WO2014183982A1 - Light transit time sensor - Google Patents

Abstract

The invention relates to a light transit time sensor (22) comprising modulation gates (Gam, Gbm) arranged in light-sensitive areas and readout fingers (Ga, Gb) which are arranged in light-insensitive areas. Said modulation gates (Gam, Gbm, GO) and readout fingers (Ga, Gb) are arranged in parallel strips which form a group of light transit time pixels (23). A light transit time pixel (23) comprises at least two readout fingers (Ga, Gb). The invention is characterised in that said light transit time sensor (23) comprises at least one row of light transit time pixels (23z) having at least three light transit time pixels (23), and that the readout fingers (Ga, Gb), which are not arranged on the edge of the (23Z) rows, are adjacent to a modulation gate (Gam, Gbm) on two sides.

Description

 LIGHT RUNNING TIME SENSOR

The invention relates to a light transit time sensor according to the preamble of the independent claim.

The light transit time sensor relates in particular to light transit time camera systems in particular

Lichtlaufzeit- or 3D TOF camera systems, the runtime information from the

Phase shift of an emitted and received radiation win. As Lichtlaufzeitbzw. 3D TOF cameras are particularly suitable PMD cameras with photonic mixer detectors (PMD), as they u.a. in the applications EP 1 777 747 B1, US Pat. No. 6,587,186 B2 and also DE 197 04 496 C2 and, for example, from the company 'ifm electronic GmbH' or 'PMD Technologies GmbH' as frame grabber 03D or as CamCube are. The PMD camera in particular allows a flexible arrangement of the light source and the detector, which can be arranged both in a housing and separately.

Furthermore, from DE 198 21 974 AI a device for detecting phase and amplitude of electromagnetic waves is known in the light-transmitting modulation gates in

Photosensitive region and charge accumulating accumulation gates are arranged in strip form.

The object of the invention is to improve the fill factor on a light transit time sensor.

The object is achieved in an advantageous manner by the inventive light transit time sensor according to the preamble of the independent claim.

Advantageously, a light transit time sensor is provided, with modulation gates in light-sensitive and read-out fingers in light-insensitive areas, wherein the modulation gates and

Auslesefmger are arranged in parallel strips, which form a group of light-propagation time pixels, and wherein a light transit time pixel has at least two Auslesefmger. The

Embodiment further provides that the light transit time sensor at least one

Has a light transit time pixel line with at least three light transit time pixels, and that the

Read-outs not arranged at the edge of the line, on two sides at one

Modulation gate adjacent.

This procedure has the advantage that it dispenses with an isolation path separating the individual pixels and that the fill factor can be increased.

In a further preferred embodiment, the read-out memories have a readout gate and a readout node, the readout node being formed as a pn-type structure with a metallization. In contrast to structures that formed over the entire finger width as a pn junction are, a structure of gate and readout node spatially smaller form, which can further improve the fill factor.

Preferably, the readout nodes are arranged in an end region of the readout finger in the vicinity of the readout circuit in order to shorten signal paths.

In a further embodiment, the readout nodes are on at least three sides of

Editing gate bordered. This makes it advantageous to avoid crosstalk or penetration of the modulation gates.

In a further embodiment, it is provided to arrange at least three modulation gates between two read-out fingers, wherein the modulation gates, which adjoin the read-out memories both have transparency in an infrared as well as in a visual range. As a result, it is advantageously possible to provide the sensor with additional functionality without impairing the performance of the 3D image acquisition.

The invention will be explained in more detail by means of embodiments with reference to the drawings.

They show schematically:

FIG. 1 shows the basic principle of a time-of-flight camera according to the PMD principle,

 FIG. 2 shows a modulated integration of the runtime-shifted generated charge carriers,

 FIG. 3 shows a cross-section of a PMD pixel,

 FIG. 4 shows a top view of a PMD pixel,

 FIG. 5 shows a top view of a PMD pixel with read-out node,

 FIG. 6 shows a plan view of a PMD pixel with bifurcated read-out fingers,

 7 shows a cross section through a PMD pixel according to the invention with two

 Modulation gates,

 FIG. 8 shows a schematic top view of a PMD pixel according to the invention,

 FIG. 9 is a cross-sectional and plan view of a passing PMD pixel row;

 Figure 10 is a plan view of a passing PMD pixel line with minimum pixel size,

FIG. 11 shows an arrangement of several passing PMD pixel lines,

 FIG. 12 shows an embodiment of a PMD pixel line with RGB filters,

 FIG. 13 shows a PMD pixel row with three modulation gates each,

Figure 14 shows two PMD pixel lines with different phase position. In the following description of the preferred embodiments, like reference characters designate like or similar components.

FIG. 1 shows a measurement situation for an optical distance measurement with a light transit time camera, as is known, for example, from DE 197 04 496 C2.

The light transit time camera system 1 comprises a transmission unit or an illumination module 10 with an illumination light source 12 and associated beam shaping optics 15 and a reception unit or light runtime camera 20 with a reception optics 25 and a light transit time sensor 22. The light transit time sensor 22 has at least one pixel, but preferably one pixel - Array, and is designed in particular as a PMD sensor. The

Receiving optics 25 typically consists of improving the imaging characteristics of a plurality of optical elements. The beam-shaping optical system 15 of the transmitting unit 10 is

preferably formed as a reflector. However, it is also possible to use diffractive elements or combinations of reflective and diffractive elements.

The measurement principle of this arrangement is essentially based on the fact that, based on the phase shift of the emitted and received light, the transit time of the emitted and reflected light can be determined. For this purpose, the light source 12 and the light transit time sensor 22 via a modulator 30 together with a specific

Modulation frequency or modulation signal applied to a first phase position a.

In accordance with the modulation frequency, the light source 12 transmits an amplitude-modulated signal having the phase a. In the case illustrated, this signal or the electromagnetic radiation is reflected by an object 40 and hits due to the distance traveled

Distance corresponding to phase-shifted with a second phase position b on the

Light transit time sensor 22. In the light transit time sensor 22, the signal of the first phase a of the modulator 30 with the received signal having the time-related second phase b, mixed, from the resulting signal, the phase shift or the object distance d is determined.

For a more accurate determination of the second phase position b and thus of the object distance d, it may be provided that the phase position a is operated with the light transit time sensor 22 in order to change preselected phase shifts Δφ. Equally effective, it can also be provided to selectively shift the phase with which the lighting is driven.

The principle of the phase measurement is shown schematically in FIG. The upper curve shows the time course of the modulation signal with the illumination 12 and the Light transit time sensor 22, here without phase shift, are controlled. The light b reflected by the object 40 strikes the phase-shifted light according to its time of flight tL

Light transit time sensor 22. The light transit time sensor 22 collects the photonically generated charges q during the first half of the modulation period in a first read-out path Ga and in the second cycle half in a second read-out path Gb. The charges will be

typically collected or integrated over several modulation periods. From the

Ratio of the charges qa, qb collected in the first and second gates Ga, Gb allows the phase shift and hence the distance of the object to be determined.

As already known from DE 197 04 496 C2, the phase shift of the light reflected from the object and thus the distance, for example, by a so-called IQ (in-phase quadrature) method can be determined. To determine the distance, two measurements are preferably carried out with modulation signals whose phase positions are shifted by 90 °. Thus, for example, (p _mo d + φο and (p _mo d + φo), where the charge difference Aq (0 °), Aq (90 °) determined in these phase positions is the phase shift of the reflected light via the known arctan or arctan2 relationship can be determined.

Ag (90 °)

 φ = arctan Aq (0 °)

To improve the accuracy of further measurements can be performed with shifted by 180 °, for example, phase positions.

Ag (90 °) - Ag (270 °)

 φ = arctan A (0 °) - A (180 °)

Of course, measurements with more than four phases and their multiples and a correspondingly adapted evaluation are conceivable.

Figure 3 shows a cross-section through a pixel of a photonic mixer as it

For example, from DE 197 04 496 C2 is known. The modulation photogates G _am , Go, Gt, _m form the light-sensitive area of a PMD pixel. In accordance with the voltage applied to the modulation gates G _on , Go, Gbm, the photonically generated charges q are directed either to one or the other read-out path G _a , Gb. The Auslesefmger is formed in the example shown as a pn junction or diode.

FIG. 3b shows a potential profile in which the charges q are directed in the direction of the first

Auslesefmgers G _a flow, while the potential according to Figure 3c, the charge q in the direction of the second read finger Gb flows. The potentials are specified according to the applied modulation signals. Depending on the application, the modulation frequencies are preferably in a range of 1 to 100 MHz. At a modulation frequency of, for example, 1 MHz results in a period of one microsecond, so that the modulation potential changes accordingly every 500 nanoseconds.

FIG. 3 a further shows a readout unit 400 which, if appropriate, may already be part of a CMOS PMD light transit time sensor. The photonically generated charges detected in the read-out finger G _a , G _b are distributed over a plurality of

Modulation periods collected. In a known manner, the voltage applied to the read-out fingers Ga, Gb can be tapped to high impedance, for example via the readout unit 400. The integration times are preferably to be selected such that the light transit time sensor or the read-out fingers or their capacitances and / or the light-sensitive areas do not saturate for the expected amount of light.

FIG. 4 shows a top view of a light transit time pixel 23 with typical length specifications. The distance or the length extension between the two readout fingers Ga, Gb is the channel distance L _K and the distance between the longitudinal axes of symmetry of the two

Readout Ga, Gb referred to as fingerpitch L _F p. According to the invention, it is provided that the pixels are not separated from one another by a field oxide, so that a further modulation gate is connected directly to the read-out finger.

The extent of the gates or the Auslesefmger parallel to the channel spacing L _K is the

Gate length L _G and the vertical extent to the gate width B _G. While the gates are essentially the same length in their widths B _G , the gate lengths are typically varied depending on the application. In the illustrated case, for example, the middle modulation gate Go longer than the two outer modulation gates G _am , Gb _m , for example, to influence a modulation contrast.

The Auslesefmgers G _a , Gb can be constructed and structured in different ways. From DE 197 04 496 C2 it is known, for example, to form the read-out memories as pn junctions or diodes. Preferably, the diode extends over the entire width of the Auslesefmger. Preferably, the Auslesefmger are formed opaque or covered with an opaque layer, preferably a metallization. The

If necessary, metallization can simultaneously form the contacting of the diode. FIG. 5 shows a preferred embodiment according to the invention, in which the read-out fingers Ga, Gb have a gate structure and an integration or read-out node. In the illustrated case, a readout gate AG _a , AG _b extends over the entire width B _{G of} the readout Ga, Gb. In one end region of the read-out flag Ga, Gb is a read-out node AK _a , AK _b . As already shown in FIG. 3, a voltage is applied to the readout gate AG _a , AGb, which forms a potential well for the charge carriers q below the gate. The charge carriers collected there flow primarily through diffusion processes to the respective read-out node AK _a , AK _b and can there be tapped by the read-out unit 400, for example as a voltage.

The read-out fingers Ga, Gb are preferably covered with an opaque layer, preferably a metal layer. The metal layer may possibly also for contacting the

Serve as readout node. In a further embodiment, it is also conceivable to influence the readout gate itself in terms of transparency, for example by siliciding.

Furthermore, variations of the size and position of the readout nodes AK _a , AK _{b are} conceivable. For example, the readout nodes AK _a , AK _{b have} the same length as the readout gates AG _a , AG _b . Also, multiple readout nodes may be distributed over the structure of the gate. In particular, readout nodes can be arranged at both ends of the gate. Also, a central readout node is conceivable. The readout nodes are preferably designed as a pn junction or as a diode.

FIG. 6 shows a further variant in which the read-out fingers or the gate structure in the region of the read-out nodes is widened. By this Aufgabelung the readout nodes AK _a , AK _{b are surrounded} by a gate structure and separated with a minimum distance from the adjacent modulation gate yarn, Gbm. By such a procedure, electrical passages from the modulation gate to the read-out nodes are advantageously reduced or inhibited.

As already mentioned, a PMD pixel is typically separated from the neighboring matrix pixel by means of a field oxide and an optical cover. This will

typically minimizes crosstalk between the individual matrix pixels. For very small pixels, however, the fill factor plays an increasingly important role. In particular, it is disadvantageous if the separating field oxide matrix has a larger dimension than the read-out fingers. If the pixel pitch is on the order of a fingerpitch, it plays

Peripheral area or separation area with respect to the filling factor an increasingly important role. The typical finger pitch L _F p, ie read-out finger plus modulation gates, lies in the Order of about 7 - 10 μηι. The height and width of the fingers and the channel length are scalable in a very wide range for PMD structures, similar to transistors.

Through the use of smaller manufacturing processes and the miniaturization of

Readout electronics, it is possible to use the pixel pitch L _PP or the size of a pixel in the

Order of magnitude of the finger pitch L _F p to bring. Furthermore, there is a desire for sensors with the highest possible resolution and small dimensions.

To solve these technical requirements, a continuous PMD pixel structure without separating field oxides is proposed.

Figures 7 and 8 show such a continuous pixel structure.

FIG. 7a shows a cross-section of such a structure with four read-out paths Ga, Gb and two modulation gates yarn, Gbm between two read-out paths Ga, Gb. The

Modulation gates yarn, Gbm are transparent as shown in Figure 3 and form a photosensitive region 26 of the light transit time pixel 23. The Auslesefmger Ga, Gb are covered with a mask and form a light-insensitive region 27. Below the Auslesefmger Ga, Gb doping regions are indicated may be n- or p-doped according to the base material of the semiconductor. As already shown in the above examples, this can also be a gate structure with read-out nodes.

FIG. 7b shows a potential curve for the structure shown in FIG. 7a, analogous to the representation according to FIG. 3c.

FIG. 8 shows a plan view of the pixel structure according to FIG. 7a. The Auslesefmger Ga, Gb are read by a readout unit 400, wherein Auslesefmger same name are performed together on an evaluation unit 420, for example, the charge differences Aq or the charge sums Σq the Auslesefmger Ga, Gb determined. The modulation signal originating from the modulator 30 is fed in counterphase via a potential supply line 35 to the modulation gates Yarn, Gbm. The modulation gates yarn, Gbm in the immediate

Neighborhood of a selection finger Ga, Gb are each at the same potential. The read-out unit 400 and the potential supply line 35, like the readout gates Ga, Gb, are covered with an opaque mask.

The light-propagation time pixel shown in FIGS. 7 and 8 is arranged according to the invention, as shown in FIG. 9, continuously and without separation in a row of pixels. The pixel line shown in FIG. 9 is divided into four individual pixels 23.1 - 23.4. The pixel boundaries are each marked with dashed dotted lines. The read-out flags that are at the edge of the light-time pixel share one photosensitive area 26a with the neighboring pixel. Depending on the potential position, the charge carriers generated there thus flow once to the neighboring pixel and the other time to the own read-out marker. However, since this process repeats on both outer reading outlets with opposite signs, this charge transfer on average recovers, so that effectively, as shown in FIG. 9, only half the common light-sensitive area 26a is active for the respective pixel.

FIG. 10 shows a smallest possible pixel structure with only two read-out paths Ga, Gb and four modulation gates yarn, Gbm, wherein, as in the pixels shown so far, the

Auslesefmger with two equal potential in the same acting modulation gates is surrounded. The pixel pitch Lpp is in this case only slightly larger than the finger pitch Lpp.

FIG. 11 shows an arrangement with a plurality of PMD pixel rows 23z, in which, for a space-saving arrangement, two pixel rows each have a surface area for the pixel array

Share potential leads 35 and readout units 400.

FIG. 12 or FIG. 12a shows a variant with three modulation gates yarn, Gbm, GO, which have different optical filters in order to generate, for example, visual color information.

For example, to detect a color information, a potential according to FIG. 12b can be applied to the pixel structure. By applying color filters Red R, Green G and Blue B over the modulation gates in combination with a suitable control of the gates, it is possible to extract color information from the sensor whose resolution is higher than the physical resolution with respect to the 3D Values.

The modulation gates on the sides of the Auslesefmger be provided on both sides with the same filters, which are permeable to infrared in addition to a transparency in one of the three primary colors. The middle gate is permeable only in the infrared.

In the color or RGB mode, the middle gates GO are preferably driven at 0 volts, while the outer modulation gates Ga, Gb are supplied with a voltage between the readout or separation gates. The charge carriers generated below the color filters are collected at the nearest readout Ga, Gb. The potential barrier applied under the middle modulation gate GO can not be overcome by the charge carriers. For a photon mixture, a potential according to FIG. 12c is preferably applied to the modulation gates, wherein the modulated illumination takes place in the infrared IR.

FIG. 13 shows a configuration in which a higher virtual pixel resolution than the physical pixel resolution can be realized. Typically, the pixels used are square, but this is not mandatory. A pixel consists of at least two read-out gates Ga, Gb and the modulation gates Yarn, Gbm, G0 or MOD. In the example shown here, the actual pixel cell has a format of 2: 1 in the ratio of width to height. By realizing the pixels as passing matrix pixels, it is possible to achieve a virtually higher pixel resolution by calculating distance values from the charge differences of all read-out factors Ga, Gb with the respective neighboring read-out memory Ga, Gb. The resulting virtual pixel pitch Lypp is then 1: 1 in the example shown, while the physical pixel pitch Lpp is 2: 1.

Typically, in a PMD sensor, four frames are recorded in temporal succession with four different phase angles. Although this allows a high local

Resolution, however, requires a low temporal resolution.

FIG. 14 shows a structure in which four different phases can be realized in a narrow spatial area by sharing the readout memory four different phases. Via the modulator 30, the upper and lower pixel line are subjected to a modulation signal, wherein the lower pixel line is applied with a shifted by 90 ° modulation signal. By suitable interconnection adjacent pixels in a row of pixels can be acted upon by a shifted by 180 ° phase position. By this procedure, a distance value can be determined in a frame by offsetting two times two neighboring pixels. This makes it possible to significantly minimize the motion artifacts without greatly affecting the spatial resolution.

LIST OF REFERENCE NUMBERS

 10 transmission unit

 12 illumination light source

 15 beam shaping optics

 20 receiving unit, TOF camera

 22 light transit time sensor

 23 light-time pixels

 23Z light time pixel line

 25 receiving optics

 26 photosensitive area

 26a common photosensitive area

27 light-insensitive area

 28a active range for readout finger Ga

28b active range for read finger Gb

30 modulator

 35 potential supply

 40 object

 400 readout unit

 420 evaluation unit

 Yarn, G0, Gbm, MOD modulation gates

Ga, Gb elite finger

AG _a , AG _b Readout gates

AK _a , AK _b Selection node

q charges

qa, qb charges on the readout finger Ga, Gb

Claims

1. Light transit time sensor (22) with modulation gates (Yarn, Gbm) in light-sensitive and Auslesefmger (Ga, Gb) in light-insensitive areas, wherein the modulation gates (Yarn, Gbm, G0) and Auslesefmger (Ga, Gb) are arranged in parallel strips, the in groups form a light transit time pixel (23),  wherein a light-propagation time pixel (23) has at least two read-out paths (Ga, Gb), characterized  in that the light transit time sensor (23) has at least one light transit time pixel line (23z) with at least three light transit time pixels (23),  and that the read-out flags (Ga, Gb) which are not arranged at the edge of the (23Z) row are adjacent to a modulation gate (yarn, Gbm) on two sides. 2. light transit time sensor (22) according to claim 1, wherein the Auslesefmger (Ga, Gb) a Readout gate (AG _a , AGBb) and a readout node (AKa, AKb), wherein the readout node is formed as a pn structure with a metallization. 3. light transit time sensor (22) according to claim 2, wherein the readout node (AK _a , AK _b ) in an end region of the Auslesefmgers (Ga, Gb) in the vicinity of the readout circuit (400) is arranged. 4. light transit time sensor (22) according to claim 3, wherein the readout node (AK _a , AK _b ) is bordered on at least three sides of the read-out gate (AG _a , AGBb). 5. light transit time sensor (22) according to claim 1, wherein between two Auslesefmgern (Ga, Gb) at least three modulation gates (yarn, Gbm, G0) are arranged, wherein the modulation gates (yarn, Gbm) to the Auslesefmger (Ga, Gb ) have both a transparency in an infrared (IR) and in a visual range (R, G, B).