I just did a brief bit of research on this. Apparently it was hypothesized by Hermann Sachse(1). Maybe he had an understanding of tetrahedral geometry of sp3 atoms? Not sure. However, his hypothesis was confirmed by a Ernest Mohr(2), through x-ray diffraction. What lead Sachse to such an idea is hard to say, I assume elemental analysis was present at the time. This was 30 or so years after kekule(3) suggested the ring structure of benzene, so some understanding of structure was present.

1)H. Sachse, Chem. Ber, 1890, 23, 1363; Z. Physik. Chem, 1892, 10, 203; Z. Physik. Chem., 1893, 11, 185–219.

2)E. Mohr, J. Prakt. Chem., 1918, 98, 315 and Chem. Ber., 1922, 55, 230.

3) Kekulé, F. A. (1865). "Sur la constitution des substances aromatiques". Bulletin de la Societe Chimique de Paris 3: 98–110. On p. 100, Kekulé suggests that the carbon atoms of benzene could form a "chaîne fermée" (a closed chain, a loop).

It was first suggested in 1890 by Hermann Sachse and later expanded on by Ernst Mohr. Cyclohexane is not flat as what Baeyer (1835-1917) assumed; instead, it is puckered into a three-dimensional conformation that relieves all strain. Their conclusion was based on combustion data experiments.

One method which avoids the need for NMR is the reduction of 4-tert-butyl cyclohexanone with NaBH4 or some other reducing agent which converts it to an alcohol. If you mix methanol, NaBH4 and the ketone then the alcohols will give two peaks in GC.

I am not sure what was invented first GC or the first primative NMR machine, but I think as a desert island chemist forced to build my own equipment that the GC machine would be more easy to create than a NMR machine. Not that it would be easy to build a GC out of nothing.

If you were to reduce the ketone and get two peaks, you might think that you have an impure product with something other than the alcohol, but if you were to convert it into an ester then you would still have two peaks. But if you dehydrate into an alkene then you would only have one peak.

While the GC observation would not show you the ring flipping from one chair to another, it would suggest that two different alcohols can be made thus in turn suggesting that the two sides of the ring are chemically different.

I just did a brief bit of research on this. Apparently it was hypothesized by Hermann Sachse(1). Maybe he had an understanding of tetrahedral geometry of sp3 atoms? Not sure. However, his hypothesis was confirmed by a Ernest Mohr(2), through x-ray diffraction. What lead Sachse to such an idea is hard to say, I assume elemental analysis was present at the time. This was 30 or so years after kekule(3) suggested the ring structure of benzene, so some understanding of structure was present.

1)H. Sachse, Chem. Ber, 1890, 23, 1363; Z. Physik. Chem, 1892, 10, 203; Z. Physik. Chem., 1893, 11, 185–219.

2)E. Mohr, J. Prakt. Chem., 1918, 98, 315 and Chem. Ber., 1922, 55, 230.

3) Kekulé, F. A. (1865). "Sur la constitution des substances aromatiques". Bulletin de la Societe Chimique de Paris 3: 98–110. On p. 100, Kekulé suggests that the carbon atoms of benzene could form a "chaîne fermée" (a closed chain, a loop).

I think it was only hypothesis at that time .theoretical chemist can better explain it

the discovery of the cyclohexane ring inversion was determined by gas electron diffraction (GED) before the use of NMR

If a copy of "Stereochemistry of Carbon Compounds" by E. L. Eliel is available, this question is addressed at the beginning of Chapter 8 and historically in an extended footnote to the first paragraph. It appears that Sache realized that the tetrahedral angle of carbon would not allow a cyclohexane ring to be flat but was a lone voice in the wilderness. This chapter is worth (re)reading as is the whole book.

The simplest to dichlorinate cyclohexane to 1,2, or 1,3 or 1,4 dichlorocyclohexane and observe the cis and trans two compound....

Colleagues,

"The invention of actual physical models of molecules has been ascribed to A. W. Hofmann. He is said to have built these models by taking a croquet set, drilling holes in the croquet balls, and using pieces of the croquet handles to connect those balls into large ball-and-stick modles that he used in a lecture demonstration given at a meeting of the Royal Institution of Great Britain, Friday, April 7, 1865, entitled "On the combining Power of Atoms". In this lecture Hoffman discussed the laws governing chemical combinations from the atomic point of view and introduced the concept of valency. However, at that time (1865) chemistry was considered to be two dimensional, and Hoffman's models were also two dimentional. These ball-and-stick modles subsequently evolved into smaller (and three-dimensional).

The earliest three-dimensional molecularmodles known to the present author were those constructed by van't Hoff in connection with his stereochemical work (1874) [1]. These models (which the author has had the privilege of holding in his hands) wer constructed of orang-colored cardboard triangles approximately one inch along an edge, which were glued together to form tetrahedra. With assemblages of these tetrahedra, van't Hoff described the geometries of simple molecules, as is now so well known. In the early 1890s, Sachse illustrated the boat- and -chair structures of cyclohexane as three-dimensional modles of essentially the Hofmann type [2]. However, cyclohexane was found to exist in only one form. To explain it Mohr (1919) suggested that these boat and chair forms continuously change into one another. Hence, they cannot be separated from each other and the concept of Sachse and Mohr is now known as the theory of strainless rings.

[1] J. H. van't Hoff, Bull. Soc. Chim. France, [2] 23, 295 (1875); the original (Dutch) version appeared in 1874.

[2] H. Sachse, Chem. Ber., 23, 1363 (1890); Z. Physik. Chem, 10, 203 (1892)"

See

--- Allinger, Norman L. , Molecular Structure : Understanding Steric and Electronic Effects from Molecular Mechanics , Wiley (2010).

I Think it is It was first suggested in 1890 by Hermann Sachse and later expanded on by Ernst Mohr based on the thermodynamics.

Regarding Mark Foreman's mention of use of t-butyl group to "anchor" the cyclohexane, please note that the best example of anchoring of cyclohexanes is in steroids. Around 1949 D.H.R. Barton proposed in a few pages "Conformation Analysis" in the scientific journal Experientia. For that he was awarded joint Nobel Prize in Chemistry in the late sixties.

H T Andrew Cheung

Xray diffraction and further supported by electron diffraction (Barton, 1950). An examination of Drieding model of cyclohexane chair showed it is devoid of any strain.