The history of the search for the search of the "perfect" oscillator is very fascinating. It was originally driven by the need for the most precise system for the marine chronometer. During the 1800s this took up a lot of mental energy. The issue is that there was a mismatch between the coefficient of elasticity for a hardened steel balance spring and the rate of contraction/expansion of the bimetallic balance. They could be matched at 50 and 90 degrees F, but the full curve resulted in a 1 second gain per day at 70. Hence, the term Middle Temperature Error.
When you are at sea for 3 years and on a ship with no environmental controls and you cross extreme latitudes in summer and winter, that 1 second per day becomes an arithmetic nuisance. It can and did result in many wrecks.
Some attacked this via materials (Frodsham experimented with glass), others took a mechanical approach where they "boosted" the balance response to temperature by affiixes such as small mercury thermometers or physical attachments (Poole,Kullberg).
Many of these watchmakers did the math and fully understood what they were doing. Finally, Guilluame, the Director of Swiss Bureau of Standards, invented Elinvar after developing Invar. He was awarded the Nobel Prize in Physics in 1922.
This was a game changer in watches because it eliminated the need for a split bimetallic balacne which was subject to Middel Temp Error and also physical forces such as centrifugal effects on the free ends.
The related alloys and the use of monometallic balances account for the incredible performance of modern mechanical wrist watches.
I had written an article about this work back in 1992.
Fast Forward. Quartz chronometers were first issued to the USN in WWII!. BIG suckers!
Mechanical chronometers were not removed from service until 1988 (I have a copy of the recall order) because quartz could not handle salt water and constant vibrations. Only GPS superseded the WWII-manufactured Hamilton M21 (which used Elinvar and the monometallic balance).
After WWII Hamilton started work on an electric watch, as did LIP in France. The idea was to eliminate changes in torque that result from a mainspring. They used a fine wire triggered by the balance position to fire a sole to impulse the balance.
This work is credited with the development of the button cell (I forget but it may have been Hamilton that created the first cell).
This watch was introduced in 1957, but it is a PITA to own and a disaster to service. There is only one person in the USA who can reliably service these because that is ALL he did. OTOH, Elgin's attempt was a disaster and had to be recalled from the market. If you just looked at the movement it fell apart.
Then came Accutron 214 which is just a pawl attached to a tuning fork that pushes an index wheel. These were actually used by NASA. The first ones did not use electronics IIRC. The issue with these is the coils do go bad. The last I heard people were trying to reinvent the winding procedures. If you have one serviced, make sure it is by someone who adjusts it for the higher voltage of the modern cells.
Then came the quartz revolution Again Hamilton was in the lead (Pulsar, I think). These cost in 1970 a couple grand. Omega came out with a quartz driven tuning fork watch that is now very rare in working condition. I do not think even Omega Switz has parts.
Finally came the democratization of quartz. I remember a trip to Canal St. in NYC in 1982 with a friend to look at selling quartz jellybean watches from China. Came in assemblies to avoid customs. Battery, strap, dialed movement, and case. Less than $2USD SHIPPED!. No wonder the Swiss speak of this period as the "crisis".
I have a friend at the US Naval Observatory who maintains the rubidium fountains. Took us for a tour. Amazing stuff.
There is a long history of engineering and material science in this field that should make any person with a physics background proud.