus of intensive research and it is reasonable to believe that new and exciting developments will occur in a very near future.
