Patent Application
20170336254
The present invention relates to a photodetector, and particularly, to a photodetector used in an analyzing devices, such as a spectrophotometer.
Photodetectors are used in analyzing devices in a wide range of fields, such as a spectrophotometer or liquid chromatograph. In an absorbance measurement of a sample using a spectrophotometer, for example, a beam of white light emitted from a deuterium lamp (D2 light source) is cast into a sample cell, and the light which has transmitted through the sample cell is separated into a range of wavelengths by a wavelength-dispersing element (e.g. a diffraction grating) and detected by a photodetector.
In an absorbance measurement, in advance of an actual measurement of a sample, a beam of white light is cast into a sample cell containing only the solvent to be used for dissolving the sample, and the transmitted light exiting from the sample cell is subjected to a measurement to obtain a reference signal. Subsequently, with the sample solution contained in the sample cell, the measurement of the transmitted light is similarly performed to obtain a sample signal. Based on the difference between the sample signal and the reference signal, the amount of absorption by the sample is calculated, and the sample concentration is determined.
One of the conventionally and commonly used photodetectors is a photodiode detector (PD detector). As shown in
Non Patent Literature 1: Tanaka, Makino, “Linear image sensor with high performance and large photosensitive element”, Sensors and Actuators A, 29, 201-207, 1991
Non Patent Literature 2: Hirofumi SUMI, Tadakuni NARABU and Shinichiro SAITO, “For the Better Image Quality of CMOS Image Sensor”, Fundamentals Review, Vol. 3, No. 3, pp. 44-51, January 2010
In the absorbance measurement, the amount of light incident on the detector may change by approximately 1000 times depending on the wavelength, due to such factors as the wavelength-intensity distribution of the light source used as well as the absorption characteristics of the sample, solvent and sample cell. Accordingly, a photodetector used in an absorbance measurement should be configured with a sufficiently wide measurable range so that it can detect a wide range of intensities of light and thereby enable accurate determination of the sample concentration. However, conventionally used PD detectors have the following problem if the measurable range is widened to increase the upper limit of the amount of incident light.
In the PD detector, the PD 101 converts incident light into electrons by photoelectric conversion and provides those electrons as photocurrent. In the detection circuit including the capacitor 102, the photocurrent is accumulated for each predetermined length of sampling period. The accumulated photocurrent is converted into voltage and sent to an external system.
The voltage value V obtained by converting the photocurrent into voltage in the detection circuit is expressed by equation (1):
V=(1/C)IΔt=(1/C)M·e (1)
where I is the photocurrent, C is the integral capacitance of the capacitor 102, Δt is the length of the sampling period, M is the number of photoelectrons generated in the PD 101, and e is the elementary charge (=1.602×10−19 C). M may also be expressed as the product of the number of photons incident on the PD 101 and the quantum efficiency η of the PD 101.
The kinds of noise in the detector are roughly divided into an optical shot noise which statistically occurs depending on the amount of incident light, and an electrical noise which occurs in the electrical circuits of the detector (for example, see Non-Patent Literature 2). For ease of explanation, a shot noise which results from dark current is not considered in the present description. The magnitude of the optical shot noise is expressed by the square root of the amount of incident light, whereas the magnitude of the electrical noise is independent of the amount of incident light. The electrical noise expressed as a voltage value Vn can be converted into the number of electrons Mn by equations (2) and (3), where Qn in equation (2) is the amount of noise charges:
Q
n
=M
n
·e=C·V
n (2)
M
n
=C·V
n
/e (3)
The signal-to-noise (S/N) ratio of the detector can be expressed by equation (4) using the number of incident electrons M and the magnitude of the electrical noise converted into the number of electrons Mn. The first term in the root symbol in the denominator of equation (4) corresponds to the optical shot noise. The second term corresponds to the electrical noise.
As can be understood from equation (3), even if the magnitude of the electrical noise Vn is unchanged, if the integral capacitance C of the capacitor 102 is increased in order to widen the measurable range, the number of noise electrons Mn increases in proportion to the integral capacitance. As a result, the value of the second term in the root symbol in the denominator of equation (4) increases, which lowers the S/N ratio. This effect becomes particularly noticeable as the amount of incident light becomes smaller, i.e. as the value of the number of incident electrons M in equation (4) becomes smaller.
A specific example which illustrates how the S/N ratio changes depending on the amount of incident light is hereinafter described. In the present example, it is assumed that the maximum signal voltage of the photodetector is 10 V, the electrical noise is 100 μV, and the upper limit of the measurement range is 109 (number of electrons). In this case, the integral capacitance C of the capacitor 102 is given by:
Since the electrical noise is 100 μV,
Q
n
=C·V
n=16·10−12·100·10−6=16·10−16 (6)
and therefore
M
n
=Q
n
/e=16·10−16/1.6·10−19=104=10000 (7)
In the calculation of the S/N ratio based on equation (4), if M equals the upper limit of the measurement range, i.e. if M=109, then Mn2=108, so that M>Mn2. By comparison, if M=107 or smaller, Mn2=108 is larger than M, which means that the electrical noise constitutes the major cause of the entire noise of the detector. For example, when M=106, the S/N ratio is 99.5, which is by more than one order of magnitude lower than the S/N ratio for which only the optical shot noise is considered (S/N=1000).
The problem to be solved by the present invention is to provide a photodetector which ensures a high S/N ratio regardless of the amount of incident light while allowing the measurable range to be widened.
The photodetector according to the present invention developed for solving the previously described problem includes:
a) a plurality of photoelectric conversion elements arranged within a photosensitive-element area constituting one photosensitive element;
b) a plurality of detection circuits each of which is provided for one of the plurality of photoelectric conversion elements, each of the detection circuits including a capacitor; and
c) a signal processing section for totaling output signals produced by the plurality of detection circuits.
For example, the photoelectric conversion elements are photodiodes.
For example, the signal processing section may include: an analog-to-digital converter for converting a signal produced by each of the plurality of detection circuits into a digital signal; a digital memory for storing digital signals produced by the analog-to-digital converter; a multiplexer for sequentially reading digital signals from the digital memory; and a digital computing unit for totaling the digital signals sequentially read by the multiplexer.
The photodetector according to the present invention may have a plurality of photosensitive elements. In a photodetector having a plurality of photosensitive elements, each of the photosensitive elements may be constructed in the previously described manner.
In a conventional photodetector, each photosensitive element has a single photoelectric conversion element and a single detection circuit for processing output signals of the photoelectric conversion element. In the photodetector according to the present invention, each photosensitive element has a plurality of combinations of photoelectric conversion elements and detection circuits, with each combination having one photoelectric conversion element and one detection circuit. As noted earlier, conventional photodetectors have the problem that increasing the integral capacitance of the capacitor in the detection circuit in order to widen the dynamic range causes an increase in the number of noise electrons and lowers the S/N ratio. By comparison, in the case of the photodetector according to the present invention, the dynamic range can be widened by simply increasing the number of combinations of the photoelectric conversion elements and detection circuits; there is no need to increase the integral capacitance of the capacitor in the detection circuit. Accordingly, it is possible to ensure a high S/N ratio while widening the measurable range.
The signal transport from the photoelectric conversion element to the capacitor in the detection circuit is carried out by an electric current. Therefore, if the wiring distance between these two elements is long, a high parasitic capacitance occurs. Similarly to the integral capacitance of the capacitor, this parasitic capacitance also constitutes a factor which increases the number of noise electrons Mn expressed by equation (3).
Accordingly, in the photodetector according to the present invention, at least the capacitor in the detection circuit should preferably be located within the photosensitive-element area. This arrangement reduces the wiring distance from the photoelectric conversion element to the capacitor and minimizes an increase in the amount of noise electrons.
The photodetector according to the present invention may also be configured so that the signal processing section totals output signals produced by the plurality of detection circuits for each of a plurality of sub-periods set by dividing the sampling period, and provides the total signal at intervals of the sampling period.
For example, in the case where K photoelectric conversion elements are provided within one photosensitive-element area (where K is an integer equal to or greater than two) and the sampling period is divided into L segments of time (where L is an integer equal to or greater than two), the dynamic range can be widened by K×L times, without increasing the integral capacitance, by reading an output signal in each segment of time, totaling the read output signals, and providing the total signal as the output signal for one photosensitive element and one sampling period.
With the photodetector according to the present invention, it is possible to ensure a high S/N ratio regardless of the amount of incident light while allowing the measurable range to be widened.
Embodiments of the photodetector according to the present invention are hereinafter described with reference to the drawings. The photodetector according to any of the present embodiments is a photodiode (PD) detector having a plurality of photosensitive elements and is intended to be used as a detector in an analyzing device.
The S/N ratio of the photodetector in the present embodiment under specific measurement conditions is hereinafter described.
For the present embodiment, the following measurement conditions are assumed: The accumulation time (which corresponds to the sampling period) is 10 ms. The maximum number of photons incident on the photosensitive-element area 11 during the accumulation time is 2×109. The quantum efficiency of the photodiode is η=0.5. The maximum value of the voltage in the detection circuit 14 (the voltage value produced upon incidence of the maximum number of photons) is 10 V. The electrical noise is 100 μV.
As already explained using equation (5), in the conventional PD detector, a capacitor having an integral capacitance of C=16 pF is used in order to be compatible with the aforementioned measurement conditions. In the present embodiment, since incident light is distributed over 100 photodiodes arrayed within the photosensitive-element area 11, the maximum number of electrons generated within each photosensitive sub-element area 12 is 107. Accordingly, the integral capacitance C of each capacitor 102 can be decreased to 0.16 pF.
The S/N ratio in the present embodiment can be expressed by equation (8):
where K is the number of photosensitive sub-element areas 12 in the photosensitive-element area 11, which is 100 in the present configuration example. Parameter m is the number of electrons generated in the photodiode placed in the sub-element area. When the number of photons M generated in the photosensitive-element area 11 is 106, the S/N ratio calculated by equation (8) is 707, which is much higher than the S/N ratio of the conventional PD detector, i.e. 99.5. The term K·mn2 representing the electrical noise in equation (8) has a value of 106.
If the photodiode is in the form of a package and mounted on a printed circuit board along with the integrator circuit having an analog IC and capacitor, the number of noise electrons increases due to the parasitic capacitance of the package of the photodiode and/or analog IC as well as the parasitic capacitance due to the wiring on the printed circuit. Since these parasitic capacitances are also on the order of pF, it is practically difficult to construct an integration amplifier having a high gain as obtained by the previously described calculation. Accordingly, at least the portion of the integration amplifier in the detection circuit should preferably be integrally formed on the same semiconductor integrated circuit board as the photodiode.
A device having such a configuration and manufactured by a CMOS process which is capable of micro-wiring is hereinafter described as the second embodiment.
In general, CMOS devices are operated at a low supply voltage. Therefore, in many cases, it is impossible to handle a high voltage of 10 V within the device, as in the first embodiment. Accordingly, the following description deals with an example in which the upper limit of the voltage that can be handled within the device is 1 V, while the electrical noise is the same as in the previous case, i.e. 100 μV.
The elements arranged within the photosensitive sub-element area 22 are as follows: a photodiode 31; a transfer transistor 32, provided at the end of a signal-reading line connected to the photodiode 31, for transferring photocharges; a floating diffusion 33, connected to the photodiode 31 via the transfer transistor 32, for temporarily accumulating photocharges and converting those charges into a voltage signal; a reset transistor 34 for removing electric charges accumulated in the floating diffusion 33; and a source follower amplifier 35 formed by two cascade-arranged transistors 351 and 352, for providing, as a voltage signal, the electric charges accumulated in the floating diffusion 33. Power sources VDD are connected to the reset transistor 34 and the transistor 351, respectively. A constant voltage (bias voltage) VBIAS is connected to the gate of the transistor 352.
Drive lines (not shown) for supplying control signals φT and φR are connected to the gate of the transfer transistor 32 and that of the reset transistor 34, respectively. Based on the control signals supplied through these drive lines, the intensity of the light incident on the photodiode 31 is detected at predetermined intervals of sampling time.
The S/N ratio in the configuration of the second embodiment is hereinafter calculated under the same conditions as in the first embodiment except that the upper limit of the voltage is 1 V.
In the second embodiment, there are 10,000 photosensitive sub-element areas 22 within the photosensitive-element area 21, with one photodiode 31 placed in each photosensitive sub-element area. Accordingly, the upper limit of the number of electrons generated in each photodiode 31 is 105. Since the upper limit of the voltage is 1 V, the integral capacitance C of the used capacitor calculated by equation (1) is 16 fF. From equation (3), the number of noise electrons Mn corresponding to an electrical noise of 100 μV is 10. Then, the term representing the electrical noise in equation (4) is K·Mn2=106. This value is the same as in the first embodiment. This result demonstrates that an S/N ratio which is as high as in the first embodiment can be ensured even in the case where high voltages cannot be handled, as in a device manufactured by a CMOS process.
In the second embodiment, since the detection circuit is located within the photosensitive sub-element area 22 as shown in
In the case where the integral capacitance is as low as in the second embodiment, the influence of the reset noise (or “thermal noise”) also needs to be considered. In the detection circuit, before the number of photoelectrons is converted into voltage through the integral capacitance, the integral capacitance needs to be reset to the reference voltage for each sampling period. In this operation, a reset noise occurs due to the fluctuation of the reset voltage. The relationship between the reset noise and the integral capacitance is expressed by equation (9):
V
n−reset=√{square root over (kT/C)} (9)
Equation (9) shows that the lower the integral capacitance C is, the higher the reset noise becomes. For example, when the integral capacitance is C=16 fF and the temperature is T=300 K, the reset noise is 509 μV, which is more than five times as high as the electrical noise.
In the second embodiment, in order to eliminate the influence of the reset noise, a correlated double sampling circuit (CDS circuit) 24 as shown in
The output signals from the CDS circuits 24 are converted into digital signals by the A/D converters 25. After being stored in the digital memories 26, those signals are sequentially read by the multiplexer 27 and sent to the digital computing unit 29 through the output port of the high-speed digital output circuit 28. The digital computing unit 29 totals the output signals from all photosensitive sub-element areas 22 and provides the total signal as the output signal for one photosensitive element.
The digital computing unit 49 performs appropriate operations according to the purpose of the analysis, such as averaging the signals detected by the detection circuits corresponding to the photosensitive-element areas 41 or calculating a difference between those signals. The digital computing unit 49 may also be configured to calculate a difference in the signal continuously produced from the same photosensitive element over a period of time based on analysis conditions specified by an analysis operator on a separate control device. A specific configuration is as follows: An output signal corresponding to only the background light is initially acquired in the first accumulation period. Subsequently, in each accumulation period, an output signal corresponding to the sum of the background light and measurement light is acquired, and the former output signal is subtracted from the latter. The result of this calculation is sent to an external system as the output signal corresponding to the measurement light.
In this PD detector 40, the predetermined interval of sampling time at which the digital computing unit 49 provides an output signal obtained by totaling the signals from the photosensitive elements is divided (i.e. frequency-divided) into sub-periods (e.g. 1000 sub-periods), while the selection gates 36 in the 1000 photosensitive sub-element areas 42 are sequentially supplied with control signal φX (a signal which indicates the sub-period) through their respective gate terminals. The photocurrents produced by the 1000 photodiodes 31 belonging to the same photosensitive-element area 41 are sequentially read from one photosensitive sub-element area 42 in each sub-period, and temporarily stored in the digital memories 46. The stored signals are totaled by the digital computing unit 49 at the predetermined intervals of sampling time and provided as the output signal.
In this PD detector 40, since the photosensitive-element area 41 is divided into 1000 photosensitive sub-element areas 42, the integral capacitance C of the capacitor is 160 fF, the electrical noise converted into the number of electrons is 100, and the value of the term K·Mn2 representing the electronical noise in equation (8) is 107 under the same conditions as in the second embodiment, i.e. the maximum number of photons incident on the photosensitive-element area 41 is 2×109, the quantum efficiency of the photodiode 31 is η=0.5, the maximum value of the voltage in the detection circuit (the voltage value produced upon incidence of the maximum number of photons) is 1 V, and the electrical noise is 100 μV. Although the term representing the electronical noise in the third embodiment has a larger value than in the second embodiment, this value can be appropriately decreased by increasing the number of divisions of the photosensitive-element area 41.
In the previously described configuration, since the number of photosensitive sub-element areas 42 arranged within the photosensitive-element area 41 is the same as the number of divisions of the sampling period, each photosensitive element is provided with one signal-reading line 43 and other related elements. It is also possible to adopt a different configuration. For example, the number of divisions of the sampling period can be decreased to one half of the number of photosensitive sub-element areas by dividing the photosensitive sub-element areas 42 arranged within each photosensitive-element area 41 into two groups and providing each group with one signal-reading line 43 and other related elements.
The configurations described in the previous embodiments not only increase the S/N ratio by decreasing the magnitude of the electrical noise but also improve the yield rate of the chip-manufacturing process. In commonly used manufacturing process techniques, the circuit wiring is formed on a semiconductor substrate by mask-pattern projection. In this process, foreign matters (e.g. particles) may adhere to the wafer surface. If this occurs, the desired circuit wiring cannot be formed, and a spot region in which the dark current or leak current specifically increases is formed. If such a spot region is formed in one of the photosensitive elements in a conventional PD array detector, the entire array detector will be treated as a defective product due to that single photosensitive element, since each photosensitive element has only one combination of the PD and the detection circuit inside.
By comparison, in the previously described embodiments, the photosensitive-element area is divided into a plurality of photosensitive sub-element areas, with each photosensitive sub-element area having a PD and detection circuit arranged for it. Even if a spot region has occurred in one of the L divisions of the photosensitive-element area, a correct output signal can still be obtained by converting the L−1 available detection signals into L signals or by performing a computing process for estimating the signal value at the spot region from the detection signals obtained at the photosensitive sub-element areas adjacent to the spot region. Accordingly, the array detector will not be treated as a defective product, so that the yield rate will be improved. Since the presence of a defective photosensitive sub-element area can be detected in the production process, it is possible, for example, to previously configure the digital computing unit 49 in the third embodiment so as to automatically perform the aforementioned computing process. That is to say, for example, the digital computing unit 49 can be configured to disregard a specific photosensitive sub-element area and perform a computing process using only the output signals from the other photosensitive sub-element areas.
Any of the previous embodiments is a mere example and can be appropriately changed without departing from the spirit of the present invention. In the previous embodiments, the A/D converters are provided so as to read signals at high speeds. However, those A/D converters are unnecessary in the case where the voltage signals are processed in analog form. The number of divisions of the photosensitive-element area in each embodiment is also a mere example and may be appropriately changed. The configuration in any of the embodiments is applicable in any type of array detector having a plurality of photosensitive-element areas arranged in a one-dimensional or two-dimensional form. The computing processes performed by the digital computing unit described in the third embodiment are also mere examples; for example, the digital computing unit may be configured to perform an appropriate computing process according to the content of the analysis using an analyzing device which includes a PD detector according to one of the previous embodiments.
In the previous descriptions, for ease of understanding, the equations which do not take into account the influence of the shot noise resulting from the dark current (which is hereinafter called the “dark-current shot noise”) were used to calculate the S/N ratio. A case in which the dark-current shot noise is considered is hereinafter described.
Adding the term Mdark representing the dark-current shot noise to equation (4) for calculating the S/N ratio for a conventional PD detector gives the following equation (10):
This equation does not include the term representing the dark current itself, because the dark current can be removed by using the previously described CDS circuit.
On the other hand, adding the term representing the dark-current shot noise to equation (8) for calculating the S/N ratio of a PD detector according to the present invention gives the following equation (11). Using this equation (11) for a PD detector according to the present invention also yields a higher S/N ratio than that of the conventional PD detector calculated by equation (10).
| Number | Date | Country | Kind |
|---|---|---|---|
| 2014-235966 | Nov 2014 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2015/069693 | 7/8/2015 | WO | 00 |
