Here it says that antenna is passive device which doesn't offer any added power to the signal. But at same time it says that antenna is responsible for increasing the amount of energy to a radio frequency (RF) signal.
So I don't get aren't they opposite things? How can it be a passive device and increase the amount of energy to a RF signal?
An antenna is a passive device.
The gain of an antenna refers to its directivity times efficiency compared to an isotropic antenna.
An isotropic antenna is a theoretical antenna that radiates equally in all directions. If this antenna were encapsulated in the center of a sphere, it would illuminate all parts of the sphere equally and uniformly.
All other real world antennas do not illuminate a sphere equally. Some areas of the surface have more power than other areas. As a result of the antenna favoring some areas of the surface compared to others, the antenna is said to have gain when compared to an evenly illuminated sphere. The area or direction with the most power is considered to be the major lobe of the radiation.
While radiation has been used in this description, due to the theory of reciprocity, it applies equally to a receiving antenna.
You can also think of this in the context of a flashlight/torch bulb. If you illuminate the bulb without its reflector, it does not appear to be very bright. But if you now place a very narrow beam reflector behind it, you can shine it at someone's eyes with nearly blinding results. The energy emitted by the bulb has not changed, but to the observer it is as if the bulb is many times brighter. This is analogous to antenna gain.
That's a good observation.
The confusion comes from what "gain" is understood to mean in the special case of antennas. For ordinary circuits, gain is generally the output power divided by the input power. You are right that passive components therefore can't have "gain" (in this case used to imply gain > 1).
However, for antennas, there are two important differences in what "gain" means:
This is just like a flashlight with a reflector that focuses the light in a particular direction. In this context, you can say the reflector has high gain in one narrow direction, but nearly 0 gain in other directions. The total emitted light power does not exceed the light the bulb is putting out. But, viewed at a distance from where the beam is focused, it can appear that way.
This makes sense when you think about it, since you're interested in gain in a particular direction. Absolute power doesn't mean much in that context. It's power per solid angle that matters. The obvious reference is then the power per solid angle you get if the power was spread evenly in all directions, called a spherical radiation pattern.
Since dipoles are the low level basic antenna, sometimes gain of a particular antenna is quoted relative to dipole in the dipole's optimal radiation direction. However, this should be stated explicitly. Assume that unqualified "gain" of a antenna is relative to a spherical radiation pattern.
So therefore the gain of a antenna is really a measure of how much it can concentrate the radiated power in a particular direction. Antennas are passive devices and can't create more output power than input power.
In their most basic form, antennas are passive devices. Products advertised as active antennas are passive devices with amplifiers added to the feed network.
However, one of the most important properties for an antenna is their gain. Gain is a directional property that is commonly expressed in decibels referenced to isotropic (dBi). An isotropic antenna radiates the same power in all directions and therefore its gain is 1 dBi in all directions. However, real antennas radiate in patterns that are larger in some directions than others. A cross section of the radiation pattern for a dipole antenna is shown below. it's shaped kind of like a doughnut.
At the peak of the radiation pattern for the dipole, it is radiating 1.64 times more power than an isotropic pattern would in that direction given the same input power. Therefore in that direction the equivalent gain is 2.15 dBi (\$10\log (1.64)\$). However you can see that almost no power is radiated above or below the antenna.
For any perfectly matched antenna with no resistive losses, the integral over the sphere of the total radiated power will be the same for the same input power. However more power will be radiated in certain directions than others.
Satellite dishes are antennas that have a lot of gain in a very narrow beam. They are used to send very lossy (high frequency) signals over large distances. The extra gain from their directivity allows them to overcome the free space loss.
It is passive, in that there are no electronics in it and it requires no power supply.
However the shape and size of the antenna affects the shape and size of the radiation pattern that is emitted from it.
At the heart all antennae are just a "dipole" - that is, two conductors heading away from each other (or a monopole and the ground (the second pole) which is what you see with a traditional FM antenna). This dipole creates a radiation pattern with a polarisation (depending on the angle of the dipole). Different lengths create different radiation patterns (often pictured as doughnuts around the antenna). Some lengths cause the current waveform inside the conductor to fold back on itself (1/4 wavelength) doubling the apparent current at any one point in the conductor, thus increasing the instantaneous power but at the cost of "height" of the radiation pattern.
Then many antennae also have external portions that shape that radiated energy into a more usable pattern. For instance a parabolic dish will form the radiated energy into more of a beam. A Yagi array (TV antenna) makes it more like the shape of a flame. There are many arrangements for different situations.
But basically:
Since there is a fixed and finite amount of energy in that radiation pattern you can now see that shaping it is merely a case of moving energy from one portion to another. You get less power behind a Yagi array than you do in front. Such antennae are then said to be directional in that they work better in one direction. Without those external shaping elements you get an omni-directional antenna (which is actually a misnomer, since it's only really in 2 dimensions that it's "omni").
