Outstanding photodetector beats competition by multiples

Logo POET Technologies

Even the POET team was surprised: POET Technologies‘ new photodetector outperforms competition’s detectors twentyfold – enough so to develop it as a stand-alone product. Moreover POET Technologies provided evidence for mass production capability of its process technology, and thus took the last hurdle to commercial production. With this result POET inventor Dr. Geoff Taylor is retiring from active business. Read more in POET’s press release as of 2016-04-04 and in this blog post.

Deutsch Dieser Text ist die englische Übersetzung meines Blogbeitrags »Herausragender Fotodetektor schlägt Konkurrenz um ein Vielfaches«.

Aside from the excitement about the excellent performance of the detector, the POET announcement also raised some questions among investors in the forums. Sure, the detector is good news, but what about the rest? How do the new statements fit together with the previously communicated business goals and the schedule? That remained largely unknown or at least open to interpretation.

On 2016-04-07 I had the opportunity to speak with Dr. Subhash Deshmukh, Chief Operating Officer (COO) at POET Technologies, to clarify open questions. In this article you will find bothstatements from the POET news release as well as Deshmukh’s explanations.

Photodetector performs beyond all expectations


Figure 1: VCSEL transceiver (schematically)

Let’s talk abound the photodetector first. It is one component of the integrated VCSEL transceiver POET Technologies is developing. VCSEL transceivers are required for optical data transmission. The transceiver converts the electrical signals of a sending computer to optical signals and sends them to the remote site. There another VCSEL transceiver receives the light signals and converts them back into electrical signals the receiving computer can handle. In a communication, both parties are senders as well as receivers, thus each VCSEL transceiver has transmit as well as receive functionality.

Figure 1 schematically shows the structure of a VCSEL transceiver. It consists of three components:

  1. The VCSEL (red) serves as a laser light source for the optical signal to be transmitted. It passes through an optical fiber to the remote site.
  2. The photodetector (green) receives the light signal. The light coming via optical fibers from the remote party falls on the detector and generates an electric current: the output of the detector.
  3. An electronic control unit (blue, „transistors“, FET) conveys the data to be transmitted onto the signal light generated by the VCSEL. It modulates the light signal, and by doing so, turns electrical signals into optical signals. At the receiving end, the control electronics processes the electrical signals delivered by the detector, if needed.

In „How does the optical POET transceiver work?“ I have explained this in some more detail.

Typische Responsivity heutiger Fotodetektoren

Figure 2: Typical responsivity of photodetectors (pin diodes) available on the market today

The photodetector is crucial for the receiving end of the transceiver. The higher the proportion of the incident light the photodetector can convert into electricity, the better. This is an essential characteristic of a photodetector and is called responsivity. It is the ratio of the power of the incident light, expressed in watts (W), and the intensity of the current generated by the detector, expressed in amperes (A). This responsivity is given in amperes per watt (A/W).

A higher responsivity means the same light input results in a higher electrical current, or, to put it in other words, a detector with a high responsivity can produce the same current from a weak light signal as a detector with a lower responsivity can do at a higher light power. A a practical application this can be used to reduce power consumption, because, if a weak light signal is sufficient, you can reduce the transmit power.

Figure 2 shows two examples of responsivity of photodetectors available on the market today, usually called pin diodes. As can be seen from the curves, the responsivity not only depends on the material, but also on the wavelength, i.e. the color of the incident light. At a certain wavelength the responsivity is at a maximum. For common pin diodes, this peak is somewhere between 0.4 and less than 1 A/W.

Responsivity des POET-Detektors

Figure 3: Responsivity of the POET detector

The POET detector’s responsivity is literally off the chart:  It responds at an impressive 13 A/W, vastly outperforming the best current photodetectors.

To illustrate the vast difference between conventional photodetectors and the POET detector, I extended the scale in Figure 2 „a little“ and charted the responsivity of the POET detector as a red dot (figure 3).

For now, we only know the POET detector’s responsivity at a particular wavelength of 850 nanometers (nm). The actual curve would probably have a very steep climb and an equally steep drop around 850 nm, because resonant cavity detectors can be tuned to a particular wavelength very accurately. This also applies to the VCSEL on the remote side: The light it emits must exactly match the detector wavelength for the communication to work.

POET has characterized the responsivity of its detector at an input optical power of less than 200 microwatts (µW) and an aperture size of 10 micrometers (µm). The latter, by the way, is only a third of what other photodetectors require. (Unfortunately, only the PDF version of the POET announcement on SEDAR has the ‚μ‘ character correctly. In the HTML versions that went through the news media, the ‚μ‘ mutated to an ‚m‘ and incorrectly transformed micro into milli.)

A photodetector is usually equipped with a so-called transimpedance amplifier (TIA), because the current supplied by the detector is too low for subsequent processing. In the conference call on 2015-09-30, Geoff Taylor stressed the fact that the POET detector’s TIA would not be a separate component, but would be integrated into the detector cost-effectively. According to Deshmukh,it appears no additional amplification may be required. While this may not result in cost savings, it does save some valuable space. With an integrated TIA, the POET detector would have been much smaller than competitive products, without it, the size can be reduced even further.

Due to the unexpectedly good results, POET Technologies will not only integrate the detector as a building block into its VCSEL transceiver, but will also market it as a standalone product. Better still, as the detector can  be produced cost effectively and en masse, it is already at a near market stage, pending further short term optimization.

The current photodetector market is characterized by stiff competition and extremely thin profit margins. In addition to its superior responsivity, POET’s production costs are significantly lower than those of all other photodetectors on the market, according to Deshmukh. What’s more, competitors‘ margins are extremely thin. Current producers have no room for any price reductions. POET could offer its own detector at a fair market price, with a comfortable profit margin, without having to be afraid competitors would or could start a pricing war. Customers will have a choice: They could either buy a standard detector or POET’s vastly superior product at a highly competitive price point. This is placing POET in a dominant competitive position to win a significant market share.

The detector is anticipated to hit the market as the company’s first product by the end of 2016, the POET announcement says. This information, however, led to some confusion among the shareholders. Wasn’t the VCSEL transceiver going to be the first POET product? And wasn’t it to be completed by the end of Q2 2016? However, if a) the detector will be the first product, and b) will come in December, hasn’t POET fallen behind on the VCSEL transceiver by six months or even more?

Q1 milestone achieved

No, not at all, Subhash Deshmukh explains. He describes the individual milestones.

For the first quarter of 2016 POET Technologies‘ goal was to  prepare the company’s manufacturing process for high-volume production. The company has already fully achieved this target in the foundry of its production partner, Wavetek. Several lots of six-inch wafers ran through the production facilities and delivered detectors, VCSELs and transistors. The detectors were produced first, so their results were available first. VCSELs and transistors are still to be tested and characterized. This should take about three weeks, Deshmukh assumed.

Ability of mass production demonstrated

The significance of this milestone can not be overstated. Based on concrete production runs POET Technologies has demonstrated that its process is working in mass production, supplying high-volumes, and delivering a sufficiently low defect rate. This demonstration was the last technical hurdle the company had to take.

Manufacturers no longer have to wonder whether POET chips can be mass produced cost-efficiently or not. Now their question is or should be “how can we deploy the POET technology for our own products – before competition does?” POET is expecting first revenues for out-licensing its process to customers in the second half of 2016 (NRE revenues).

Schedule remains unchanged

The schedule, as presented in the conference call on 2016-02-16, remains as it is, see figure 4. Achieved goals are marked in green; yellow is in progress; red denotes future goals.

Zeitplan POET Technologies 2016

Figure 4: POET Technologies‘ schedule 2016


So what are the next steps for the VCSEL transceiver?

  • Now that the three transceiver components have been manufactured separately in the first quarter, they will now be optimized and integrated. The result of the second quarter should be an integrated chip containing VCSEL, detector and control electronics (NFET). The components should operate and function in concert and demonstrate transmission and receiving capabilities.
  • Further processing and optimization steps will follow in the third quarter. At the end, there’s the tape-out, the transfer of all files relevant for production to the foundry.
  • Production is expected to begin in the fourth quarter. Initially shorter runs will help ensure that all stages of the production process are functioning as planned and that the result matches the specifications. Once everything is aligned, and any last problems have been solved, high-volume production will commence.

Dr. Subhash Deshmukh

Dr. Subhash Deshmukh, COO POET Technologies

In December, two POET products are expected to appear on the market: the detector and the VCSEL transceiver.

POET Technologies is going to keep investors informed as substantive progress warrants… As an example, Deshmukh says, a functioning VCSEL would not be newsworthy. That would not mean anything special to the POET COO, because there are already a lot of working VCSELs on the market. However, should the VCSEL have similar outstanding characteristics as the detector, that would be another story. And of course a functioning integrated VCSEL transceiver would be newsworthy, because it would represent the first such device of its kind on the market.

Subhash Deshmukh impresses as a person with very high standards. „Working components are not enough“, he says. He wants more. He wants excellent quality. He wants to make the most of the options on hand. Will the integrated VCSEL transceiver function? For Deshmukh this is already a given. Will the transceiver be better than competing products? Of course, says Deshmukh. The question is not whether the POET transceiver will outpace the competition, it is “by how much?” My impression is that Deshmukh will rather spend extra time and effort to ensure that a product achieves its full potential. With the detector, he has certainly delivered a very impressive example.

Geoff Taylor is suffering from knee problems

Dr. Geoff Taylor

Dr. Geoff Taylor, POET-Erfinder

With these outstanding Q1 results and his “baby” reaching the production stage, POET inventor Dr. Geoff Taylor is retiring by the end of April. Health problems may have accelerated this decision. He suffered from complications with a knee replacement, requiring multiple surgeries so that Taylor was unable to perform his duties.

Since the 1980s he had researched the POET technology and developed the critical components. With the handover of leadership as well as technical and scientific responsibility to Suresh Venkatesan and Subhash Deshmukh, Geoff Taylor believes the company is in good hands and positioned for future growth. „The company is now on the verge of successful growth under its new leadership and I believe is well positioned to produce critical prototypes necessary for POET to be recognized as a leader in the data communications market. I have every confidence that this will be accomplished by the new management in place and I am closely monitoring the progress as they move forward with their demonstrations“, Taylor says.

Addressing corporate communications weaknesses

The press release confirms that POET Technologies is well on schedule, their detector results surpass expectations and a new product is in the pipeline. Ironically, investor reaction in the forums indicates that the message was understood only partially and vaguely.

However, if the communication does not even reach the company’s own investors optimally and the latter do not fully appreciate the technological advances, it is very, very difficult to find new investors and get them into the POET stock.

What are the reasons? Let me compare it to optical data transmission: In order to communicate successfully, transmitter and receiver must be tuned as closely as possible to the same wavelength. Figure 5 shows a communication where this is not the case. The transmitter (red curve) is sending a signal mainly at a specific wavelength, namely the maximum of the red curve. Right and left of this wavelength the signal strength decreases. The receiver (green curve), however, is tuned to a different wavelength. It could receive the signal well, if only it would be where the green curve reaches its maximum. A transmission will take place only where the two curves overlap (yellow area). In this example, it falls far short of its potential.

Figure 5: Weak communication due to different „wavelengths“

Communication between POET Technologies and its shareholders are similar. POET is sending on the red curve. Consider the following passage: „At a bias voltage of 3.3V, POET detectors have demonstrated a room temperature thyristor enhanced saturated responsivity of 13A/W at an input optical power threshold of less than 200mW and at an aperture of 10mm.  This is about 20X higher on a 3X smaller device relative to typical 850nm PiN diodes.“ It contains a lot of strong signals to recipients who are tuned to the red curve, i.e. semiconductor professionals and laypersons who have familiarized themselves intensively with the subject.

The vast majority of current and potential POET investors, however, is tuned to the green curve and is expecting news on this wavelength. Terms such as „bias voltage“, „responsivity“ (or even „thyristor enhanced saturated responsivity“), „input optical power threshold“ or „PiN diodes“ are substantially going over their heads. The layman has no idea what the numbers and units would mean or what he could compare them to.  While the message comprises a lot of good information, it does not effectively reach the recipients, because they are sent at the wrong „wavelength“ and do not cause enough resonance.

On the other hand, the investor is lacking the information he is actually expecting and to which it is perfectly tuned. He has, for example, a burning desire to hear something about how the development of the VCSEL transceiver is progressing. However, the word „transceiver“ is not mentioned at all in the whole POET announcement. „VCSEL“ appears only once, and the context raises questions.

What should be done? Recommending training in semiconductors, semiconductor manufacturing and optoelectronics to thousands of investors and other interested persons so that they can better understand and interpret POET messages, is not feasible.

No, POET Technologies must identify the problem, and get closer to the target group when communicating. It must attune itself to the receivers‘ wavelength. Message authors should try to put themselves in their readers‘ position, and perceive their knowledge and their expectations, possibly with some assistance. The technical specifications are important of course, but POET should also explain what they mean and how they fit into the overall context. As the company optimizes detectors and VCSELs, it should also optimize its communications, to give them maximum effect.

Figure 6: Improved communication due to more similar „wavelengths“

Figure 6 shows improved communication. The red curve has approached the green one. Although the wavelengths do not yet match perfectly, much more information is transmitted to the recipient, as can be seen by the size of the yellow area.

Title photo: Helmholtz resonators, Rudolph Koenig, Paris. Photo: Teylers Museum, Haarlem, Netherlands

On our own behalf

My blog posts about POET Technologies are free and should remain so, if possible. However, searching, preparing and writing posts, cost time and effort. If you value the result, I’d be happy to not only receive feedback, but also a donation of your choice to my bank account IBAN DE93360100430323898437, BIC PBNKDEFF. Subject: POET blog. Thank you so much!

Please note the information regarding risks and liabilities!


Kommentar verfassen

Trage deine Daten unten ein oder klicke ein Icon um dich einzuloggen:


Du kommentierst mit Deinem WordPress.com-Konto. Abmelden /  Ändern )


Du kommentierst mit Deinem Twitter-Konto. Abmelden /  Ändern )


Du kommentierst mit Deinem Facebook-Konto. Abmelden /  Ändern )

Verbinde mit %s

Diese Seite verwendet Akismet, um Spam zu reduzieren. Erfahre, wie deine Kommentardaten verarbeitet werden..