Why does accretion of planetesimals, in forming the planets, only work for solids?
Gases would fly off. Gases consist of atoms or molecules that are unbound. In a room, the air is free to move, being limited only by the walls which trap the gas in the room. What would hold a gas together when a few gas molecules collide? Only gravity, if the gravitational attraction between the molecules is stronger than the tendency of the gas molecules to move apart, i.e. if gravitational attraction is stronger than thermal pressure.
Liquids, like gases, are unbound atoms and molecules; however, the motion of the materials in liquids is much slower so the material takes up less volume. Typically, the motions of liquids are slow enough that walls are not needed to keep the material from dispersing (although walls of a container are needed keep gravity from distributing the liquid into the flattest possible form).
A solid contains such slow moving atoms and molecules that the atoms/molecules are bound together by the sharing of electrons. The energy of these atoms/molecules are too low to enable them to break these electronic bonds and move about freely, unbound.
Whether a material is in the gas, liquid or solid phase is determined by the ambient temperature and pressure. At standard temperature and pressure (room temperature at the surface of the Earth), water is a liquid. At the same temperature, but on the Moon, without 30 pounds per square inch of atmospheric pressure to hold moving water molecules down, water in a bucket would evaporate into the gas phase. But at the distance of Saturn, water would freeze into the solid phase.
There were no liquids in the solar nebula. Why? In an interstellar cloud, the pressure and temperature are both very low. If we were to find water molecules in the gas phase in such a cloud, and if we were to slowly cool them off, they would go directly from the gas to the solid phase, because of the low pressure. Thus, in the solar nebula, which was a small fragment of an interstellar cloud, we need only think about solids and gases.
H and He were always in the gas phase. No interstellar cloud in a galaxy can every get cold enough for hydrogen or helium to go through a phase change from gas to solid (or liquid). Light from stars in our and other galaxies, and even light from the Big Bang itself, is sufficient to keep the H and He warm enough to stay in the gas phase.
Water(H2O), Carbon Dioxide (CO2), Methane (CH4) and Ammonia (NH3). These materials are compounds of hydrogen, oxygen, carbon and nitrogen, the four most abundant materials in the universe (ignoring helium). At what tempertures would they be gases and solids in the solar nebula? Observations of comets tell us about the behavior of water. At and beyond Jupiter's orbital distance from the Sun, comets do not develop comae and tails. But once they cross to asteroid belt distances, the water ice begins to sublimate. This distance from the Sun, the distance at which water ice begins to sublimate -- or thinking about a comet receding from the Sun, the distance at which a vapor cloud around a comet would begin to freeze back as frost onto the surface of the comet -- is called the snow line. Thus, inside of about 4 AU from the Sun, if I tried to make a planetesimal (or a comet), the water ice would be unstable as a solid and would sublimate. Only outside of 4 AU could I make solids, or planetesimals that contain water! Therefore, if at 1 AU I can only make a planet out of solids, I will have little, if any, water to work with. Similarly, carbon dioxide, methane and ammonia remain in the gas phase out to much greater distances from the Sun (i.e., they don't freeze out until much colder temperatures than for water).
The Jovian advantage. Our observations of stars and galaxies reveal that 98% of everything is H and He. The other 2% is the sum total of all the O, C, N, iron, nickel, gold and all other elements put together. Now, imagine the zone in the ecliptic plane at about 1 AU from the young Sun. Herein, the Earth is forming, and for starters 98% of the mass in this zone is unusable because it is in the gas phase (H, He). Worse, the next three most abundant elements in the universe - O, C, N - readily form molecules with each other (O2, CO, CO2, N2) or with hydrogen (water, methane, ammonia). All of these easily formed molecules are volatiles - that is, they sublimate at relatively low temperatures, making them very difficult to find as solids. Consequently, most of the remaining 2% of the mass is trapped in volatile substances, trapped in the gas phase. What's left for making solids? At temperatures below about 2000 K, metal oxides (e.g., Al2O3) can condense from the gas phase; at temperatures below 1300 K slicates ("rocks" containing SiO4) can condense; below 1000 K, we start to get very common materials like iron-oxide (FeO), olivine (Fe2SiO4 and Mg2SiO4). Thus, the solids available to make the terrestrial planets were (first, where it was warmest) metal oxides [think: Mercury] and next (where it is a bit cooler) silicates and metal oxides [think: Venus, Earth, Mars]. As we move outwards from the Sun, as the temperature presumably continues to decrease, no other significant new materials are added to the mix until we add water. And we don't get water until we get out past 4 AU. But when we do pass 4 AU, we suddenly have water ice to work with. And the amount of water ice would be enormous compared to the tota mass of metal oxides and silicates. Therefore, at the present day distance of Jupiter, the process of accretion has an enormous advantage in that it has much more material in the solid phase to work with.
Forming Giant, gaseous planets. Current theory suggests that giant planets form through a two stage process, whereas the rocky, terrestrial planets require but a single step (accretion). The gas giants first undergo accretion from solids, taking advantage of the presence of icey solids in the outer solar system to accrete large 'cores' quickly. In fact, as best as we can tell, Jupiter, Saturn, Uranus and Neptune all have cores containing about 10 Earth masses of rock and iron, so-called 'rocky cores.' Once these protoplanets grow to about 10 Earth masses by accretion, they are capable of collecting gas using gravity. This becomes possible because the escape velocity from these protoplanets becomes greater than the thermal velocity of the gas.
So, provided there remains some gas to collect, these rocky cores sweep up all the nearby gas as they orbit the Sun. Since 98% of the gas is H and He, they will become more and more Sun-like in composition, as they sweep up more and more gas. Jupiter and Saturn did this very successfully. Uranus and Neptune appear to have failed to sweep up much gas, perhaps because there was no gas left to collect by the time their cores were big enough to begin the sweep-up process.
Forming comets. The snow line concept restricts our ideas on where comets formed, principally by defining the region inside of 5 AU as off-limits for comet formation: comets must have formed at distances form the Sun greater than 5 AU. Despite the present day existence of the Kuiper Belt and the Oort Cloud, the consensus among astronomers is that comets did not form in these locations. Rather, they formed in the giant planet zone, from 5 AU - 30 AU. Comets formed as icey planetesimals; then, they experienced close gravitational encounters with the giant planets, not close enough for them to be accreted by the giant planets, but close enough for the giant planets to toss them into dramatically different orbits. Jupiter and Saturn, being so much more massive than Neptune and Uranus, were able to toss enormous numbers of comets far out, creating the Oort Cloud. Neptune and Uranus tossed objects out into the Kuiper Belt. Many of the comets that were tossed out of the inner solar system probably escaped into interstellar space; however, gravitational forces from surrounding stars and the remaining molecular cloud out of which the Sun formed should have slowed the outward rush of some of the comets such that they went into orbit around the Sun and became the Oort Cloud.