Mercury doesn't have any atmosphere in the normal sense of the word, because what little gas it has is trillions of times thinner than the air at the surface of the Earth and therefore completely undetectable by any normal means, and because, given its low surface gravity and proximity to the Sun, any gases that currently exist at the surface will escape into space within a fairly short period of time.
      However, there are various sources that provide a very small, somewhat unsteady supply of gases to the surface of the planet, and as a result various kinds of atoms and molecules can persist at the surface for a short time. The major sources are (1) the Solar Wind, which delivers free electrons and bare atomic nuclei (primarily of hydrogen and helium) to the planet, (2) meteoritic impacts which vaporize portions of the surface rocks and of the impacting objects, providing small amounts of oxygen, silicon, iron, and other components of vaporized rock to the gaseous mixture, and (3) very occasional "burps" of volcanic gases through cracks in the crust of the planet, which of course deliver various kinds of sulfurous gases, similar to, though probably different in detail from the volcanic gases we are familiar with on the Earth. (All these to be discussed in more detail at a later date.)

Planetary Atmospheres Lecture Notes
(The following is not particularly thorough, because I said more in class than I wrote on the screen, and what lies below essentially represents what I wrote on the screen; still, the discussion below expands a bit on the brief comments above.)
      Mercury has no significant atmosphere. The amount of gas is measured in trillionths of a percent of the amount in our atmosphere (so if you were standing on the surface of Mercury you couldn’t tell there was any air there at all, without sensitive measuring tools). Also, any gas that is there at any given time would float off into space and be lost, or combine with the surface rocks and removed from the ‘atmosphere’, in a relatively short time -- a few months to a few years depending upon how picky you want to be about how much of it has to disappear before you don't count it anymore.
      So, for any air to be on Mercury, you need some source of fresh air, so to speak, that is continually replacing the lost material.
      There are three major sources --
      (1) The Solar Wind
      (2) Meteoroid bombardment
      (3) Outgassing
      The Solar Wind is a VERY thin gas streaming away from the Sun in all directions at hundreds of miles per second. This wind is VERY thin -- billions of trillions of times thinner than the air at the surface of the Earth -- but there is some of it in every part of the Solar System, and as it flows away from the Sun, Mercury is in the way, so some of it piles up in the magnetic field of Mercury (as in the van Allen radiation belts of the Earth), and runs into the surface.
      The amount of such ‘air’ must be very, very small, and it is made almost entirely of hydrogen because it’s coming from the Sun, and that’s what the Sun is made of, and as you will see hydrogen escapes more easily and faster from Mercury than any other material, so the amount of hydrogen that can be there at any given time is very low:

      Since there’s not much stuff in the Solar Wind and the hydrogen escapes pretty quickly, the amount of hydrogen at the surface of Mercury must be very, very small, but there should be ‘some’ there. It’s just a question of how much is there, and how easily, if the word easily has any meaning in this context, we can detect it.
      AND UNTIL THE MESSENGER SPACECRAFT ARRIVED AT MERCURY WE COULDN’T DETECT IT, because the temperature is too low for it to make its presence known, at all easily, and there’s not much there in the first place.

      SIMILARLY, meteoroids (small bits of ice or rock moving through space, in orbit around the Sun) are always running into Mercury.
      The Earth runs into several tons of meteoric material every day (different numbers in different references, because of different ways of averaging the amount of stuff per unit of time)
      Mercury, being smaller, only runs into a few hundred pounds of meteoric material each day. BUT when this stuff hits Mercury, it does so with VERY high speeds. Mercury is going around the Sun at about one hundred thousand miles an hour. And any other object moving around the Sun near Mercury’s orbit (either because it’s usually there, which is a very unlikely thing; or because it has fallen toward the Sun, which is more normal) will be moving at somewhere between 100 and 150 thousand miles an hour.
      Depending upon what direction the other object runs into Mercury, these speeds could result in a relative collision speed as much as 250 thousand miles per hour. And even if the speed is only a few tens of thousands of miles an hour, hitting something at that speed, and VERY SUDDENLY stopping, creates a LOT OF HEAT, and vaporizes the incoming object, and a large part of the surrounding countryside, producing an impact (or, more accurately, explosion) crater 10, 20, or even 30 times larger than the incoming object. And with that size, having a thousand to ten thousand times the mass of vaporized material, as in the incoming object.
      SO you can take the few hundred pounds of incoming material, and convert that to a few hundred to a few thousand tons of vaporized rock every single day at the surface of Mercury, because of bombardment by interplanetary debris. This produces a gas made primarily of the same things that rocks are made of -- about half oxygen, a quarter silicon, and a quarter other metal atoms, primarily iron, aluminum, potassium, calcium, magnesium, and ‘most importantly’, SODIUM.
      WHY IS SODIUM IMPORTANT?
      NOT because it is a very significant fraction of the vaporized rock. It is only a small percentage of that. But sodium, unlike hydrogen, is an atom that LOVES to make its presence known, even at ‘low’ temperatures, by absorbing and emitting light. SO much so that even though the sodium vapor produced in this way is only a small fraction of the vaporized rock, and that is only a few hundreds or thousands of tons (compared to a ton of air sitting on every square foot of the Earth’s surface, which is a heck of a lot of tons, overall), we can actually observe the fluorescence (glowing) of sodium vapor at the surface of Mercury from the Earth, on some occasions.
      Because of this, any discussion of the ‘atmosphere’ of Mercury will mention sodium. Not because it’s an important material in terms of the amount, but because it’s easy to observe. Similarly, you will read about sodium vapor surrounding Jupiter, in the orbit of Io. And in the case of the Jovian planets, about methane and ammonia in their atmospheres. HYDROGEN is the main component of their atmospheres, but does NOT make its presence known at the low temperature of those atmospheres, whereas the methane and ammonia do.
      IN OTHER WORDS, sometimes we talk about what is really there, and other times about what we can observe. And you’ll need to keep that in mind when reading about the atmospheres of the planets in your textbook and most other references. It is far more common for textbooks to discuss what we can detect, than what is really there.
      Outgassing of volcanic gases from the interior of the planet. There is VERY LITTLE of this going on, but apparently some, because we can detect the results on rare occasions.
      The gases that result would probably be similar to volcanic gases on the Earth, and among things a small proportion would consist of sulfur compounds, and as a result there is a very small amount (a few trillionths of an Earth atmosphere) of sulfur gas in the ‘atmosphere’ of Mercury. And like sodium, sulfur loves to absorb sunlight and fluoresce, and we can easily detect such sulfurous emissions from a cloud of gas surrounding Jupiter (in the orbit of Io) and on occasions, from Mercury.