top of page

The Photoelectric Effect in Silicon Image Sensors: A Defense [Part: 2]

In the previous part of this blog [LINK], we investigated the issues with the photoelectric effect with respect to solid state physics, showing that there was only three possible mechanisms and that the photoelectric effect was the only effect likely to be important with wavelengths within the visible spectrum. The other processes being Compton Scattering and Pair Production.

This set of posts was sparked by a rather curious outburst from an academic in my past, mainly that the photoelectric effect was "irrelevant" for the operation of solid-state imaging technologies such as Si CMOS Imagers or SPAD detectors.

Source: In this post I'll focus on the rather deep (both technically and philosophically) text book:

"The Nature of Light: What is a Photon?", C. Roychoudhuri, A. Kracklauer and K. Creath (Editors), CRC Press, 2008, ISBN13: 978-1-4200-4424-9

Lets systematically go through some of the sections relating to the photoelectric effect:

  • Page 400 (Chapter Authors: C. Roychoudhuri and N. Tirfessa):

  • " Einstein's 1905 paper on photoelectric effect reflects the experimental observations of Hertz (1887) and others after him. In all these early experiments electrons were released free from metal plates and measured as a current through a collection plate whose voltage was manipulated to measure the kinetic energy of the free electrons. Einstein correctly formulated the observed results as if a photon carries a packet of energy hv which is expended to provide the binding energy of the electron in the metal (work function) and the rest is used by the electron as its kinetic energy (KE) as a free particle. This is a bound-free transition. "

  • This certainly sounds like it is based on the vacuum tube designs of the time and indeed is the basis of what we will later define as the " external photoelectric effect ", i.e. photons promote electrons with enough energy that they leave the surface of the metal.

  • " In contrast, electrons in modern photo-detectors undergo bound-bound transitions (valence to conduction band). These detectors, including "single-photon" counters are essentially semiconductor p-n junction devices where electrons experience quantum mechanical level transition from valence to the conduction band after absorbing energy from an incident EM field. "

  • This certainly looks to me like direct, published, robustly edited material that relates to what we will define later as the " internal photoelectric effect ".

Now, the picture becomes muddied by this debate over the energy imparted to the electron, and this indeed may cause issues with the exact physical interpretation of the "photoelectric" nomenclature. The external photoelectric effect with its requirement of electrons leaving the surface of the metal uses the metal's work function as this value. In contrast the internal photoelectric effect only requires a photon energy equal to the energy gap between the valence and conduction band in semiconductors. As metals do not have energy gaps and instead have a sea of de-localized electrons, the nomenclature gets further muddied between metallic photo detection and semiconductor photo detection.

A further issue in this context, is the different nomenclature given to bound-free transitions (photoelectric effect) and the bound-bound transitions, which in the above source are described only as "photo-detections". It may be the case that different fields use the terms in different ways or their use is more relaxed in different fields. I suspect that the nomenclature of the photo-electric effect will be appropriately tightly bound and well defined in solid-state physics, but that same nomenclature may be less well defined in the engineering field. If we acknowledge different fields use words in a different manner, we then accept that the same field should use their peer reviewed nomenclature uniformly. As a student of engineering, and indeed the academic in question being in the same field, one would expect a common agreement on the use of the term " photoelectric effect ", and indeed agreement with the appropriate peer reviewed literature.

At this point lets demonstrate the disagreement in the nomenclature by refering to some dictionary definitions of the photoelectric effect, see link.

  • " Photoemission is often referred to as the external photoelectric effect. In contrast to photoemission, any photoelectric effect that results from electronic transitions from bound to quasi-free states within a solid is called an internal photoelectric effect. "

  • " Because of the very high conductivity of metals, no internal photoelectric effects are observed in metals, and only photoemission occurs. "

  • Source: The Great Soviet Encyclopedia, 3rd Edition (1970-1979). © 2010 The Gale Group, Inc.

This Soviet encyclopedia entry seems to be in agreement with the manner in which the photoelectric effect promotes an electron to free it from its bound shell (conduction band), and in agreement with the C. Roychoudhuri, A. Kracklauer and K. Creath textbook above. The difference being that the book by Roychoudhuri stops short of linking bound-bound transitions to the nomenclature of the "internal photoelectric effect".

Note, the use of quasi-free state in the above encyclopedia entry may relate to the ability of the electron to move within the lattice structure of the solid once it is promoted into the conduction band (i.e. free to convey a current, but not free of the material).

Also note the destination above between metals and semiconductors. In a discussion clearly relating to CMOS imaging sensors, SPADs, Silicon or III-V photo-detectors, we are clearly not talking about metal band structures and are talking about the band-gaps formed within semiconductors.

While I will be discussing the scientific literature (in the relevant Engineering field) in a later post, a quick google search of "internal + photoelectric effect + transition", yields a number of scientific texts. These are worth mentioning here as we have clearly got to the stage where this debate is one of word usage within differing fields.

Source: "Semiconductor Physical Electronics", Seng S. Li, Springer, 2012 [Link]

  • " In this chapter, we are concerned with fundamental optical properties and internal photoelectric properties in semiconductors. The optical properties associated with the fundamental and carrier-free processes and internal photoelectric effects such as photo-conductive, photo-voltaic and photo-magneto-electric effects in a semiconductor will be depicted. Many fundamental physical properties, such as the energy band structure and recombination mechanisms can be understood by studying the optical absorption process and photoelectric effects in a semiconductor. Practical applications using internal photoelectric effects such as photo-voltaic and photo-conductivity effects have been extensively reported in the literature. Future trends are moving towards the development of various optoelectronic devices and optoelectronic integrated circuits (OEICs) for use in computing, communications, signal processing and data transmission "

  • " The photo detectors, which employ internal photoelectric effects to detect photons in a semiconductor device, are described in section 12.4. A p-n junction or a Schottky barrier photo detector can be very fast and sensitive when operated under a reverse-bias condition. "

  • " In this section the basic principals and general characteristics of a p-n junction photodiode, are discussed. Figure 12.15.a shows the schematic diagram of a revers-bias p-n junction photodiode. Electron-hole pairs are generated by the internal photoelectric effect in the diode to a depth of the order 1/a, where a is the optical absorption coefficient at the wavelength of interest. These photo-generated electron-hole pairs are separated in the depletion region by the built-in electric field and collected as a photocurrent in the external circuit."

The above source, has now just proved my point. That within the engineering, optical detection community, the internal photoelectric effect is in use and does indeed relate exactly to semiconductor solid-state circuits for the express applications of communications and is expressly linked in the literature to p-n junctions.

I have to admit, this is starting to sound like the academic in question was taking his view of the photoelectric effect too much from the physics community's nomenclature and not from the nomenclature of the optical electronics and sensors engineering community. This is hardly fitting given the discussion was a) most certainly engineering and b) that it was most certainly a discussion of solid-state semiconductor imaging technologies. The overriding feeling at the time, was that the discussion should be based on literature and nomenclature used within the current field of discussion. If this is not the case, then one would equally, logically be able to say that lawyers or social scientists use a word in a different way to engineering or physics, and that an engineer or physicist should use that meaning of the word, rather than their own, in a publication intended for the engineering or physics community. No, quite the contrary, we should always use words that are accepted in our own field and indeed the accepted meaning as used by our field of work.

Source: "Fiber Optics Standard Dictionary", M. Weik, Springer, 1997 [Link]

  • " Internal Photoelectric Effect: The changes in the characteristics of a material that occur when incident photons are absorbed by the material and excite the electrons in various energy bands in the molecules composing the material. Note 1: Characteristic changes include changes in a) electrical conductivity, i.e. the photoconductive effect, b) electric potential development, i.e. the photovoltaic effect and c) photosensitivity. Note 2: Changes in emmisivity, i.e. the photoemmissive effect, are not internal photoelectric effects. Note 3: Electrons may move from a valence band to a conduction band when the material is exposed to radiation thereby increasing their mobility. The material includes the intrinsic material and the impurities, including dopants. The internal photoelectric effect includes both intrinsic and extrinsic effects. "

  • " Internal Photoelectric Effect Detector: A photodetector, a) in which incident photons raise electrons from a lower to a higher energy level, resulting in an altered state of the electrons, holes or electron-hole pairs generated by the transition, which is then detected, or b) may be used as a photodetector in a fibre optic receiver."

So, this source also seems to directly link the mechanism the chap agreed with, with the nomenclature of the "internal photoelectric effect", further it links that effect to detectors for receivers.

For future reference, a quick google of the terms "SPAD + single-photon + internal + photoelectric effect", yielded a number of peer reviewed papers from world renowned author's within the CMOS SPAD community [Link]. I won't dwell on this here, as it will be a separate blog post, however here are just a few of the papers that directly link SPADs to the internal photoelectric effect and therefore nullify this particular academic's view that I, and I alone, was incorrect in linking the effect to silicon solid-state imaging and photon counting technologies. "Irrelevant" indeed. TOSH...

  • [Link] "Progress in Silicon Single-Photon Avalanche Diodes", Ghioni et al, IEEE Journal of Selected Topics in Quantum Electronics, 2007

  • [Link] "Advanced Photon Counting: Applications, Methods, Instrumentation", P. Kapusta, M. Wahl, R. Erdmann, Springer, 2015

  • Andreas Bulter: "Detectors are core components in every setup based on photon counting. For the spectral range between approx. 300 and 1000nm, there are essentially two detector classes available: detectors based on the external photoelectric effect, such as photomultiplier tubes, microchannel plate photomultipliers or hybrid photomultiplier tubes, or detectors based on the internal photoelectric effect such as single-photon avalanche diodes."

  • Andreas Bulter: "In contrast to photomultiplier tubes, which are based on the external photoelectric effect, i.e. the generation of photoelectrons through a photocathode, avalanche photodiodes are based on the internal photoelectric effect, i.e. the generation of photoelectro0ns inside the device"

Lets take an example that is separate to the IEEE archived Ghioni paper referenced above.

  • [Link] "Schottky barrier photodiodes with antireflection coating", M. V. Schneider, The Bell System Technical Journal (Volume:45 , Issue: 9), 1966.

  • Abstract: "Schottky barrier diodes can be used for fast and efficient photodetectors if the incident light is coupled into the depletion layer of the diode and if electron-hole pairs are created by the internal photoelectric effect in the depletion layer. Fast response of the diode is achieved by designing a Schottky barrier with a small RC product. High quantum efficiency is obtained by coupling the light through a thin metal layer into the depletion region of the diode and by using an antireflection coating on the metal layer for matching the incident light beam."

I will close this set of sources from the photonics community and the optical detection engineering community with a patent. This is for an array of SPADs and crucially also links SPADs to the internal photoelectric effect. Being a company's patent one would assume the patent is robustly written and thought out, with authors or people behind the scene that have researched the area. Likewise the patent is a legal entity, therefore it is robustly checked for novelty and correctness. One would assume a company would get their facts exactly right, and in line with the rest of the community, prior to publishing something a) in the public domain and b) as a legal declaration of their research and IP. Indeed the question would be if a company would risk including not well checked or facts of debatable accuracy into a document their competitor would have access to?

  • [Link] "Device and method for detecting light", Patent: US 20130015331 A1, H. Birk, V. Seyfried, 2013, Leica Microsystems Cms Gmbh (i.e. the scientific imaging arm of the world renowned camera manufacture Leica cameras).

  • " Avalanche photodiodes (APD) are also known in the field, these being highly sensitive and fast photodiodes. They exploit the internal photoelectric effect to generate charge carriers and the avalanche breakdown (avalanche effect) for internal amplification. APDs are regarded as the semiconductor equivalent of photomultipliers and are used, inter alia, in the detection of very low optical power levels and have cut-off frequencies up to the gigahertz range. "

In the next blog post on this issue, and indeed on this particular academic's rather unprofessional response to the use of the photoelectric effect with relation to solid-state imaging sensors, despite the body of literature in engineering, I will look at some of the world leading companies in the field. For example we all know that Canon produce many of the world's photographic cameras, many now using CMOS silicon image sensors and older cameras using silicon CCD image sensors. What is their take on the situation? Likewise what do companies such as Olympus, Panasonic or Nikon have to add into this debate?

Stay tuned...

bottom of page