But I find it to be absolutely incorrect. How can a BJT transistor even work without a base? Shouldn't a transistor without base be just a semiconductor (PP, NN)? Does there exist some special NASA-army-grade experimental BJT transistor without a base?
So who is right, me or the teacher?
Clearly the source of confusion here is what base means: you assume it's "base region", and your teacher seems to assume "base terminal".
What defines a BJT transistor is the NPN or PNP structure. How many of these regions are exposed through terminals is irrelevant: most transistors will have three, but phototransistors may have only two, and parasitic transistor structures may not be exposed at all, but they still behave similarly and are considered to belong to the same device class.
Phototransistors are a bit special in that they have no base current in the strict sense, as carriers are generated in the base itself. But since the structure still behaves quite similarly to a regular BJT, this fact is usually ignored and photodiode current generated by light is considered to be base current.
I think you and your teacher are both right - you are just not using the same definitions and thus your wording disagrees.
From your description:
Shouldn't be transistor without base be just a semiconductor (PP,NN)?
You define the "base" to be the semiconductor in the middle of a normal BJT - the P-doped region in a NPN or the N-doped region of an PNP. In this regard you are right: A phototransistor still has this PNP or NPN structure.
Your teacher might define the base differently - the terminal you put current through to achieve a collector-emitter current. When you look at it this way, he is also right - most phototransistors (I know of none myself that break this rule, apart from some optocouplers) have no "base" lead. The only way to trigger a current flow is through photons hitting the base region, and causing the release of electrons by doing so, turning the device on.
Both make sense, although I would argue that you need to watch out when thinking that "no base" means "pp or nn" semiconductor. Many complex semiconductor structures exist that have many regions, yet have few leads. Look at Triacs or IGBTs!
The base connection is often not used and would thus serve mostly to pick up noise.
You can find optocouplers (4-pin) which have no base connection and others (eg. 4N35) which bring the base connection out (where it is sometimes used).
Similarly, individual phototransistors are often packaged in an LED-like packages with lens and 2-pin leadframe.
Your teacher right. BJT photo-transistors rarely have base connection.
It's the light hitting the base region that controls the current flow from collector to emitter. The more light, the larger the current flow. That's what makes it a photo-transistor and not just a transistor.
Base is already present. Connection to base is present too, but it works not from electricity - from the light. No matter how the energy comes into transistor, but it MUST come someway to make it work, so you can think of base connection as the wireless to the Sun )
I use the base pin on a phototransistor to handle DC currents from the sun, or 60Hz currents from incandescent lighting.
simulate this circuit – Schematic created using CircuitLab
By nulling out the DC error, you can use high value collector resistors thus achieve high gain.
An alternative is that current source I drew in parallel with the collector resistor, to be activated by Vout lower than e.g. VDD/2
