US20100159461A1

US20100159461A1 - Ion sensitive field effect transistors

Info

Publication number: US20100159461A1
Application number: US12/591,847
Authority: US
Inventors: Christofer Toumazou; Bhusana Premanode; Leila Shepherd
Original assignee: DNAE Group Holdings Ltd
Current assignee: DNAE Group Holdings Ltd
Priority date: 2004-07-13
Filing date: 2009-12-02
Publication date: 2010-06-24
Also published as: JP4801064B2; CN101949883A; CN101292156A; JP2008506943A; CN101292156B; ES2674618T3; US20080265985A1; JP5509118B2; GB0415633D0; DE602005026528D1; EP2295957A1; EP1774306B1; ES2361324T3; GB2416210A; CN101949883B; GB2416210A8; EP1774306A1; JP2011099877A; WO2006005967A1; ATE499602T1

Abstract

According to a first aspect of the present invention there is provided a digital signal processing circuit, one or more switches of the circuit being provided by an ion sensitive field effect transistor.

Description

This application is a continuation-in-part of U.S. patent application Ser. No. 11/572,050 filed Jan. 12, 2007, entitled “Signal Processing Circuit Comprising Ion Sensitive Field Effect Transistor and Method of Monitoring a Property of a Fluid”, the entire contents of which are hereby incorporated herein by reference.

FIELD OF THE INVENTION

The present invention relates to ion sensitive field effect transistors and to processing and control systems utilising ion sensitive field effect transistors.

BACKGROUND TO THE INVENTION

The ion-sensitive field effect transistor (ISFET) is based on a MOSFET with a remote gate (or “reference electrode”) lying beneath a chemically-sensitive insulator. The surface of the insulator is exposed to an electrolyte upon which measurements are to be made. A typical ISFET use scenario is illustrated in FIG. 1. Field effects from ionic charge interaction at the insulator surface cause shifts in the ISFET drain current vs gate to source voltage (I_D-V_GS) characteristic as illustrated in FIG. 2. The insulator in contact with the electrolyte is chosen according to its chemical properties and sensitivity to a particular ion. For example, the insulator surface may be the gate oxide in ISFETs fabricated in a dedicated process which removes the polysilicon gate contact, or can be the Si₃N₄passivation layer inherent in the mass-producible CMOS process, where the polysilicon gate is left electrically floating and coupled to the sensing area via a metal stack. The rest of the transistor is encapsulated to isolate the bond wires from the electrolyte.
For ISFETs designed to measure the pH of an electrolyte, i.e. the H+ ion content of the electrolyte, silicon nitride and aluminium oxide membranes are commonly used to insulate the gate. ISFETs can be made sensitive to ions other than H⁺ through the choice of the ion-sensitive membrane, thus adding an element of ion-selectivity. ISFETs whose membrane is modified to be selective to a particular ionic species are known as ChemFETs, with a further variation, known as EnFETs, using enzymes in close proximity to the membrane surface. It has also been shown that even conventional pH-ISFETs with unmodified Si₃N₄membranes exhibit a limited but measurable sensitivity to K⁺ and Na⁺ ions. This said, practical and commercial applications of the ISFET for applications other than pH sensing are rare. Nonetheless, in the following discussion, the term ISFET is used both specifically to refer to a pH sensor and generally to refer to all ion and enzyme sensitive FETs operating on similar principles.
The attractiveness of ISFETs and their FET-based counterparts is that they are compatible with the standard fabrication processes used to mass produce computer chips and can therefore be reliably and cost-effectively produced. Importantly, processing circuitry can be integrated onto the same chip as the ISFET device itself. The integration of intelligent circuitry with the sensing device itself is what is required for the development of so-called “smart sensors” which require robustness to non-ideal sensing conditions, as well as to provide electronics to discriminate between chemicals “on-chip”.
The normal operating mode of an ISFET is the strong inversion region of the I_D-V_GScharacteristic. In this region, the gate to source voltage exceeds the threshold voltage V_TH, resulting in a strong inversion of the channel underlying the gate. For this mode of operation, drain current is related to the gate voltage by a square law or linear relationship.
Referring again to FIG. 1, any voltage applied to the reference electrode of an ISFET is capacitively-coupled via the electrolyte to the insulator surface, where a pH-dependent charge from ions on this interface modulates the channel current, causing the observed shifts in the ISFET transfer characteristic, thereby modulating the ISFET threshold voltage V_th:
V _th(ISFET) =V _th(MOS) +V _chem (1)
where V_th(MOS)is the threshold voltage of the basic MOSFET and V_chemis a pH-dependent chemical potential between the insulator surface and the reference electrode. A linear simplification of the chemically-dependent parameter V_chemis given by:
V _chem=γ+2.3αU _TpH (2)
where γ is a pH-independent grouping of chemical potentials, α is a dimensionless sensitivity parameter and U_Tis the thermal voltage. As with a standard FET, an ISFET can operate in either strong or weak inversion.
Assuming that the ISFET is operated in the constant drain current mode, with a constant drain-source voltage, the gate to source voltage directly reflects the pH-sensitive interfacial potential at the gate interface, that is:
pH=pH_cal +V _gs /S, (3)
where pH_calis the pH of a calibration liquid at 37° C. and S is the pH sensitivity of the ISFET. The derivation of this relationship is detailed further in “ISFET, Theory and Practice”, P. Bergveld, IEEE Sensor Conference, Toronto, October 2003. However, this approach assumes a constant temperature, and in any practical approach temperature compensation must be applied.
The conventional approach to compensating measurements for temperature effects is to model the temperature dependence of a system, measure the temperature in parallel with the pH, and correct the measured pH on the basis of the model and the measured temperature. Whilst effective, this approach has a number of disadvantages. Firstly, it relies upon the provision of a temperature sensor, typically comprising a temperature sensitive resistor integrated onto the same chip as the ISFET. Secondly, processing power must be provided to perform the correction. Thirdly, the process of correcting the measured pH values takes time. In a typical system, pH and temperature values are converted into their digital equivalents, prior to carrying out the further processing with a microprocessor or CPU. If necessary, digital control outputs are converted into analogue equivalents prior to application to a device to be controlled.
It has long been recognised that a key area in which ISFETs can be applied is that of implantable and wearable sensors. The requirements of conventional ISFET design outlined in the previous paragraph do not sit well with such sensors which require to be small, to consume low levels of power, and to be extremely accurate. Especially where the sensors form part of a control loop, e.g. controlling a drug delivery system, they must also be extremely accurate.

SUMMARY OF THE INVENTION

According to a first aspect of the present invention there is provided a digital signal processing circuit, one or more switches of the circuit being provided by an ion sensitive field effect transistor.
The digital signal processing circuit may further comprise biasing means for biasing the or each ion sensitive field effect transistor in the weak inversion region.
The or each ion sensitive field effect transistor may comprise an analyte sensitive membrane which is exposed in use to a medium to be monitored.
The digital signal processing circuit may be arranged to implement one or more of the following functions: AND, NAND, OR, XOR, NOR. The digital signal processing circuit may use CMOS logic.
The circuit may be configured to operate as a comparator for comparing a value of a parameter measured by the ion sensitive field effect transistor with a threshold value, the circuit comprising an ion sensitive field effect transistor and a metal oxide semiconductor transistor arranged in an inverter configuration.
One of the ion sensitive field effect transistor and the metal oxide semiconductor transistor may be an n-channel device and the other a p-channel device.
According to a second aspect of the present invention there is provided a device for detecting a plurality of chemical reactions and evaluating a logical function having the result of each of the plurality of chemical reactions as its inputs. The device comprises:

- at least one reaction chamber for each of the plurality of chemical reactions; and
- each reaction chamber being provided with an ion sensitive field effect transistor configured to operate as a switch for detecting one of the plurality of chemical reactions, the ion sensitive field effect transistors being coupled together to form a digital signal processing circuit for evaluating the logical function.

According to a third aspect of the present invention there is provided a method of evaluating a logical function having as one of its inputs the value of a parameter of a medium. The method comprises:

- configuring an ion sensitive field effect transistor to operate as a switch of a logic circuit; and
- exposing the ion sensitive field effect transistor to said medium.

The method may further comprise configuring the ion sensitive field effect transistor to switch when a change in the concentration of hydrogen ions in the medium indicates incorporation of a nucleotide into a sample of genetic material.
According to a fourth aspect of the present invention there is provided a method of detecting a plurality of chemical reactions and evaluating a logical function having the result of each of the plurality of chemical reactions as its inputs. The method comprises:

- for each of the plurality of chemical reactions, configuring at least one ion sensitive field effect transistor to operate as a switch of a logic circuit, and exposing the ion sensitive field effect transistor to the chemical reaction.

According to a fifth aspect of the present invention there is provided a method of detecting the presence of a unique genetic sequence in a target sample of single-stranded genetic material. The method comprises:

- performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to complement the unique genetic sequence;
- configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;
- exposing the ion sensitive field effect transistor to a solution comprising a product of the hybridisation reaction and an excess of all nucleotides; and
- monitoring the output of the ion sensitive field effect transistor to determine whether switching occurs and to thereby determine whether the probe has hybridised with the target sample.

Given that the probe is configured to complement the unique genetic sequence (i.e. is made up of a sequence of complementary nucleotides), the probe will only hybridise with a target sample in which the unique genetic sequence is present. Furthermore, when the product of the hybridisation reaction is exposed to an excess of all four nucleotides, chain extension will only occur if the probe and the target have hybridised. As such, switching of the ion sensitive field effect transistor due to chain extension is an indication that the unique genetic sequence, identified by the probe, is present in the target sample.
According to a sixth aspect of the present invention there is provided a method of detecting the presence of a plurality of unique genetic sequences in a target sample of single-stranded genetic material. The method comprises:

- for each of the plurality of unique genetic sequences:
  - performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to complement the unique genetic sequence;
  - configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;
  - exposing the ion sensitive field effect transistor to a solution comprising a product of the hybridisation reaction and an excess of all nucleotides; and
- configuring the ion sensitive field effect transistors into a digital processing circuit to provide an output of a logical function as an indication of the presence or absence of each of the plurality of unique genetic sequences.

According to a seventh aspect of the present invention there is provided a method of identifying a nucleotide present at a specific location in a target sample of single-stranded genetic material. The method comprises:

- performing a hybridisation reaction involving said target sample and a nucleotide probe in order to form a probe-target hybrid, the nucleotide probe being configured to uniquely identify the specific location;
- for each candidate nucleotide that may be present at the specific location:
  - configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;
  - exposing the ion sensitive field effect transistor to a solution comprising the probe-target hybrid and a nucleotide complementary to said candidate nucleotide; and
- monitoring the output of the ion sensitive field effect transistors to determine which switches and to thereby determine which of said complementary nucleotides has been incorporated at the specific location.

When the probe-target hybrid is exposed to a nucleotide, chain extension will only occur if that nucleotide complements the nucleotide at the specific location identified by the probe. Therefore, switching of the ion sensitive field effect transistor due to chain extension is an indication that the nucleotide present in the solution complements the nucleotide present at the location identified by the probe. Furthermore, identification of the complementary nucleotide implicitly identifies which of the candidate nucleotides is present at the specific location.
According to an eighth aspect of the present invention there is provided a method of identifying a nucleotide present at a specific location in a target sample of single-stranded genetic material. The method comprises:

- for each candidate nucleotide that may be present at the specific location:
  - performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to uniquely locate a nucleotide complementary to the candidate nucleotide at the specific location;
  - configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;
  - exposing the ion sensitive field effect transistor to a solution comprising the products of the hybridisation reaction and an excess of all nucleotides; and
- monitoring the outputs of the ion sensitive field effect transistors to determine which switches and to thereby determine which probe has hybridised with the target sample.

Each of the probes is configured to substantially complement the genetic sequence present in the sample. However, each probe will include a different nucleotide at the specific location. As such, only the probe including the nucleotide that is complementary to the nucleotide present at the specific location will hybridise with a target sample. Furthermore, when the product of the hybridisation reaction is exposed to an excess of all four nucleotides, chain extension will only occur if the probe and the target have hybridised. As such, determining which of the ion sensitive field effect transistors switches due to chain extension, determines which of the probes have hybridised with the target, and therefore which nucleotide complements that nucleotide present at the specific location. Identification of the complementary nucleotide implicitly identifies which of the candidate nucleotides is present at the specific location.
According to a ninth aspect of the present invention there is provided a method of detecting the presence of a single nucleotide polymorphism in a target sample of single-stranded genetic material. The method comprises:

- performing a hybridisation reaction involving said target sample and a nucleotide probe in order to form a probe-target hybrid, the nucleotide probe being configured to uniquely identify the expected location of the single nucleotide polymorphism;
- configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of a nucleotide into genetic material;
- exposing the ion sensitive field effect transistor to a solution comprising the probe-target hybrid and a complementary nucleotide associated with the single nucleotide polymorphism; and
- monitoring the output of the ion sensitive field effect transistor to determine whether switching occurs and to thereby determine whether the complementary nucleotide has been incorporated at the specific location.

When the probe-target hybrid is exposed to a nucleotide, chain extension will only occur if that nucleotide complements the nucleotide at the specific location identified by the probe. Therefore, switching of the ion sensitive field effect transistor due to chain extension is an indication that the nucleotide present in the solution complements the nucleotide present at the location identified by the probe. As such, the ion sensitive field effect transistor will only switch if the single nucleotide polymorphism is present at the expected location identified by the probe.
According to a tenth aspect of the present invention there is provided a method of detecting the presence of a single nucleotide polymorphism in a target sample of single-stranded genetic material. The method comprises:

- performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to uniquely locate a complementary nucleotide associated with the single nucleotide polymorphism at the expected location of the single nucleotide polymorphism;
- configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;
- exposing the ion sensitive field effect transistor to a solution comprising the products of the hybridisation reaction and an excess of all nucleotides; and
- monitoring the outputs of the ion sensitive field effect transistor to determine whether switching occurs and to thereby determine if the probe has hybridised with the target sample.

The probe is configured to substantially complement the genetic sequence present in the sample, but to also include a complementary nucleotide associated with the single nucleotide polymorphism at the expected location of the single nucleotide polymorphism. As such, the target sample will only hybridise with the probe if the target sample includes the single nucleotide polymorphism. Furthermore, when the product of the hybridisation reaction is exposed to an excess of all four nucleotides, chain extension will only occur if the probe and the target have hybridised. As such, the ion sensitive field effect transistor will only switch if the target sample includes the single nucleotide polymorphism.
According to an eleventh aspect of the present invention there is provided a method of detecting the presence of a plurality of single nucleotide polymorphisms in a target sample of single-stranded genetic material. The method comprises:

- for each of the single nucleotide polymorphisms:
  - performing a hybridisation reaction involving said target sample and a nucleotide probe in order to form a probe-target hybrid, the nucleotide probe being configured to uniquely identify the expected location of the single nucleotide polymorphism;
  - configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of a nucleotide into genetic material;
  - exposing the ion sensitive field effect transistor to a solution comprising the probe-target hybrid and a complementary nucleotide associated with the single nucleotide polymorphism; and
- configuring the ion sensitive field effect transistors into a digital processing circuit to provide an output of a logical function as an indication of the presence or absence of each of the plurality of single nucleotide polymorphisms.

According to a twelfth aspect of the present invention there is provided method of detecting the presence of a plurality of single nucleotide polymorphisms in a target sample of single-stranded genetic material. The method comprises:

- for each of the single nucleotide polymorphisms:
  - performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to uniquely locate a complementary nucleotide associated with the single nucleotide polymorphism at the expected location of the single nucleotide polymorphism;
  - configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of a plurality of nucleotides into genetic material;
  - exposing the ion sensitive field effect transistor to a solution comprising the hybridisation reactants and an excess of all nucleotides; and
- configuring the ion sensitive field effect transistors into a digital processing circuit to provide an output of a logical function as an indication of the presence or absence of each of the plurality of single nucleotide polymorphisms.

The hybridisation reaction of any of the above methods may occur during a nucleic acid amplification cycle. Furthermore, the hybridisation reaction may be part of a thermocycling amplification cycle, the cycle comprising denaturing, hybridisation and extension.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates schematically an ISFET in a typical use scenario;

FIG. 2 shows the I_D-V_GScharacteristics of the ISFET of FIG. 1 at various pH levels;

FIG. 3 illustrates schematically a p-channel silicon MOSFET;

FIG. 4 illustrates a current mirror comprising an ISFET;

FIG. 5 illustrates an HCell for measuring hydrogen ion concentration;

FIG. 6 shows the IV characteristic of the Hcell of FIG. 5;

FIG. 7 illustrates schematically a CMOS inverter;

FIG. 8 shows the switching characteristic of the inverter of FIG. 7;

FIG. 9 illustrates the shift in switching threshold caused by changing pH in a CMOS inverter comprising an ISFET;

FIG. 10 shows the switching properties of an ISFET inverter at differing input voltages;

FIG. 11 shows an ISFET based NAND gate and the corresponding truth table;

FIG. 12 shows an ISFET based NOR gate and the corresponding truth table;

FIG. 13 illustrates changes in output voltage of a standard ISFET source-drain follower readout circuit,

FIG. 14 illustrates schematically a CMOS inverter comprising an ISFET

FIG. 15 a illustrates pH response due to nucleotide incorporation during primer extension;

FIG. 15 b illustrates the simulated output of the CMOS inverter of FIG. 14;

FIG. 16 a illustrates the simulated voltage response of the ISFET based NOR gate of FIG. 12;

FIG. 16 b illustrates the simulated voltage response of a first ISFET of the ISFET based NOR gate of FIG. 12; and

FIG. 16 c illustrates the simulated voltage response of a second ISFET of the ISFET based NOR gate of FIG. 12.

DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS OF THE INVENTION

An n-channel FET such as that illustrated in FIG. 3 is a four terminal device, consisting of a p-type silicon substrate (B) and two highly-doped n-type wells known as the source (S) and the drain (D). The silicon surface is covered with a silicon dioxide insulator. A polysilicon gate contact (G) controls the charge within the region under the insulator surface between source and drain, known as the channel.
As the voltage V_Gapplied to the gate is increased, positive charge is initially repelled from the channel forming a depletion layer with no mobile charge carriers and a net negative charge. As the gate voltage is further increased, this depletion layer widens until electrons begin to be drawn from source and drain into the channel, forming an inversion layer. The transistor is usually operated above a certain threshold voltage for which the channel is strongly inverted and the mobile electrons in the inversion layer drift across the channel when a potential difference is applied between drain and source. As already noted, for this mode of operation drain current is related to the gate voltage by a square law or linear relation.
The so-called “weak inversion” mode of operation involves maintaining the gate voltage lower than the threshold voltage such that the channel is depleted and only a thin inversion layer exists. In weak inversion, the mobile charge in the thin inversion layer is too low to contribute significantly to any drift current across the horizontal electric field. Drain current in weak inversion is due to the diffusion of electrons across a concentration gradient between source and drain. Since the electron concentrations at source and drain and along the channel are related to the barrier potentials at those points by the Boltzmann distribution, it follows that the drain current is exponentially related to Vs, Vd and Vg relative to Vb, scaled by the thermal voltage U_T=kT/q or RT/F. That is:
I _d =I ₀exp(V _G /nU _T)[exp(−V _S /U _T)−exp(−V _D /U _T)] (4)
where I₀is the pre-exponential multiplier and n is the sub-threshold slope factor.
For the ISFET, the reference electrode acts as a remote gate and the chemically sensitive membrane deposited on top of the SiO₂insulator is exposed directly to the sample solution. The extent of inversion in the channel beneath the insulator is dependent not only on the voltage applied to the reference electrode, but also on the accumulation of charge from ions in solution on the sensing membrane. The build-up of ions at the surface of the membrane is related to the concentration of the ionic species in the sample by the site binding and Gouy-Chapman double layer models. Since any positive charge build-up on the membrane surface must be mirrored by a negative charge build-up in the channel, changes in ionic concentration of the sample will be reflected directly in the ISFET's weak inversion drain current.
A knowledge of the relationship between membrane surface charge and species concentration, together with the fact that weak inversion ISFET current is proportional to membrane surface charge, means that electronic circuits performing simple mathematical manipulation can be used to obtain a direct relation between species concentration and current even in ChemFETs and EnFETs. Furthermore, the sensitivity of the weak inversion ISFET current to ion concentration is independent of temperature, since the temperature-scaled Boltzmann distributions of both the electrons in the channel and the ions in solution cancel each other out.
The large-signal equation for drain current in a MOSFET biased in the weak inversion region is given by:
$\begin{matrix} \begin{matrix} I_{D} = I_{o} \exp (\frac{V_{GS}}{{nU}_{T}}) (1 - \exp (\frac{- V_{DS}}{U_{T}})) \\ = [2 n \frac{KW}{L} U_{T}_{2} \exp (\frac{V_{GS} - V_{T 0}}{{nU}_{T}}) \exp (\frac{(n - 1) V_{BS}}{{nU}_{T}})] \\ (1 - \exp (\frac{- V_{DS}}{U_{T}})) \end{matrix} & (5) \end{matrix}$
where β=KW/L, V_TOis the threshold voltage for V_BS=0 and n is the sub-threshold slope factor and saturation is assumed for V_DS>4U_T. This equation also holds for a weakly-inverted ISFET, since all additional chemical phenomena are represented as the modulation of its threshold voltage by a potential across the electrolyte which is linearly proportional to pH. Since pH is exponentially related to hydrogen ion concentration, a direct relation between hydrogen ion concentration and weak inversion drain current can be developed.
Any circuit which extracts the potential across the electrolyte and converts it to a weak inversion current signal is of significant interest for real-time chemical signal processing since hydrogen ion concentration is a more natural parameter for signal processing than pH. The current mirror, illustrated in FIG. 4, is the simplest of such circuits.
If a diode-connected ISFET is biased with a current source and its reference electrode connected to the gate of a MOSFET as in FIG. 4, and assuming geometrically and electrically-matched devices and ignoring ΔV_DSerror, then I₁≠I₂because of the ΔV_thbetween the ISFET and the MOSFET, which is known from (1) to be equal to V_chem.
Using equation (5) for perfectly matched, saturated devices:
$\begin{matrix} \frac{I_{D 2}}{I_{D 1}} = \exp (\frac{V_{chem}}{{nU}_{T}}) & (6) \end{matrix}$
Substituting in (2) for V_chemand the logarithmic relation between pH and hydrogen ion concentration, pH=−log₁₀[H⁺], we find that the current ratio I_D2/I_D1is proportional to a known power of hydrogen ion concentration and is independent of temperature effects:
$\begin{matrix} \begin{matrix} \frac{I_{D 2}}{I_{D 1}} = \exp (\frac{γ}{{nU}_{T}}) \exp (\frac{2.3 U_{T} α p H}{{nU}_{T}}) \\ = \exp (\frac{γ}{{nU}_{T}}) \exp (\frac{- α \ln [H^{+}]}{n}) \\ = K_{chem} \cdot {[H^{+}]}^{- α / n} \end{matrix} & (7) \end{matrix}$
This significant result shows that the drain current in a weakly-inverted ISFET is controlled exponentially by its gate-source and bulk-source potentials, and is scaled by a temperature-independent parameter which is proportional to a known power (less than unity since 0<α<1 and n>1) of the hydrogen ion concentration, that is:
I _D(ISFET) =I _D(MOS) ·K _chem ⁻¹·[H+]^α/n (8)
The temperature insensitivity of the ISFET biased in the weak inversion region makes it ideal for use in the transduction stage of a chemical sensor. Use of MOSFETs in weak inversion is also advantageous for on-chip processing circuitry since the exponential relation between drain current and terminal voltages can be exploited to implement mathematical manipulation using very simple, low-power circuits.
In order to obtain an output current which is directly proportional to [H⁺], some further manipulation of equation (7) is required. Translinear circuits which exploit the linear relation between transconductance and drain current in sub-threshold MOS transistors can be used to perform multiplication, division and power law functions on current signals (although bipolar transistors exhibiting this relationship might alternatively or additionally be used).
The circuit illustrated in FIG. 5 is presented as an example of an input stage whose output is a current I_outdirectly proportional to the number of H⁺ ions (protons) in the bulk of the electrolyte. This circuit is referred to here as an “HCell”. It is illustrative of how apt it is to exploit the exponential behaviour of field effect transistors in weak inversion to obtain real chemical parameters directly. The ISFET used in the HCell has its ion-sensitive membrane chemically tailored such that its pH-sensitivity gives α/n=0.5.
Assuming saturation and ignoring ΔV_DSerror, it can be shown from equation (5) that the drain currents between the ISFET X2 and MOSFET M1 are related by:
$\begin{matrix} \frac{I_{xis}}{I_{b 1}} = \frac{n_{2} β_{2}}{n_{1} β_{1}} \exp (\frac{- Δ V_{T 0}}{{nU}_{T}}) \exp (\frac{(n - 1) (V_{b 2} - V_{b 1})}{{nU}_{T}}) \exp (\frac{V_{S 1} - V_{S 2}}{U_{T}}) where {ΔV}_{T 0} = V_{T 0 (ISFET)} - V_{TO (MOS)} = V_{chem} & (9) \end{matrix}$
For the circuit shown in FIG. 5, assuming geometrically and electrically matched devices with equal device dimensions and bulks connected to ground, equation (9) becomes:
$\begin{matrix} \frac{I_{xis}}{I_{b 1}} = \exp (\frac{- γ - 2.3 α U_{T} p H}{{nU}_{T}}) \exp (\frac{- V_{ref}}{U_{T}}) = {K_{chem} [H^{+}]}^{0.5} \exp (\frac{- V_{ref}}{U_{T}}) where K_{chem} = \exp (\frac{- γ}{{nU}_{T}}) & (10) \end{matrix}$
The reference voltage V_refis used to set the bias point such that both M1 and X2 operate in weak inversion. The limit on pH input range is the voltage range of the operable weak inversion region and the pH sensitivity S. For weak inversion regions of approximately 400 mV V_GSrange, typical ISFETs with sensitivities of 50 mV/pH have 7 or 8 pH units of dynamic range.
Transistors M3 to M6 form a translinear loop for which, using the translinear principle, we obtain the relation:
$\begin{matrix} I_{out} = \frac{I_{xis}^{2}}{I_{b 2}} & (11) \end{matrix}$
Substituting in the result from (10)
$\begin{matrix} I_{out} = \frac{I_{b 1}^{2} \exp (\frac{- 2 V_{ref}}{{nU}_{T}}) K_{chem}^{2} [H^{+}]}{I_{b 2}} & (12) \end{matrix}$
Setting Ib1=Ib2 we get a direct relation between the current ratio Iout/Ib1 and [H+]:
$\begin{matrix} \frac{I_{out}}{I_{b 1}} = \exp (\frac{- 2 V_{ref}}{{nU}_{T}}) K_{chem}^{2} [H^{+}] & (13) \end{matrix}$
This ratiometric approach to signal processing reduces the temperature dependence of the circuit assuming that the devices are in close proximity. Using a Proportional To Absolute Temperature (PTAT) reference voltage V_ref, and assuming that K_chemand α are temperature-independent to a first order approximation, we see that the current ratio is directly proportional to hydrogen ion concentration with inherent temperature compensation.
The circuit of FIG. 5 was simulated on AMS 0.8μ technology using a simplified version of the Martinoia macromodel for a silicon nitride ISFET with pH_pzc=3 and perfectly linear pH sensitivity (constant α for all pH and temperature). It was found that n≈1.43 and a was chosen to be 0.715 corresponding to a sensitivity of 42.6 mV/pH at T=300K.
Vdd=1.8V, Ib1=Ib2=10 nA, and Vref was chosen to be 50 mV in order to centre the region of operation on pH 7. The output current, shown in FIG. 6 gives good linearity for the pH range 5 to 9, for which
$\frac{\partial \ln (Iout)}{\partial p H} = - 2.303 \pm 0.19 .$
Large dimensions (W=432 μm, L=8 μm) were chosen for X2 and M1 to minimise the effect of matching errors. W/L for transistors M3 to M6 was 40 μm/8 μm.
Extending the principles presented herein to a circuit with several ISFETs each coding a different ionic concentration, would allow any chemical equation involving products, quotients and power law relations of ionic concentration to be processed in real time. Exploiting all four terminals of the ISFET by use of the bulk or ‘back gate’ as a second input for translinear manipulations further increases the flexibility of these principles. Furthermore, the inclusion of capacitors opens up this principle to a whole field of reaction kinetics differential equations.
The ISFET-MOSFET mirror is the simplest current mode input stage and has been presented here to illustrate how apt it is to convert the threshold voltage modulation caused by chemical phenomena to a current. Replacing the MOSFET in this configuration with an almost pH-insensitive ISFET (known as a REFET) would permit the use of a solid-state quasi-reference electrode since the unstable and unknown electrode potential is common to both ISFET and REFET and is cancelled in the current mirror topology. The difference in threshold voltage ΔV_THis smaller between an ISFET and a REFET than between an ISFET and a MOSFET, decreasing ΔV_DSerrors and reducing the required V_bias, thereby reducing power consumption. Matching could also be improved. For a more robust circuit with higher SNR, a fully differential input stage should be used.
Smart sensing concepts based on weak inversion operation of the transistor applications beyond chemical discrimination using the selectivity of the membranes include:

- Real-time reaction monitoring and analytical data processing
- DNA-sequencing
- Fast acid-base titrations using kinetic models to extract the analytical information sought, well before reaching the equivalence point
- Implementation of “chemical decision trees” using ISFETs and other sensors directly as logic (true/false) devices with adjustable thresholds
- Medical diagnostics using real-time monitoring of blood and urine metabolite ratios
- Neural bridges
- Impurity detection

In addition to its suitability for use in a transducer input stage, the ISFET operating in the weak inversion can provide a basic building block for the digital processing of chemically related signals.
The standard CMOS inverter illustrated in FIG. 7 has the voltage transfer characteristic shown in FIG. 8, when operated in the weak inversion region. For the weak inversion region, in saturation and assuming V_BS=0, the drain current is given by:
$\begin{matrix} I_{D} = 2 n β U_{T}^{} \exp (\frac{V_{GS} - V_{T 0}}{{nU}_{T}}) & (14) \end{matrix}$
where n is the sub-threshold slope parameter, β=KW/L, U_Tis the thermal voltage and V_T0is the intrinsic threshold voltage.
At the switching threshold V_inth, M1 and M2 have equal drain currents.
$\begin{matrix} V_{inth} = \frac{V_{DD} - \langle V_{TOp} \rangle + \ln (\frac{I_{sn}}{I_{sp}}) V_{TOn}}{1 + \ln (\frac{I_{sn}}{I_{sp}})} & (15) \end{matrix}$
where sub-threshold slopes for NMOS and PMOS are assumed equal.
If the NMOS M1 of FIG. 7 is replaced by an n-channel ISFET, then, given that V_th(ISFET)=V_th(MOS)+V_chem, the switching threshold will be given as:
$\begin{matrix} V_{inth} = \frac{V_{DD} - \langle V_{TOp} \rangle + \ln (\frac{I_{sn}}{I_{sp}}) [V_{TOn} + V_{chem}]}{1 + \ln (\frac{I_{sn}}{I_{sp}})} & (16) \end{matrix}$
i.e. a decrease in switching threshold for pH<11 and an increase for pH>11.
FIG. 9 illustrates the shift in switching threshold caused by changing pH. A consequence of this characteristic is that if vi is fixed, then switching will occur exclusively due to changes in pH. Moreover, the pH threshold at which this switching occurs can be set by the choice of vi as shown in FIG. 10. For illustration, it is shown that the output is high for pH≧4 when vi is fixed at 0V and for pH≧8 when vi=100 mV.
The significance of this is that the circuit of FIG. 7 can be used as a comparator whose output is high for a pH greater than a threshold defined by the fixed input voltage on the gate, vi.
Logic gates may also be constructed using the ISFET operating in the weak inversion region. In the circuit of FIG. 11 for example, v_GAsets the pH threshold p_thAand v_GBsets p_thB. The n-channel ISFET is in saturation for pH<p_th. The output Y will only be 0 when both ISFETs are in saturation; for all other states the output is 1 as shown in the NAND truth table. Viewed from another perspective, we can say that the output is high for pH_A≧p_thAOR pH_B≧p_thB
By a similar analysis, the output in the circuit of FIG. 12 is high only when pH_A≧p_thAAND pH_B≧p_thBas can be deduced from the NOR truth table (assuming again that V_GAsets the pH threshold p_thAand v_GBsets p_thB).
It has been shown that ISFETs can be used to implement the basic logic gates, triggered not by 0 and 1 on the gate input, but by a pH <or > than a chosen threshold. ISFETs are therefore well-suited to the direct implementation of more complex logic functions.
Whilst the preferred operating mode for ISFETs/MOSFETs forming such digital circuits is the weak inversion mode, this is not essential, and they might alternatively be operated in the saturation mode. Circuits may use a combination of weak inversion mode and saturation mode devices.
It is further proposed here to make use of such ISFETs to create nucleic acid (e.g. DNA/RNA) “switches” that may perform in-vitro computation. Such a fully electronic method has the advantage that the instrumentation is inherently smaller and less complex than that used in traditional techniques for nucleic acid detection, such as the optical instrumentation used to detect fluorescent tagging. It also provides the advantage that digital computation can be performed on the same chip as that used to detect nucleic acid with a technique using very little static power even for large arrays, thus making it very suitable for a point-of-care environment.
According to these methods, ion-sensitive field effect transistor (ISFET) technology is applied to the detection of unique sequences within a sample of nucleic acid, using biochemical signals generated when the extension of a specific primer or “probe” occurs in the presence of the sample. In particular, these methods can be used to detect variations in DNA, such as a single nucleotide polymorphism (SNP).
DNA is a twisted double-helix structure consisting of nucleotides including one of four nucleobases: adenine (A), thymine (T), cytosine (C), and guanine (G). The double-helix structure is formed from two complementary DNA strands connected via hydrogen bonds. These bonds will only form between complementary bases that form a base pair. Adenine forms a base pair with thymine, and guanine with cytosine. In the nucleus of each human cell, there are 23 pairs of coiled DNA molecules called chromosomes, and the sequence of bases within an individual's DNA within these chromosomes holds all of their genetic and hereditary information.
The genetic sequences of different individuals are remarkably similar. For example, the DNA sequences of two different individual's can be identical for hundreds of bases. An SNP occurs when there is a difference of an individual base between two genetic sequences. More specifically, an SNP is defined as a mutation of a single base affecting at least 1% of a defined population, and there are currently over 1.8 million identified SNPs in the human genome. When these genetic variants occur in regions near each other they tend to be inherited together. These regions of linked variants are known as haplotypes.
Detection of unique sequences of DNA, or of single bases as in SNP detection, makes use of DNA synthesis which happens during the process DNA replication. According to this process, the double-stranded structure of DNA breaks into two, whereby each strand then acts as a ‘template’. If the template strand is hybridised to a short complementary ‘primer’ strand which serves as the starting point for DNA chain extension, nucleotides are added to the to the primer's 3′ end in such as way as to copy the template strand in the correct sequence according to the rules of complementary pairs (i.e. A can only pair with T and C can only pair with G) in a process known as ‘chain extension’ or ‘DNA synthesis’. Each single strand of DNA thus becomes double-stranded DNA, resulting in DNA replication. This incorporation of complementary nucleotides into a growing strand of DNA produces or consumes hydrogen ions depending on the pH at which the reaction occurs. An ISFET, as described above, can therefore be used to detect an SNP as the resulting change in pH can be measured as a change in the ISFET threshold voltage.
It is hypothesised that the incorporation of complementary nucleotides into a growing strand of DNA yields pyrophosphate (HP₂O₇ ⁻) which under appropriate biochemical conditions is hydrolysed to orthophosphate (H₂PO₄ ⁻):
HP ₂0₇ ⁻+H ₂0
2H₂PO₄ ⁻+H⁺ (17)
It is this hydrolysis reaction that produces or consumes hydrogen ions depending on the pH at which it occurs. As such, in-vitro optimization of the reaction environment to maximize the rate of pH change therefore leads to the production of hydrogen ions during nucleotide incorporation, resulting in a pH decrease.
An alternative hypothesis for the H+ release mechanism is pH-specific nucleotide hydrolysis in a growing oligomer:
dNTP+DNA
PPi+DNA+nH⁺ (18)
The nH+represents the pH-dependent mass balance of the proton release. The value of n, the number of protons, can be positive or negative, and is decided by a complex function of the starting pH and the pKa of dNTP, PPi and DNA under specific experimental conditions. The pKa of each individual component determines the number of protons dissociated or remaining at a given pH. For a given pH, n can then be calculated by subtracting the number of un-dissociated protons of all reactants from those of the products. If n is positive, hydrogen ions are liberated, and pH decreases. If n is negative, pH increases.
For the purpose of detection of a specific SNP, a single-stranded DNA probe can be designed to bond with the target DNA sample, forming a probe-target hybrid, in order to uniquely identify the location of the SNP within the target DNA being analysed. For example, if the probe length is n nucleotides, then the identity of the nucleotide at position n+1 within the target can be interrogated by adding each of the different nucleotides, with bases A, T, C and G, in parallel reaction chambers, each with its own ISFET sensor. Only a complementary nucleotide will be incorporated (extend the probe-target hybrid), thereby releasing protons and leading to a measurable change in ISFET threshold voltage. A non-complementary nucleotide will not be incorporated. It should be noted that the DNA probe can be comprised of either native nucleotides or modified nucleotides. Examples of modified nucleotides are phosphothioate oligonucleotides, fluorescently labelled oligonucleotides, or other nucleotides with modifications/replacements on the phosphodiester backbone, 3′ OH, 2′H, 2′OH, or nucleoside base.
As an alternative, the DNA probe could include the SNP to be detected. In this case, in order to identify the nucleotide at position n within the target, four separate probes could be used in four parallel reaction chambers. Each probe would have one of the four nucleotides (with bases A, T, C or G) at position n, and each reaction chamber would contain an excess of all four nucleotides (i.e. dATP, dCTP, dGTP, dTTP). In the reaction chambers where the probe matched the target DNA, multiple-base extension would occur for as many bases as the target is longer than the probe (e.g. for a target length of 50 bases (from single-stranded amplification product) and a probe length of 30 bases, including the SNP to be discriminated, this would give an extension of 20 bases). In homozygous individuals only one reaction chamber would show a signal. However, in heterozygous individuals two reaction chambers would show a signal.
The examples described above make use of four reaction chambers, each differentiated by a single nucleotide. In many cases, only two parallel reaction chambers are required for the application of SNP identification, since it is only necessary to discriminate a common “mutant” form from a “wild type”, as opposed to needing to identify the specific nucleotide mutation. In this case, only two probes, each with a different base at position n are required (or equivalently, for the single base extension detection method, two chambers each containing a different dNTP). However, another two reaction chambers may or may not be included in the implementation to act as a negative control.
Using these methods it can be determined which nucleotides have been incorporated by observing which ISFET reaction chamber gives a discrete pH fluctuation. From this it can then be determined which nucleotide is present at the location of interest and whether this is a mutation (SNP) or “wild type” (no SNP).
For the purpose of detection of a nucleic acid sequence, such as identifying a particular pathogen, a similar principle is used. Here, the probe is designed to be complementary to a unique sequence in the target nucleic acid (i.e. DNA/RNA). If the target nucleic acid is present, hybridisation occurs, and in the presence of all four nucleotides, extension will occur, causing hydrogen ions to be released and trigger the threshold of the ISFET-based switch. In such circumstances, one or more control probes should be used in parallel reaction chambers.
In the case of SNP detection and subtyping (where typically, there would be a single-base mismatch at the 3′ end in the case of target DNA not matching the probe), those probes that do not match the target DNA still bind to the target but to a lesser extent than those probes that do match the target DNA. However, the SNP is detected by measuring pH changes that occur due to extension of the probe-target hybrid and, whilst extension can occur when the DNA target does not match the probe, this is regulated by enzyme fidelity. With pathogen detection, probes are designed such that they are highly specific to a unique “signature” sequence within the target DNA of the pathogen. As such, mismatch probes would not bind to the target DNA.
To further improve allele-specific discrimination, probes with multiple mismatch pairings within the probe other than at position n (“the 3′ end”) can be introduced [1]. A single mismatch can sometimes be weakly extended by DNA polymerase, known as “pseudo-positive” extension (i.e. extension of a probe with a mismatch at the 3′ end which should not hypothetically extend). It is proposed in [1] that the problem of A/C or T/C mismatch extension is alleviated by including two extra mismatches just before the last base at the 3′ end of the probe. When the electrical signal from a probe which matches the target (i.e. a “match probe”) is compared to electrical signal from a probe which is chosen not to match the target (i.e. a “mismatch” or “control probe”), the additional mismatches should be included in both probes. Discrimination is thus improved because background noise from pseudo-positive reactions with the control probe is suppressed, whereas some extension of the match probe will still occur because there is a match at the end of the chain, despite the mismatches just before the end of the probe.
In practice, the target DNA (which is the product of a process involving DNA extraction from the cell, amplification by a process such as the polymerase chain reaction (PCR) and then conversion to single-stranded DNA) can be hybridised to the probe, to produce the probe-target hybrid, prior to exposing an ISFET to a solution containing the probe-target hybrid together with the other reaction ingredients. These reaction ingredients would include, but are not limited to, the DNA polymerase enzyme, dNTP, MgCl₂and NaCl. Alternatively, an ISFET can be coated with a reagent membrane (i.e. immobilised and/or freeze-dried and/or stored in soluble beads or foil blisters located in close proximity to the ISFET) and the target DNA introduced onto the membrane, thereby being more amenable to a fully automated solution in a disposable cartridge format comprising sensors, reagents, thermal elements, fluidics, and temperature control.
FIG. 13 illustrates changes in output voltage of a standard ISFET source-drain follower readout circuit, for the case of nucleotide incorporation (SNP present in target, as shown by the grey line) and no nucleotide incorporation (no SNP present in target, as shown by the black line).
These concepts can be extended to implement nucleic acid logic using the discrete pH changes caused by events such as primer extension for SNP detection. This has a practical utility in the direct transduction of a nucleic acid input to a binary output, such as that shown in Table 1. As described above, a standard CMOS inverter (FIG. 7) can be adapted to perform a pH-thresholding operation by replacing one of the MOSFETs with an ISFET. The circuit can be implemented in strong or weak inversion, but should be restricted to weak inversion if low power operation is desired. Also as described above, from a behavioural point of view, the ISFET can be considered to be a MOSFET with a chemically-sensitive threshold voltage, which varies approximately linearly with pH. Therefore, if the NMOS M1 of FIG. 7 is replaced by an n-channel ISFET, as shown in FIG. 14, the circuit can be used as a pH comparator. Simulation results have shown the uncertainty region around the threshold pH to be approximately ±0.3 pH units (16 mV). This is therefore a useful technique for array-based thresholding operations where qualitative trends from discrete changes caused by reaction intermediaries are being monitored, rather than precise quantification of targets.
FIGS. 15 a and 15 b illustrate simulations showing how the measured data from nucleotide incorporation could be converted to a digital output by the ISFET-based DNA logic NOT gate circuit of FIG. 14. FIG. 15 a illustrates the pH response due to nucleotide incorporation during primer extension when no SNP is present (as shown by the black line) and when a SNP is present (as shown by the grey line), and FIG. 15 b illustrates the simulated output from the DNA logic inverter circuit to this change in pH.
These principles can be further extended to multiple nucleic acid inputs, creating other nucleic acid logic circuits that perform simple logical operations providing a yes/no answer. This is particularly useful for detecting the presence or absence of a combination of multiple SNPs, such as those of a haplotype, which can give information about the risk of genetic predisposition to certain diseases such as heart disease or Type II diabetes. Furthermore, some SNPs are linked to drug metabolism and therefore can be used to predict whether a patient should take a high or low dose of certain prescription drugs. For example, in the case of warfarin, an anticoagulant blood-thinning agent, a high dose in people with certain SNPs could cause excessive bleeding. Two SNPs in particular have been found to correspond with lower dose requirements, due to the fact that they indicate slower metabolism of the drug. These SNPs are on the alleles notated “CYP2C9*2” and “CYP2C9*3” which code for liver enzymes, and a patient with either or both of these SNPs will require a lower dose of warfarin. Table 2 shows a simplified truth table illustrating the NOR relationship of warfarin dose to the CYP2C9 genotype.
Alternatively, ISFET-based nucleic acid detection circuits could be configured to provide a weighted score. For example, a third SNP, notated as “VKORC1”, can also be used in warfarin-dose calculation [2] and rather than the Boolean answer (high dose/low dose) described by the 2-input NOR example, drug dose can be calculated as part of a hardware-coded algorithm with multiple ISFET switches providing a weighted score. By way of further example, this would also be particularly useful when detecting the few alleles that indicate an individual's risk of developing breast cancer, providing a weighted score to indicate the risk.
This NOR relationship can be implemented using ISFET-based DNA Logic, using a circuit such as that shown in FIG. 12, where the classical MOSFETs are replaced with ISFETs at the inputs, provided that appropriate DNA probes and biochemical reagents are used to detect nucleotide incorporation during primer extension. FIGS. 16 a, 16 b and 16 c illustrate the simulated voltage response of the NOR circuit (V_out), ISFET 2 (V_B) and ISFET 1 (V_A) of FIG. 12 respectively. Each of FIGS. 16 a, 16 b and 16 c are separated, by dashed vertical lines, into four results corresponding to the presence or absence of SNP A and SNP B in a DNA sample, as shown in the Table 3. When one or both of the ISFET sensors detects a SNP the output will be zero. Only if neither ISFET detects a SNP will the result be a digital one.
The operation of the ISFET as a means of implementing a “DNA logic” unit capable of calculating truth-table type outputs can therefore be used for sensor array-based processing for on-chip information extraction from multiple biomarkers. This also provides that the conversion to digital 0 or 1 signals occurs as soon as possible, thereby providing that the signal is more robust to degradation from interconnects and sources of interference between the sensor chip and any off-chip interfacing. Such DNA logic circuits could of course be used with any number of ISFETs to detect and perform functions involving any number of SNPs. Furthermore, these principles could be expanded to arrays of such logic circuits, wherein each logic circuit computes a function of multiple DNA inputs and combinations of these various logic circuits perform DNA computations. Such arrays could implement a truth table associated with 10s, 100s, 1000s or even 1,000,000s of SNPs or other DNA markers (i.e. a large array of sensors and a large array of logic gates). A further advantage of using “DNA logic” operating on a similar principle to standard CMOS logic is the low static power consumption of the detection array. If biased appropriately at a certain reaction pH, the only static current drawn will be the leakage currents of multiple ISFET input devices, which can be minimised through biasing of their bulk terminals, with their gates only switched “on” by a pH change over a certain threshold due to a particular target-probe match. In addition to performing logic functions on the elements of an array providing truth-table type outputs, it is also possible to include a weighting function on each gate so that the result of the computation is not necessarily a 0 or 1. For example, such weighting could be used to provide a value representing a coefficient by which the dose of a drug could be adjusted.
The use of DNA to switch on and off transistors to create a synthetic form of DNA logic is useful for CMOS ISFET lab-on-chip based technologies, providing the advantages proffered by silicon of disposability, portability, scalability to ever-increasing numbers of devices (i.e. according to a ‘Moore's law’ now applied to chemical sensing elements), and leveraging the economy of scale of the semiconductor industry for mass fabrication of devices using mature techniques rather than requiring development of custom manufacturing facilities. In addition, it creates DNA computation as an extension to the work already being pioneered within cellular computing. This method provides for rapid, label-free nucleic acid detection and interpretation, scaleable to large arrays and with ease of integration of the results with an IT network, providing advantages over existing optical methods of nucleic acid detection of reduced cost, increased flexibility, connectivity, manufacturability, simplicity of data processing and reduced power consumption.
The methods described above can be used with or without nucleic acid amplification, either by thermocycling e.g. polymerase chain reaction, or isothermal methods e.g. transcription-mediated amplification, ligase chain reaction or strand displacement methods. The methods described above provide means for developing a fully portable, low-power unit for nucleic acid detection which could be operated outside a laboratory, providing clinicians with a technology which would save time, money and lives by making important diagnostic information available on-demand.
Whilst the methods described above are based upon detection of SNPs in DNA, they can also be applied to other applications including the identification and interpretation of any set of DNA-based biomarkers. In addition, these methods can similarly be used to perform signal/logic processing for any other reactions that lead to a change in pH, such as those associated with mRNA, proteins, antibody/antigen interactions, and interactions between biological cells.
Furthermore, these applications can be achieved with and/or without the use of an intermediate label, wherein the pH can be generated either directly from the target itself (e.g. primer extension releasing), or by using a label or ‘reporter’ of some kind, which releases or consumes protons as part of a secondary reaction which is triggered by the presence of the target, and which is used to indicate the presence of the analyte of interest. In other words, the change in pH can also come from a secondary pathway wherein the binding of a nucleic acid probe to the target can act as a trigger for more pH change from a secondary reaction. For example, a nucleic acid probe can also be an aptamer, which blocks the activity of a proton generation enzyme such as urease. The result of the binding of the aptamer, and the extension following the binding, becomes the trigger to activate the urease. Urease then catalyses the conversion of urea to ammonia and the production of ammonia results in a much larger pH change.
It will be appreciated by those of skill in the art that various modifications may be made to the above described embodiments without departing from the scope of the invention. In one modification, the single gate ISFET described above is replaced with a multi-gate ISFET. In another modification, the “back-gate” or substrate is used as an additional input of the device.
It will also be appreciated that “dynamical” mathematical systems including instantaneous companding systems can be created by adding capacitors to the gates of ISFETs (and other MOSFETs of the associated circuitry), thus converting the characteristics into large signal non-liner time domain biochemical functions, e.g. log domain filters and processors. Such functionality relies upon the exponential/logarithmic characteristics of the weak inversion MOSFET.

TABLE 1

Truth Table of DNA Logic NOT Gate

pH < pHth	DNA INPUT	OUTPUT

0	No SNP detected	1
1	SNP detected	0

TABLE 2

Truth Table of Warfarin dose predicted from CYP2C9 genotype

CYP2C9*2	CYP2C9*3	Dose

Absent (0)	Absent (0)	High (1)
Absent (0)	Present (1)	Low (0)
Present (1)	Absent (0)	Low (0)
Present (1)	Present (1)	Low (0)

TABLE 3

Truth Table of DNA Logic NOR Gate

pH < pHthA	pH < PHthB	Vout
SNP A present	SNP B present	OUTPUT

0	0	1
0	1	0
1	0	0
1	1	0

REFERENCES

[1]H. Yaku et al. “Design of allele-specific primers and detection of the human ABO genotyping to avoid the pseudopositive problem”, vol 29, pp 4130-4140, 2008
[2] B. F. Gage, C. Eby, P. E. Milligan, G. A. Banet, J. R. Duncan, and H. L. Mcleod, “Use of pharmacogenetics and clinical factors to predict the maintenance dose of warfarin,” Thrombosis and Haemostasis, vol. 91, no. 1, pp. 87-94, 2004.

Claims

1. A digital signal processing circuit, one or more switches of the circuit being provided by an ion sensitive field effect transistor.

2. A circuit according to claim 1, wherein the or each ion sensitive field effect transistor comprises an analyte sensitive membrane which is exposed in use to a medium to be monitored.

3. A circuit according to claim 1, wherein the circuit is configured to operate as a comparator for comparing a value of a parameter measured by the ion sensitive field effect transistor with a threshold value, the circuit comprising an ion sensitive field effect transistor and a metal oxide semiconductor transistor arranged in an inverter configuration.

4. A circuit according to claim 3, wherein one of the ion sensitive field effect transistor and the metal oxide semiconductor transistor is an n-channel device and the other is a p-channel device.

5. A circuit according to claim 1, wherein the digital signal processing circuit is arranged to implement one or more of the following functions: AND, NAND, OR, XOR, NOR.

6. A circuit according to claim 1, wherein the digital signal processing circuit uses CMOS logic.

7. A circuit according to claim 1, wherein the digital signal processing circuit comprises biasing means for biasing the or each ion sensitive field effect transistor in the weak inversion region.

8. A device for detecting a plurality of chemical reactions and evaluating a logical function having the result of each of the plurality of chemical reactions as its inputs, the device comprising:

at least one reaction chamber for each of the plurality of chemical reactions; and

each reaction chamber being provided with an ion sensitive field effect transistor configured to operate as a switch for detecting one of the plurality of chemical reactions, the ion sensitive field effect transistors being coupled together to form a digital signal processing circuit for evaluating the logical function.

9. A device according to claim 8, wherein one or more of the ion sensitive field effect transistors comprise an analyte sensitive membrane that is exposed within the reaction chamber.

10. A device according to claim 8, wherein the digital signal processing circuit is arranged to implement one or more or a combination of the following functions: AND, NAND, OR, XOR, NOR.

11. A device according to claim 8, wherein the digital signal processing circuit uses CMOS logic.

12. A device according to claim 8, wherein the digital signal processing circuit comprises biasing means for biasing one or more of the ion sensitive field effect transistors in the weak inversion region.

13. A method of evaluating a logical function having as one of its inputs the value of a parameter of a medium, the method comprising:

configuring an ion sensitive field effect transistor to operate as a switch of a logic circuit; and

exposing the ion sensitive field effect transistor to said medium.

14. A method according to claim 13, the method further comprising configuring the ion sensitive field effect transistor to switch when a change in the concentration of hydrogen ions in the medium indicates incorporation of a nucleotide into a sample of genetic material.

15. A method of detecting a plurality of chemical reactions and evaluating a logical function having the result of each of the plurality of chemical reactions as its inputs, the method comprising:

for each of the plurality of chemical reactions, configuring at least one ion sensitive field effect transistor to operate as a switch of a logic circuit, and exposing the ion sensitive field effect transistor to the chemical reaction.

16. A method according to claim 15, the method further comprising configuring one or more of the ion sensitive field effect transistors to switch when a change in the concentration of hydrogen ions in the medium indicates incorporation of a nucleotide into a sample of genetic material.

17. A method of detecting the presence of a unique genetic sequence in a target sample of single-stranded genetic material, the method comprising:

performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to complement the unique genetic sequence;

configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of one or more nucleotides into genetic material;

exposing the ion sensitive field effect transistor to a solution comprising a product of the hybridisation reaction and an excess of all nucleotides; and

monitoring the output of the ion sensitive field effect transistor to determine whether it switches and to thereby determine whether the probe has hybridised with the target sample.

18. A method of detecting the presence of a plurality of unique genetic sequences in a target sample of single-stranded genetic material, the method comprising:

for each of the plurality of unique genetic sequences:

configuring the ion sensitive field effect transistors into a digital processing circuit to provide an output of a logical function as an indication of the presence or absence of each of the plurality of unique genetic sequences.

19. A method of identifying a nucleotide present at a specific location in a target sample of single-stranded genetic material, the method comprising:

performing a hybridisation reaction involving said target sample and a nucleotide probe in order to form a probe-target hybrid, the nucleotide probe being configured to uniquely identify the specific location;

for each candidate nucleotide that may be present at the specific location:

configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of a nucleotide into genetic material;

exposing the ion sensitive field effect transistor to a solution comprising the probe-target hybrid and a nucleotide complementary to said candidate nucleotide; and

monitoring the output of the ion sensitive field effect transistors to determine which switches and to thereby determine which of said complementary nucleotides has been incorporated at the specific location.

20. A method of identifying a nucleotide present at a specific location in a target sample of single-stranded genetic material, the method comprising:

for each candidate nucleotide that may be present at the specific location:

performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to uniquely locate a nucleotide complementary to the candidate nucleotide at the specific location;

configuring an ion sensitive field effect transistor to operate as a switch, and to switch in the presence of the incorporation of a plurality of nucleotides into genetic material;

monitoring the outputs of the ion sensitive field effect transistors to determine which switches and to thereby determine which of the probes has hybridised with the target sample.

21. A method of detecting the presence of a single nucleotide polymorphism in a target sample of single-stranded genetic material, the method comprising:

performing a hybridisation reaction involving said target sample and a nucleotide probe in order to form a probe-target hybrid, the nucleotide probe being configured to uniquely identify the expected location of the single nucleotide polymorphism;

exposing the ion sensitive field effect transistor to a solution comprising the probe-target hybrid and a complementary nucleotide associated with the single nucleotide polymorphism; and

monitoring the output of the ion sensitive field effect transistor to determine whether it switches and to thereby determine whether the complementary nucleotide has been incorporated at the specific location.

22. A method of detecting the presence of a single nucleotide polymorphism in a target sample of single-stranded genetic material, the method comprising:

performing a hybridisation reaction involving said target sample and a nucleotide probe, the nucleotide probe being configured to uniquely locate a complementary nucleotide associated with the single nucleotide polymorphism at the expected location of the single nucleotide polymorphism;

monitoring the outputs of the ion sensitive field effect transistor to determine whether it switches and to thereby determine whether the probe has hybridised with the target sample.

23. A method of detecting the presence of a plurality of single nucleotide polymorphisms in a target sample of single-stranded genetic material, the method comprising:

for each of the single nucleotide polymorphisms:

configuring the ion sensitive field effect transistors into a digital processing circuit to provide an output of a logical function as an indication of the presence or absence of each of the plurality of single nucleotide polymorphisms.

24. A method of detecting the presence of a plurality of single nucleotide polymorphisms in a target sample of single-stranded genetic material, the method comprising:

for each of the single nucleotide polymorphisms:

25. A method according to claim 17 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

26. A method according to claim 18 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

27. A method according to claim 19 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

28. A method according to claim 20 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

29. A method according to claim 21 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

30. A method according to claim 22 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

31. A method according to claim 23 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.

32. A method according to claim 24 wherein the hybridisation reaction occurs during a nucleic acid amplification cycle.