This application claims priority to UK application 1813110.2 filed on Aug. 10, 2018, the entire contents of which are hereby incorporated by reference.
The invention relates to improvements of semiconductor devices for light detection, and particularly to improvements in lens layers for semiconductor devices.
Integrated photo sensors are widely used in semiconductor devices for conversion of light signals into electrical signals with analog or digital signal output. Such optical sensors (e.g. Photodiodes, Image sensors or Avalanche Photodiodes) have optical active regions for converting incident light into an electric signal.
To improve the fill factor of integrated photo sensors, lenses are applied in the light path. In pixel sensors, refractive micro lenses are often used. In large area solar cells, Fresnel lenses are used instead, as refractive optics are too bulky for this application.
If a larger area is to be covered with a lens, and or multiple different sizes are required at the same time (i.e. on the same chip), established refractive micro lens processes fail. If a binary or Fresnel lens is used instead, then there are more degrees of freedom, but all manufacturing processes still impose limitations especially in terms of the minimum achievable dimensions (i.e. the critical dimension, CD). CMOS processes are limited in the achievable resolution for lithography and patterning. Binary lenses do require small features to be effective, especially for short distance and large area devices (i.e. devices having a diameter of more than a few μm).
Aspects of the present invention provide semiconductor devices for light detection as set out in the accompanying claims. In one embodiment the invention is a semiconductor device for light detection including a semiconductor layer having an optical active region for receiving incident light and a peripheral region around the optical active region. The device further includes a lens layer including a first lens for directing light into the optical active region, the first lens being located in a first region of the lens layer which overlaps a part but not the whole of the optical active region in the semiconductor layer.
Certain embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings.
Embodiments described herein provide lens designs for device configurations for which known lens designs may be unsuitable or difficult to manufacture. The described lens designs can improve the effective light concentration and/or manipulate the angle of incidence, without changing the given manufacturing limits. Also the embodiments can enable fixed designs for devices having flexible dimensions, which can simplify the lens layout and the manufacturing process, as well as the process control. The approach can be scaled, so that even very large areas can be covered with a lens of this kind, while no modification of the production process is required.
The purpose of a lens is to modify the light propagation along its path. The described embodiments focus on two main aspects.
The first aspect relates to how to improve the sensitivity of a photodetector. All detectors have a sensitive area (i.e. an optical active region of the device) and a footprint larger than that which includes an optically insensitive peripheral region (also referred to as the perimeter). The fill factor defines the ratio of the area of the optical active region to the whole detector area. Embodiments described herein can increase the sensitivity and improve the fill factor by bringing light incident on the peripheral region (outside the optical active region) into the optical active region. That is, the embodiments can change the light propagation from the peripheral region towards the optical active region. As the optical active region does not necessarily need a lens, the lens can be applied to the perimeter of the device only, in which case the critical dimension of the process is sufficient.
In the case of photodiodes, the optical active region 4 is usually quite large, whereas the insensitive perimeter is limited in extent and fixed. As a consequence, even for devices of different sizes, the perimeter stays substantially constant. Hence, a single lens design for optimal collection from the perimeter could be used for all sizes of photodiodes without requiring further modifications. Therefore a lens layout can be crafted once and then used in multiple products, which can reduce manufacturing costs. Also, the modelling of detector devices becomes easier, as the perimeter of the device is fixed, which makes modelling of flexible device sizes and shapes possible. If the binary lens is applied to the perimeter, the fill factor is increased and the lens only needs to cover a small device portion. It is enough to get the light into the optical active region, a focus to the center is not necessary. However, for light already incident on the optical active region, an oblique incidence can be a benefit, especially if the silicon surface is structured (for example with a Shallow Trench Isolation, STI, pattern). Light having an “oblique incidence” refers to light having an angle of incidence which is not normal or parallel to an interface (at the optical active region in this case).
The second aspect relates to how to modify the angle of incidence of light into the detector. The effective absorption width of detectors is limited but can be increased by oblique incidence of light. In a CMOS process, for example, a thin epi-layer is used for detection. If the light enters this layer with 45° or 60° instead of 0°, the path length is increased by 65% or 120% respectively, which causes longer wavelength light (e.g. IR radiation) to be detected more effectively. In addition, charges generated closer to the center of the active region are more likely to diffuse to the (correct) photodetector onto which the light fell. Crosstalk can therefore be suppressed by directing light into the center of the device (i.e. away from the edges of the optical active region). To achieve some of these benefits it is not necessary to focus all light to a single point. A scalable and manufacturable solution proposed herein, is to apply section lenses that cover a certain area and are repeated as often as required by the device area.
A lens can comprise multiple binary lenses, which are ideally placed concentrically and the layout of one lens stops when the CD limit is reached. Then the next lens starts until the CD limit is reached again and so forth. This pattern of consecutive lenses can continue from the center of the device out into the perimeter. The pattern of each binary lens, i.e. the widths of peaks and troughs, can be exactly repeated for each concentric lens. Each lens then causes the light to bend by an equal amount, so that light rays directed by different lenses are substantially parallel. Because the lenses are located at different distances from the center, but bend the light by substantially the same angle, the focal point of each lens is different. A lens closer to the periphery will therefore generally have a greater focal length than a lens closer to the center. The focal length is the vertical distance from the plane of the lens to its focal point. That is, individual binary lenses, having substantially the same pattern and bending angle but different focal lengths, are located (repeated) over the optical active region. Because the design of the lens is fixed in its pattern, it easier to layout, produce and control. Also the crosstalk of devices is reproducible as the light stimulus is the same for all sizes of devices, which give a much better predictability. The light is bent in an equal way across the whole optical active area and thus the same path length and light intensity is maintained regardless of the area of the device. The light is concentrated towards the center which can reduce crosstalk. Also the IR sensitivity is increased because of the longer effective absorption length in the photosensitive EPI layer, due to the increased angle of incidence. Because of the equal incidence behavior across the whole active area, the device can be scaled while the behavior stays relatively stable. If in contrast there would be one large lens or multiple lenses that focus all the light to the center (i.e. to the same point) of the device, then small devices would show a different sensitivity and crosstalk behavior compared to large ones and the lens elements would require critical dimensions LCD) that cannot be manufactured.
Combining a lens at the perimeter and multiple central binary lenses can improve the fill factor, increase the IR response, lower the crosstalk and enable fabrication with the limitations of the backend and critical dimensions (CD) of a typical CMOS process.
While specific embodiments of the invention have been described above, it will be appreciated that the invention as defined by the claims may be practiced otherwise than as described. The descriptions above are intended to be illustrative, and not limiting. It will be apparent to one skilled in the art that modifications may be made to the invention as described without departing from the scope of the claims set out below. For example, although the described embodiments make use of etching to form the binary lenses other possible techniques include imprint lithography and injection molding. Even spherical lenses could be formed accordingly (cut to half and arranged as described).
Each feature disclosed or illustrated in the present specification may be incorporated in the invention, whether alone or in any appropriate combination with any other feature disclosed or illustrated herein.
The invention may also be defined, in one case, according to any one of the following numbered clauses:
a semiconductor layer having an optical active region for receiving incident light and a peripheral region around said optical active region; and
a lens layer comprising a first lens for directing light into said optical active region, said first lens being located in a first region of said lens layer which overlaps a part but not the whole of said optical active region in the semiconductor layer.
a semiconductor layer having an optical active region for receiving incident light and a peripheral region around said optical active region;
a lens layer comprising a lens for directing light into said optical active region, said lens being located in a region of said lens layer which overlaps said peripheral region but not said optical active region.
providing a semiconductor layer having an optical active region for receiving incident light and a peripheral region around said optical active region;
providing a lens layer on top of said semiconductor layer; and
forming a first lens for directing light into said optical active region in a first region of said lens layer which overlaps a part but not the whole of said optical active region in the semiconductor layer.
providing a semiconductor layer having an optical active region for receiving incident light and a peripheral region around said optical active region;
providing a lens layer on top of said semiconductor layer; and
forming a lens for directing light into said optical active region, said lens being located in a region of said lens layer which overlaps said peripheral region but not said optical active region.
Number | Date | Country | Kind |
---|---|---|---|
1813110 | Aug 2018 | GB | national |
Number | Name | Date | Kind |
---|---|---|---|
4367366 | Bloss | Jan 1983 | A |
4418238 | Lidorenko | Nov 1983 | A |
5121979 | Cohen | Jun 1992 | A |
5121980 | Cohen | Jun 1992 | A |
6010647 | Nomura | Jan 2000 | A |
6046847 | Takahashi | Apr 2000 | A |
6229562 | Kremen | May 2001 | B1 |
6399874 | Olah | Jun 2002 | B1 |
7499094 | Kuriyama | Mar 2009 | B2 |
7916204 | Wells | Mar 2011 | B2 |
9099580 | Hirigoyen | Aug 2015 | B2 |
9116271 | Tajiri | Aug 2015 | B2 |
9754986 | Takahashi | Sep 2017 | B2 |
10249779 | Doci | Apr 2019 | B2 |
20040004723 | Seko | Jan 2004 | A1 |
20040257460 | Kuriyama | Dec 2004 | A1 |
20050029433 | Sakoh et al. | Feb 2005 | A1 |
20050121071 | Repetto | Jun 2005 | A1 |
20070035721 | Toshikiyo | Feb 2007 | A1 |
20070069108 | Inaba | Mar 2007 | A1 |
20070277869 | Shan | Dec 2007 | A1 |
20070284510 | Li et al. | Dec 2007 | A1 |
20090020690 | Toda | Jan 2009 | A1 |
20090141153 | Onozawa | Jun 2009 | A1 |
20090262408 | Ogasawara | Oct 2009 | A1 |
20100214432 | Nakata et al. | Aug 2010 | A1 |
20100214456 | Tokuda | Aug 2010 | A1 |
20120247534 | Nakaya | Oct 2012 | A1 |
20120260986 | Chen | Oct 2012 | A1 |
20130105666 | Saitou | May 2013 | A1 |
20140174501 | Jost | Jun 2014 | A1 |
20140182659 | Davis | Jul 2014 | A1 |
20150090862 | Matsui | Apr 2015 | A1 |
20150156478 | Ono | Jun 2015 | A1 |
20160133771 | Pelletier | May 2016 | A1 |
20190191144 | Arbabi | Jun 2019 | A1 |
Number | Date | Country |
---|---|---|
S62-28704 | Feb 1987 | JP |
2002280532 | Sep 2002 | JP |
WO-2007026588 | Mar 2007 | WO |
WO 2011-112842 | Sep 2011 | WO |
Entry |
---|
Machine translation, Moriya, Japanese Pat. Pub. No. JP S62-28704, translation date: Jan. 1, 2021, Espacenet, all pages. (Year: 2021). |
Machine translation, Urabe, Japanese Pat. Pub. No. JP 2002-280532, translation date: Dec. 31, 2020, Espacenet, all pages. (Year: 2020). |
Abstract, Ogasawara, WIPO Pub. No. WO 2007-026588A1, Foreign Patent Retrieval Service, all pages. (Year: 2007). |
GB, Intellectual Property Office, Combined Search and Examination Report, Application No. GB2014914.2, 5 pages (dated Oct. 30, 2020). |
GB, Combined Search and Examination Report under Sections 17 and 18(3); Patent Application No. GB1813110.2; 7 pages (dated Jan. 31, 2019). |
Number | Date | Country | |
---|---|---|---|
20200052026 A1 | Feb 2020 | US |