US 20120168826 A1
To reduce the pixel size to the smallest dimensions and simplest form of operation, a pixel may be formed by using only one ion sensitive field-effect transistor (ISFET). This one-transistor, or 1T, pixel can provide gain by converting the drain current to voltage in the column. Configurable pixels can be created to allow both common source read out as well as source follower read out. A plurality of the 1T pixels may form an array, having a number of rows and a number of columns and a column readout circuit in each column.
1. A device comprising:
an array of chemically-sensitive field-effect transistors (chemFETs), at least some chemFETs in the array of chemFETs comprising:
a first source/drain terminal and a second source/drain terminal; and
a floating gate coupled to a passivation layer;
a plurality of first conductive lines coupled to chemFETs in the array, wherein each first conductive line in the plurality of first conductive lines is directly connected to the first source/drain terminals of chemFETs in a column in the array; and
a plurality of second conductive lines coupled to chemFETs in the array, wherein each second conductive line is directly connected to the second source/drain terminals of chemFETs in a row in the array;
bias circuitry coupled to the array to apply a read bias to selected chemFETs via the first conductive lines and second conductive lines; and
sense circuitry coupled to the array to sense charge coupled to the floating gates of selected chemFETs.
2. The device of
3. The device of
the bias circuitry applies a read voltage to a selected first conductive line in the plurality of first conductive lines during a read interval; and
the sense circuitry includes a sample circuit to read a selected chemFET connected to the selected first conductive line and a selected second conductive line during the read interval.
4. The device of
5. The device of
6. The device of
7. The device of
the read bias includes a read voltage on a selected first conductive line applied to boost by capacitive coupling the floating gate of a selected chemFET to a boosted voltage level, thereby inducing current flow through the selected chemFET to establish a voltage level on a selected second conductive line; and
the read bias further includes a voltage on first conductive lines connected to unselected chemFETs connected to the selected second conductive line, applied to bias by capacitive coupling the floating gates of the unselected chemFETs to an inhibit voltage level.
8. The device of
9. A device comprising:
an array of chemically-sensitive field-effect transistors (chemFETs), at least some chemFETs in the array of chemFETs comprising:
a first source/drain terminal and a second source/drain terminal; and
a floating gate coupled to a passivation layer;
a plurality of first conductive lines coupled to chemFETs in the array, wherein each first conductive line in the plurality of first conductive lines is directly connected to the first source/drain terminals of a corresponding first plurality of chemFETs in the array; and
a plurality of second conductive lines coupled to chemFETs in the array, wherein each second conductive line in the plurality of second conductive lines is coupled to the second source/drain terminals of a corresponding second plurality of chemFETs in the array.
10. The device of
11. The device of
12. The device of
inducing current flow between the selected second conductive line and the selected first conductive line through the selected chemFET during a read interval, thereby establishing an output voltage level on the selected second conductive line;
compensating for leakage current through unselected chemFETs connected to the selected second conductive line during the read interval; and
reading the selected chemFET based on the output voltage level on the selected second conductive line.
13. The device of
pre-charging the selected second conductive line to a pre-charge voltage level prior to the read interval;
applying a read voltage to the selected first conductive line during the read interval to boost by capacitive coupling the floating gate of the selected chemFET to a boosted voltage level, thereby inducing the current flow to establish the output voltage level on the selected second conductive line.
14. The device of
15. The device of
16. The device of
17. The device of
18. A device comprising:
an array of chemically-sensitive field-effect transistors (chemFETs) arranged in a plurality of rows and a plurality of columns, chemFETs in the array of chemFETs comprising:
a first source/drain terminal and a second source/drain terminal; and
a floating gate coupled to a passivation layer;
a plurality of column lines coupled to chemFETs in the array, wherein each column line in the plurality of column lines is directly connected to the first source/drain terminals of chemFETs arranged in a corresponding column in the plurality of columns;
a plurality of row lines coupled to chemFETs in the array, wherein each row line in the plurality of row lines is directly connected to the second source/drain terminals of chemFETs arranged in a corresponding row in the plurality of rows;
circuitry coupled to the plurality of column lines and row lines for reading a selected chemFET connected to a selected column line and a selected row line, the circuitry comprising:
bias circuitry coupled to the row lines to apply a read voltage to the selected row line in the plurality of row lines during a read interval, and to apply voltages to corresponding row lines and corresponding column lines connected to unselected chemFETs connected to the selected column line; and
sense circuitry to sense an ion-concentration of an analyte solution coupled to the floating gate via the passivation layer in the selected chemFET based on a sampled voltage level on the selected column line, wherein the sampled voltage level is established based on the current flowing on the selected column line.
19. The device of
20. The device of
a pre-charge circuit to pre-charge a selected column line to a pre-charge voltage level prior to a read interval; and
a sample circuit to read the selected chemFET based on the sampled voltage level on the selected column line during the read interval.
This application is a continuation of U.S. application Ser. No. 13/174,207 filed Jun. 30, 2011, which claims the benefit of priority to previously filed U.S. provisional patent application Ser. No. 61/360,493 filed Jun. 30, 2010, U.S. provisional application Ser. No. 61/360,495 filed Jul. 1, 2010, U.S. provisional application Ser. No. 61/361,403 filed Jul. 3, 2010, and U.S. provisional application Ser. No. 61/365,327 filed Jul. 17, 2010, the disclosures of all of which are incorporated herein by reference in their entireties.
Electronic devices and components have found numerous applications in chemistry and biology (more generally, “life sciences”), especially for detection and measurement of various chemical and biological reactions and identification, detection and measurement of various compounds. One such electronic device is referred to as an ion-sensitive field effect transistor, often denoted in the relevant literature as an “ISFET” (or pHFET). ISFETs conventionally have been explored, primarily in the academic and research community, to facilitate measurement of the hydrogen ion concentration of a solution (commonly denoted as “pH”).
More specifically, an ISFET is an impedance transformation device that operates in a manner similar to that of a MOSFET (Metal Oxide Semiconductor Field Effect Transistor), and is particularly configured to selectively measure ion activity in a solution (e.g., hydrogen ions in the solution are the “analytes”). A detailed theory of operation of an ISFET is given in “Thirty years of ISFETOLOGY: what happened in the past 30 years and what may happen in the next 30 years,” P. Bergveld, Sens. Actuators, 88 (2003), pp. 1-20 (“Bergveld”), which publication is hereby incorporated herein by reference in its entirety.
Details of fabricating an ISFET using a conventional CMOS (Complementary Metal Oxide Semiconductor) process may be found in Rothberg, et al., U.S. Patent Publication No. 2010/0301398, Rothberg, et al., U.S. Patent Publication No. 2010/0282617, and Rothberg et al, U.S. Patent Publication 2009/0026082; these patent publications are collectively referred to as “Rothberg”, and are all incorporated herein by reference in their entirety. In addition to CMOS, however, biCMOS (i.e., bipolar and CMOS) processing may also be used, such as a process that would include a PMOS FET array with bipolar structures on the periphery. Alternatively, other technologies may be employed wherein a sensing element can be made with a three-terminal devices in which a sensed ion leads to the development of a signal that controls one of the three terminals; such technologies may also include, for example, GaAs and carbon nanotube technologies.
Taking a CMOS example, a P-type ISFET fabrication is based on a P-type silicon substrate, in which an N-type well forming a transistor “body” is formed. Highly doped P-type (P+) regions S and D, constituting a source and a drain of the ISFET, are formed within the N-type well. A highly doped N-type (N+) region B may also be formed within the N-type well to provide a conductive body (or “bulk”) connection to the N-type well. An oxide layer may be disposed above the source, drain and body connection regions, through which openings are made to provide electrical connections (via electrical conductors) to these regions. A polysilicon gate may be formed above the oxide layer at a location above a region of the N-type well, between the source and the drain. Because it is disposed between the polysilicon gate and the transistor body (i.e., the N-type well), the oxide layer often is referred to as the “gate oxide.”
Like a MOSFET, the operation of an ISFET is based on the modulation of charge concentration (and thus channel conductance) caused by a MOS (Metal-Oxide-Semiconductor) capacitance. This capacitance is constituted by a polysilicon gate, a gate oxide and a region of the well (e.g., N-type well) between the source and the drain. When a negative voltage is applied across the gate and source regions, a channel is created at the interface of the region and the gate oxide by depleting this area of electrons. For an N-well, the channel would be a P-channel (and vice-versa). In the case of an N-well, the P-channel would extend between the source and the drain, and electric current is conducted through the P-channel when the gate-source potential is negative enough to attract holes from the source into the channel. The gate-source potential at which the channel begins to conduct current is referred to as the transistor's threshold voltage VTH (the transistor conducts when VGS has an absolute value greater than the threshold voltage VTH). The source is so named because it is the source of the charge carriers (holes for a P-channel) that flow through the channel; similarly, the drain is where the charge carriers leave the channel.
As described in Rothberg, an ISFET may be fabricated with a floating gate structure, formed by coupling a polysilicon gate to multiple metal layers disposed within one or more additional oxide layers disposed above the gate oxide. The floating gate structure is so named because it is electrically isolated from other conductors associated with the ISFET; namely, it is sandwiched between the gate oxide and a passivation layer that is disposed over a metal layer (e.g., top metal layer) of the floating gage.
As further described in Rothberg, the ISFET passivation layer constitutes an ion-sensitive membrane that gives rise to the ion-sensitivity of the device. The presence of analytes such as ions in an analyte solution (i.e., a solution containing analytes (including ions) of interest or being tested for the presence of analytes of interest), in contact with the passivation layer, particularly in a sensitive area that may lie above the floating gate structure, alters the electrical characteristics of the ISFET so as to modulate a current flowing through the channel between the source and the drain of the ISFET. The passivation layer may comprise any one of a variety of different materials to facilitate sensitivity to particular ions; for example, passivation layers comprising silicon nitride or silicon oxynitride, as well as metal oxides such as silicon, aluminum or tantalum oxides, generally provide sensitivity to hydrogen ion concentration (pH) in an analyte solution, whereas passivation layers comprising polyvinyl chloride containing valinomycin provide sensitivity to potassium ion concentration in an analyte solution. Materials suitable for passivation layers and sensitive to other ions such as sodium, silver, iron, bromine, iodine, calcium, and nitrate, for example, are known, and passivation layers may comprise various materials (e.g., metal oxides, metal nitrides, metal oxynitrides). Regarding the chemical reactions at the analyte solution/passivation layer interface, the surface of a given material employed for the passivation layer of the ISFET may include chemical groups that may donate protons to or accept protons from the analyte solution, leaving at any given time negatively charged, positively charged, and neutral sites on the surface of the passivation layer at the interface with the analyte solution.
With respect to ion sensitivity, an electric potential difference, commonly referred to as a “surface potential,” arises at the solid/liquid interface of the passivation layer and the analyte solution as a function of the ion concentration in the sensitive area due to a chemical reaction (e.g., usually involving the dissociation of oxide surface groups by the ions in the analyte solution in proximity to the sensitive area). This surface potential in turn affects the threshold voltage of the ISFET; thus, it is the threshold voltage of the ISFET that varies with changes in ion concentration in the analyte solution in proximity to the sensitive area. As described in Rothberg, since the threshold voltage VTH of the ISFET is sensitive to ion concentration, the source voltage VS provides a signal that is directly related to the ion concentration in the analyte solution in proximity to the sensitive area of the ISFET.
Arrays of chemically-sensitive FETs (“chemFETs”), or more specifically ISFETs, may be used for monitoring reactions—including, for example, nucleic acid (e.g., DNA) sequencing reactions, based on monitoring analytes present, generated or used during a reaction. More generally, arrays including large arrays of chemFETs may be employed to detect and measure static and/or dynamic amounts or concentrations of a variety of analytes (e.g., hydrogen ions, other ions, non-ionic molecules or compounds, etc.) in a variety of chemical and/or biological processes (e.g., biological or chemical reactions, cell or tissue cultures or monitoring, neural activity, nucleic acid sequencing, etc.) in which valuable information may be obtained based on such analyte measurements. Such chemFET arrays may be employed in methods that detect analytes and/or methods that monitor biological or chemical processes via changes in charge at the chemFET surface. Such use of ChemFET (or ISFET) arrays involves detection of analytes in solution and/or detection of change in charge bound to the chemFET surface (e.g. ISFET passivation layer).
Research concerning ISFET array fabrication is reported in the publications “A large transistor-based sensor array chip for direct extracellular imaging,” M. J. Milgrew, M. O. Riehle, and D. R. S. Cumming, Sensors and Actuators, B: Chemical, 111-112, (2005), pp. 347-353, and “The development of scalable sensor arrays using standard CMOS technology,” M. J. Milgrew, P. A. Hammond, and D. R. S. Cumming, Sensors and Actuators, B: Chemical, 103, (2004), pp. 37-42, which publications are incorporated herein by reference and collectively referred to hereafter as “Milgrew et al.” Descriptions of fabricating and using ChemFET or ISFET arrays for chemical detection, including detection of ions in connection with DNA sequencing, are contained in Rothberg. More specifically, Rothberg describes using a chemFET array (in particular ISFETs) for sequencing a nucleic acid involving incorporating known nucleotides into a plurality of identical nucleic acids in a reaction chamber in contact with or capacitively coupled to chemFET, wherein the nucleic acids are bound to a single bead in the reaction chamber, and detecting a signal at the chemFET, wherein detection of the signal indicates release of one or more hydrogen ions resulting from incorporation of the known nucleotide triphosphate into the synthesized nucleic acid.
However, traditionally, ion concentration in the analyte solution is measured by measuring an instantaneous voltage at an output of the ISFET. The signal-to-noise ratio provided by the instantaneous voltage may not be as high as desired in a lot of situations. Further, with the scaling of ISFET sensor array designs, more ISFET sensors are packed on a chip. Thus, there is a need in the art to provide a better SNR than the instantaneous voltage measurement and also a need for on-chip data compression.
Moreover, with the scaling of ISFET sensor array designs, more and more ISFET sensors are packed on a chip. Thus, there is a need in the art to provide a readout scheme to output measured data from a chip at a high speed.
A floating gate (FG) transistor may be used to detect ions in close proximity to the gate electrode. The transistor may be configured with other transistors to form a pixel that can be placed into an array for addressable readout. In the simplest form, the ancillary transistors are used solely to isolate and select the floating gate transistor for readout in an array. The floating gate transistor may be a chemically-sensitive transistor, and more specifically, a chemically-sensitive field effect transistor (ChemFET). The ChemFET may be designed with a metal-oxide-semiconductor field-effect transistor (MOSFET) containing self-aligned source and drain implants fabricated using standard complementary metal-oxide-semiconductor (CMOS) processing. The ChemFET may be an ion sensitive FET (ISFET), and may be a PMOS or an NMOS device.
To reduce the pixel size to the smallest dimensions and simplest form of operation, the ancillary transistors may be eliminated to form an ion sensitive field-effect transistor (ISFET) using one transistor. This one-transistor, or 1T, pixel can provide gain by converting the drain current to voltage in the column. Parasitic overlap capacitance between terminals of the transistor limits the gain. The capacitance ratios also allow consistent pixel-to-pixel gain matching and relatively constant current operation which justifies the use of a row selection line which can sink the necessary current without causing unacceptable variation. Derivatives of this allow for increased programmable gain through a cascoded transistor enabled during readout. Configurable pixels can be created to allow both common source read out as well as source follower read out.
In a one-transistor pixel array, such as the one shown in
In one embodiment, the 1T ion pixel 100 may work by boot-strapping the row selection line R to the floating gate G while at the same time providing a source of current for the column line bias. In the simplest form, this bootstrapping occurs without adding any extra capacitors. The gate to drain overlap capacitance Cgd, as shown in
The operation of this pixel depends on the fact that the signal range of any given pixel is small compared to the supply voltage or read range of the source follower. For example, the useful signal range may be only 100 mV and the supply voltage may be 3.3V. When a row is selected, the R line is driven to an active high voltage VH, while all other row lines are held at an active low voltage VL. The voltage VL is selected to be approximately equal to the nominal voltage on the column line C during the readout of any given pixel. Because the signal range is small, this voltage is known to within 100 mV in this example. Therefore, the drain to source voltage of all inactive pixels is always held to small values. This point is only critical if the gate to source voltage of inactive pixels is near the threshold of the device. For the row driven to VH, the FG voltages for that row are significantly higher than the other rows because of the bootstrapping that occurs when the row line transitions to VH. After the column line switch Sb is open, the final value on the column line will be determined almost entirely by the active row because the circuit operates according to the winner take-all mode.
There are two sources of current from other rows that can distort the signal value (one that adds current and one that takes away current) and there must be enough bootstrapping available to successfully read pixels without significant interaction from the other rows that produce these sources. The analysis to determine how much bootstrapping is needed is as follows. By the time the pixel is sampled, the device has entered the subthreshold region of operation which has a transconductance slope, for example, of approximately 100 mV/decade. This means that for every 100 mV of change in gate voltage, the current changes by 10 times. In order to effectively read a single pixel, a criteria is set so that 99% of the current on the column line is attributable to the active row and only 1% is attributable to the inactive rows (distortion current). From here it can be determined how much bootstrapping is necessary. With only 2 rows in the pixel array, a 200 mV difference in the floating gate voltages is needed according to the subthreshold slope. Since a signal range of about 100 mV is also needed to be accounted for, the total requirement is about 300 mV. If there are 10 rows, there may be 10 times more contribution from inactive rows. Therefore an extra 100 mV is needed. If the array is increased to 100 rows, another 100 mV is needed. If the array is increased to 10̂n rows, 300+100*n mV is needed. As an example, a 10000 (10̂4) row pixel array only requires a total of 700 mV (300+100*4) of bootstrapping. This amount of bootstrapping can be achieved from the overlap capacitance of the gate and drain. If more capacitance is needed, extra coupling can be facilitated in the mask layout. The above analysis only applies to pixels contributing to the readout current.
Pixels can also take current away from the column line and sink it through the deactivated row lines. Since the deactivated row line is set to approximately the level of the column line, this current draw will be minimal but it must still be quantified and controlled. To accomplish this, the final current on the column line should not be allowed to diminish beyond a certain level. This is ensured by loading the column with a small current sink such as 1 uA. For a W/L (width to length) ratio of 1, a transistor biased at its threshold will have a saturation current of about 0.1 uA. This current decreases by a factor of 10 for every 100 mV of reduction in gate to source voltage. If less than 1% contribution of current is required, the VGS of inactive pixels needs to be kept to 100+100*n mV below the threshold voltage where 10̂n is the number of pixels in the row. Thus, for a 10000 row pixel array, VGS needs to be kept to 500 mV below threshold. A typical 3.3V NMOS transistor has a VT of 600 mV. Therefore, VGS should be less than 100 mV for inactive pixels. Assuming that the FG has a nominal voltage of 0V when the row (R) and column (C) lines are at 0V, this condition is met even as R and C couple to the FG. If the FG has a larger nominal voltage than 0V (for example, due to the trapped charge), more bootstrapping is necessary to cause the column line to reach a level within 100 mV of the FG. As long as the nominal FG voltage is sufficiently low, the second criteria for minimizing distortion current is not a limiting factor. Finally, enough bootstrapping is needed to produce a current on the column line that matches the bleeding current so that the pixel can produce a measurable voltage on the column line. If VG is nominally 0v, then 700 mV is needed for bootstrapping. Therefore, for an NMOS with VT as large as 600 mV, the amount of bootstrapping required is simply limited by the VT. In order to readout the pixel with margin, a good target for bootstrapping is 1V. This leaves 300 mV of range for variation. Achieving 1V of bootstrapping is practical within a 3.3V supply.
All the current from the column readout is distributed through the row line. This causes significant droop in the voltage of the row line if the column current is also significant. The voltage droop affects the bootstrapping level but is not detrimental to the readout of the source follower because variation in drain voltage has only a second order effect. Since pixels are read out with multiple samples, offsets are canceled such that the droop does not affect the sensitivity of the pixels.
It should be noted that the same layout can be used for both source follower readout and common source readout as long as optimizations are not made for either. Only accommodations that need to be made are in the column circuits. This makes for a flexible readout architecture and either readout method may be used depending on the necessary signal range. If the signal needs a high gain, the common source mode should be used. Otherwise, the source follower mode may be used.
The following analysis is given for the gain of the source follower readout. Referring to
In one embodiment, the present invention obtains voltage gain by reading out with the common source configuration. It is desirable to achieve both a reduction in pixel size as well as an increase in signal level. The present invention eliminates the ancillary transistors in other pixel designs (e.g., 2T and 3T discussed below) and uses the source of the ISFET as the selection line to achieve both of these goals. The common source mode is a gain mode and a current mode.
The schematic of an array of pixels with column readout switches according to one embodiment of the present invention is shown in
The pixel array can be loaded with a current source with finite output resistance or another load device such as a resistor. Normally the row selection lines will be held at an active high voltage VH. When a row is selected for readout, its row selection line is pulled low to VL. The value of VL is set such that the nominal current level is about 1 uA. If the FG has a value of 100 mV higher than the norm, 10 times this current will result on the column line. If the value of FG is 100 mV lower than the norm, the current will be 10 times lower. The settling time of the signal on the column line will be signal dependent. The voltage gain is achieved with the selection of the value of R and it can be configurable to achieve programmable gain. For example, if R is 100 k ohms, then the 100 mV, translates to 1V at the output.
The actual circuit is more complicated than just a simple common source amplifier because of the parasitic capacitance involved. Since the FG node is not driven, but rather capacitively coupled to the output, there is a feedback mechanism that limits the gain. This limit is roughly equal to the total capacitance at the FG node to the gate to drain capacitance. This ratio may be about 3. It could be designed to achieve higher gain such as 10 times with careful mask operations to reduce source and drain extensions.
The overlap capacitance created by the LDD regions can be reduced by skipping the LDD implants at the drain for the device.
In the 1T pixel shown in
The pixel in common source readout configuration is shown in
Since A is large compared to the loop gain, the negative input terminal may be considered as a virtual ground node and the gain of the circuit may be determined as Vo/Vi=−Cc/Cgd. Since this ratio is known from the analysis or measured values of the source follower configuration, the gain may be determined to be about 6.5. However compared to the source follower, the gain is Vo/Vi=2/(Asf-G). In this example, a gain of 10 is realized over the source follower configuration. A lower bound on this gain is given by assuming that the input capacitance of the source follower is solely due to Cgd and that the Asf is equal to 1. In this case the gain is about 3. Since neither of these conditions is realistic, the gain is expected to always exceed this number. Thus, if the gain of the source follower configuration of a pixel is known, the gain of the common source configuration of this pixel is also known. In addition, the higher the gain, the more sensitive the pixel is. This makes the common source configuration preferable.
Flicker noise can be reduced by using a channel doping of the same type as the minority carrier. For example, an NMOS with a n-type implant produces a buried channel transistor. To shift the workfunction of the device, a P+ gate electrode can be used.
One derivative of the one-transistor pixel allows for increased programmable gain through a cascoded transistor enabled during readout.
Since the gain of the common source readout is limited by the Cgd capacitance, as shown in
Higher gain and variable gain may be produced in the 1T configuration by bringing the cascode device outside the pixel to the column line.
In this case, the cascode forces the drain of the pixel to remain at a fairly steady voltage over the range of inputs. This causes the pixel to push nearly all of the change in current through the cascode device at the base of the array and into the current load. This reduces the negative feedback from Cds, which would otherwise limit the gain. Given that the current load has infinite output resistance and there is effectively no coupling capacitor to the FG node, the gain of the pixel is now −(gmlrO1+1)gm2rO2, wherein gml is the transconductance of the cascode device at the base of the column line and gm2 is the transconductance of the pixel and rO1 and rO2 are the small signal output resistances as seen at the drain. The value of the output resistance is determined by channel length modulation. Longer gate lengths produce higher output resistance because the effect of channel length modulation is minimized. Since this gain is so large, it can be limited and configured by variation of the current source output resistance, which is shown as Radj in
Various layout choices can be made to implement a 1T and 2T transistor. In order to reduce the size of the pixel the source and drains of adjacent pixels can be shared. In this way a single row selection line enables 2 rows at a time. This reduces the row wiring: two columns are then read out at once for a given column pitch. Such a scheme is shown in
In one embodiment, the cascoded device is gain-enhanced with a differential amplifier in feedback to control a transistor that maintains a constant voltage on the column line.
In a pixel array, a row selection device may be used for selection and isolation.
When a row selection line is activated, the row selection device (a MOSFET) forms a channel due to the gate voltage exceeding a threshold voltage and acts like a switch. When the row selection is deactivated, the channel is diminished. It is important to note that a row selection device never really completely turns “on” or “off”. It only approximates a switch. When the gate is substantially lower than the source of the row selection transistor, good isolation is achieved and the pixel with the active row selection can be read effectively without input from deactivated pixels. With many rows in an array of pixels, it is necessary to achieve a given level of isolation for each row selection device. That is, the requirements for the row selection device depend on the number of rows.
Both ISFET 1101 and the row selection device 1102 are shown as NMOS, but other types of transistors may be used as well. The 2T pixel 1100 is configured as the source follower readout mode, although 2T pixels may be configured as the common source readout mode.
The right column, including a pixel consisting of 1405RS and 1405IS, a pixel consisting of 1406RS and 1406IS, a pixel consisting of 1407RS and 1407IS, and a pixel consisting of 1408RS and 1408IS, is coupled to column traces cb, ct, and cb in substantially the same manner as described above.
The pixel array 1400 has high density because of continuous diffusion, shared contacts, mirrored pixels, and one ct (column top) line and 2 cb (column bottom) line per physical column. A global bulk contact may be implemented by using a P+ wafer with P− epitaxy region.
The arrangement of pixel array 1400 provides for high speed operation. Row lines rs and rs are selected together and readout through cb and cb. This leads to a 4 times faster readout due to twice the number of pixels enabled for a single readout and half the parasitic load of a continuous array, allowing each column to settle twice as fast. In an embodiment, the full array is separated into a top half and a bottom half. This leads to another 4 times faster readout time due to twice the number of pixels readout at a time (both out the top and the bottom) and half the parasitic load of a continuous array. Thus, the total increase in speed over a single row selected continuous array is 16 times.
In an embodiment, both top and bottom halves of the pixel array may be enabled at the same time during readout. This can allow a multiplexing of readout between the top half and the bottom half. For example, one half can be doing a “wash” (e.g., flushing out reactants from the wells over the pixel devices) and the other half can be performing the readout. Once the other half is read, the readout for the two halves is switched.
In an embodiment, a 2T pixel design can incorporate two chemically-sensitive transistors (e.g., ISFETs) rather than one chemically-sensitive transistor and one row select device as described with respect to
In one embodiment, a column circuit allows column lines to be swapped to a sampling circuit such that either source-side or drain-side row selection can be made in either source follower mode or common source mode.
One or more charge pumps may be used to amplify the output voltage from a chemically-sensitive pixel that comprises one or more transistors, such as those described above.
At time t0, all switches are off.
At time t1, φ1 switches 1501, 1502, 1503 and 1504 are turned on. The track phase may start. An input voltage Vin, which may be from an ion sensitive pixel, may start to charge capacitors 1507 and 1508.
At time t2, φ1 switches 1501, 1502, 1503 and 1504 are turned off, and capacitors 1507 and 1508 are charged to Vin-Vref1.
At time t3, φ2 switches 1505 and 1506 are turned on, while φ1 switches 1501, 1502, 1503 and 1504 remain off. The boost phase may start. The capacitor 1507 may start to discharge through the capacitor 1508. Since the capacitors are in parallel during the track phase and in series during the boost phase, and the total capacitance is halved during the boost phase while the total charge remains fixed, the voltage over the total capacitance must double, making Vout approximately two times Vin.
A source follower SF may be used to decouple the gain circuit from the following stage.
The charge pump 1500 may provide a two times gain without a noisy amplifier to provide a virtual ground.
At time t0, all switches are off.
At time t1, φ1 switches 1501, 1502, 1503, 1504, 1601 and 1602 are turned on. The track phase may start. An input voltage Vin, which may be from an ion sensitive pixel, may start to charge capacitors 1507, 1508 and 1604.
At time t2, φ1 switches 1501, 1502, 1503, 1504, 1601 and 1602 are turned off, and capacitors 1507, 1508 and 1604 are charged to Vin-Vref1.
At time t3, φ2 switches 1505 and 1603 are turned on, while φ1 switches 1501, 1502, 1503, 1504, 1601 and 1602 remain off. The boost phase may start. The capacitor 1507 may start to discharge through the capacitors 1508 and 1604, and the capacitor 1508 may start to discharge through the capacitor 1604. Since the capacitors are in parallel during the track phase and in series during the boost phase, and the total capacitance is divided by three during the boost phase while the total charge remains fixed, the voltage over the total capacitance must triple, making Vout approximately three times Vin.
Additional series charge pumps can be added to increase the gain further. In a multi-stage charge pump, the capacitor values do not have to be the same size from stage to stage. It can be observed that the total area consumed by capacitors increases with the square of the gain. Although this feature may, in some cases, be undesirable with respect to area usage, power consumption, and throughput, the charge pump can be used without these penalties when the total noise produced by the ion sensitive pixel and associated fluidic noise is larger than the charge pump KT/C noise when a reasonable capacitor size is used.
At time tO, all switches are off.
At time t1, a switch φsp is on, providing an input voltage Vin to the input of the charge pump 1500.
From time t2 to time t5, the charge pump 1500 operates to push the output voltage Vout to 2(Vin-Vref1), as described before with reference to
From time t6 to t7, the switch φfb is on, feeding the output voltage 2(Vin-Vref1). back to the input of the charge pump 1500, and the first cycle ends.
During the second cycle, the charge pump 1500 amplifies the output voltage by 2(2(Vin-Vref1)). The process repeats, with the output being amplified during each cycle.
An ion sensitive MOS electrode is charge coupled to adjacent electrodes to facilitate both confinement and isolation of carriers. Measurements of ion concentration are made by discrete charge packets produced at each pixel and confined by potential barriers and wells. The ion sensitive electrode can act as either a barrier level or as a potential well. Working in the charge domain provides several benefits, including but not limited to: 1) increased signal level and improved signal to noise through the accumulation of multiple charge packets within each pixel, 2) better threshold matching of the MOS sensing and reference structures, 3) reduction in flicker noise, and 4) global-snap shot operation.
A floating electrode is used to detect ions in close proximity to the electrode. The electrode is charge coupled to other electrodes and to other transistors to form a pixel that can be placed into an array for addressable readout. It is possible to obtain gain by accumulating charge into another electrode or onto a floating diffusion (FD) node or directly onto the column line. It is desirable to achieve both a reduction in pixel size as well as increase in signal level. To reduce pixel size, ancillary transistors may be eliminated and a charge storage node with certain activation and deactivation sequences may be used.
The ion sensitive (IS) accumulation pixel contains some of the following concepts:
The basic IS accumulation pixel is shown in
The basic charge accumulation scheme is shown in
In alternative embodiments, the order of electrodes may be switched, and/or the IS electrode may be used as the barrier rather than the well. Transistors may be added to this accumulation line to enable a large array of pixels. The ancillary transistors are used to increase speed. However, it should be noted that no transistors are necessary to enable a full pixel array of the accumulation line. Instead, an array can be partitioned such that no transistors are needed. In an embodiment, the FD nodes are connected to the column line. Before a pixel is read out, the column line is reset to VDD. Then a row is selected by accumulating charge for that row directly onto the column line. After many cycles, the column discharges to a value directly proportional to the ion concentration. Since the capacitance of the column line depends on the total number of rows, the amount of accumulation required, depends on the number of rows. The array can be partitioned into sub arrays to make timing scalable. For example, every 100 rows can contain a local source follower buffer that is then connected to a global array. This hierarchical approach can be used in general with all readout schemes to make massive arrays of pixels with fast readout.
Due to the thermal activity of carriers, charge packets cannot be generated without noise. Each fill and spill operation produces charge error proportional to KTC (thermal noise in the floating diffusion capacitor), where C is equal to Cox times the area of the ion sensitive electrode. During the fill operation charge can flow freely between the source of electrons and the confinement well. However, during the spill operation, the device enters the subthreshold mode and carriers move by diffusion, mainly in only one direction, which results in half of the thermal noise of a resistive channel. The total noise in electrons for each charge packet is therefore sqrt(KTC/2)/q where q represents the charge of one electron in coulombs (1.6×10e−19). The signal in electrons is equal to VC/q. The signal to noise ratio after n cycles is equal to V*sqrt(2nC/KT). Note that the signal to noise ratio improves by the square root of the number of cycles of accumulation. For small signal levels, the amount of accumulation will be limited to the threshold mismatch between the VR reference electrode and the ion sensitive electrode. Since there is a reference electrode in every pixel and the electrodes are charge coupled, the relative threshold mismatch between each pair of electrodes is small. Assuming, this difference is about 1 mV, over 1000 accumulation cycles should be feasible, thereby improving the signal to noise by more than 30 times. By way of example, if the signal is 1 mV and the electrode area is 1 square micron with Cox=5fF/um̂2, the signal to noise ratio after 1000 cycles is 50 to 1. Since the signal level then reaches 1V, it is expected that no other noise source is relevant. For clarity, the dominant noise is simply the charge packet thermal noise which is well known.
Several design permutations are available depending on the desired mode of operation. The CCD channels are surface mode and are built in standard CMOS technology preferably below 0.13 um. Extra implants can be added to avoid surface trapping and other defects. A channel stop and channel can be formed from donor and acceptor impurity implants. The channel can be made of multiple implants to produce a potential profile optimal for the mode of operation.
The sample and hold circuit 2901 may include a switch SH, a switch CAL, a capacitor Csh, and an amplifier Amp. The switch SH's input is coupled to the output of the column buffer 2903, and its output is coupled to a voltage VREF through the switch CAL, the upper part of the capacitor Csh, and the input of the amplifier Amp. The amplifier is biased by a voltage VB2. The output of the amplifier is coupled to a switch 2904 controlled by a signal ColSeln from a column selection shift register. The output of the switch 2904 is buffered by an output buffer 2905 before reaching the output terminal Vout. The output buffer is biased by a voltage VB3.
Several embodiments of the present invention are specifically illustrated and described herein. However, it will be appreciated that modifications and variations of the present invention are covered by the above teachings. In other instances, well-known operations, components and circuits have not been described in detail so as not to obscure the embodiments. It can be appreciated that the specific structural and functional details disclosed herein may be representative and do not necessarily limit the scope of the embodiments. For example, some embodiments are described with an NMOS. A skilled artisan would appreciate that a PMOS may be used as well.
Those skilled in the art may appreciate from the foregoing description that the present invention may be implemented in a variety of forms, and that the various embodiments may be implemented alone or in combination. Therefore, while the embodiments of the present invention have been described in connection with particular examples thereof, the true scope of the embodiments and/or methods of the present invention should not be so limited since other modifications will become apparent to the skilled practitioner upon a study of the drawings, specification, and following claims.
Various embodiments may be implemented using hardware elements, software elements, or a combination of both. Examples of hardware elements may include processors, microprocessors, circuits, circuit elements (e.g., transistors, resistors, capacitors, inductors, and so forth), integrated circuits, application specific integrated circuits (ASIC), programmable logic devices (PLD), digital signal processors (DSP), field programmable gate array (FPGA), logic gates, registers, semiconductor device, chips, microchips, chip sets, and so forth. Examples of software may include software components, programs, applications, computer programs, application programs, system programs, machine programs, operating system software, middleware, firmware, software modules, routines, subroutines, functions, methods, procedures, software interfaces, application program interfaces (API), instruction sets, computing code, computer code, code segments, computer code segments, words, values, symbols, or any combination thereof. Determining whether an embodiment is implemented using hardware elements and/or software elements may vary in accordance with any number of factors, such as desired computational rate, power levels, heat tolerances, processing cycle budget, input data rates, output data rates, memory resources, data bus speeds and other design or performance constraints.
Some embodiments may be implemented, for example, using a computer-readable medium or article which may store an instruction or a set of instructions that, if executed by a machine, may cause the machine to perform a method and/or operations in accordance with the embodiments. Such a machine may include, for example, any suitable processing platform, computing platform, computing device, processing device, computing system, processing system, computer, processor, or the like, and may be implemented using any suitable combination of hardware and/or software. The computer-readable medium or article may include, for example, any suitable type of memory unit, memory device, memory article, memory medium, storage device, storage article, storage medium and/or storage unit, for example, memory, removable or non-removable media, erasable or non-erasable media, writeable or re-writeable media, digital or analog media, hard disk, floppy disk, Compact Disc Read Only Memory (CD-ROM), Compact Disc Recordable (CD-R), Compact Disc Rewriteable (CD-RW), optical disk, magnetic media, magneto-optical media, removable memory cards or disks, various types of Digital Versatile Disc (DVD), a tape, a cassette, or the like. The instructions may include any suitable type of code, such as source code, compiled code, interpreted code, executable code, static code, dynamic code, encrypted code, and the like, implemented using any suitable high-level, low-level, object-oriented, visual, compiled and/or interpreted programming language.