Why are PNP transistors used?
We use pnp mainly for convenience and to simplify certain circuits. npn is a device used to reduce current, while pnp is a device used to find power. for a current source, it is clear to use pnp for a current source connected to the positive power rail while using npn for a current source connected to the negative power rail.
for the signal, npn generally generates a signal referenced to the positive supply rail, while the pnp generates a signal referenced to the negative supply rail. it is therefore easier to design a circuit with pnp than without it. However, in some integrated circuits, it can be very expensive to use pnp. in this case, you can think twice before using one. you would try to design first using all the npn.
Discrete bipolar transistors are inexpensive and widely available. They come in a wide variety of speeds, rated voltages and power handling characteristics. High trans conductance makes it a good choice for linear audio applications.
(fets also have advantages – a zero DC grid current is a big problem! it makes them very well suited for dense digital circuits.)
many circuits use the npn and pnp polarities. you can not create a complementary output stage without both. an input stage for a typical audio amplifier would use a differential pair of npns, loaded with a current mirror constructed from pnps. (or vice versa.) Look at the diagram of any opamp, you will probably see a lot of PNN and PNP. you can see or not see the fets – some opamps use fet inputs, others are purely bipolar.
the truth is that often you can get by without pnp transistors but they can allow more practical designs that are also likely to be more efficient than the cicles performing the same function without pnp transistors. indeed, a pnp transistor assumes a current in the opposite direction to that of a npn in a similar configuration.
To be specific, consider an amplifier to drive a conventional loudspeaker. often a combination of pnp and npn is used to drive the current through the speaker with one of these couples able to send power in one direction through the voice coil of the speaker and thus to apply a force in one direction to move the loudspeaker cone in a particular direction, while the other member of the pair sends the current in the opposite direction, exerting a force in the opposite direction and eventually moving the cone in the opposite direction, or sometimes simply by slowing down the movement previously caused by his partner in the pair.
a similar need arises in integrated circuit operational amplifiers where another type of load (for example a resistor) must be supplied with current in one direction or the other, by means of the load or the voltage of the one or the other sign appearing at the exit.
there are many intelligent ideas exploiting the two different polarities to perform some tasks in a simpler way, too numerous to describe here, but these two polarities are one of the reasons why semiconductors have advantages over vacuum tubes in circuit design.