A history on chemical valence as a contribution to chemistry education.
It’s common in an elementary organic chemistry course to talk about Kekulé and his insights. He argued that just as every element has an atomic weight, it also has only one atomicity. By atomicity, he meant valence. Based on this assumption, which, in fact, had already been questioned by other chemists like Couper, Frankland, etc., he built the chemistry of carbon [see card 1]. His firm belief that carbon only exhibited a chemical valence of four in its compounds certainly constrained him to hypothesize chains of carbon–carbon bonds in organic molecules. Today we know that the chemical valences of an element are almost always more than one and that carbon is no exception. So when we present this period of chemical history to students, we can’t help but highlight an epistemological point that isn’t isolated, namely that revolutionary discoveries can come from a wrong hypothesis. But it’s also true that, in the case of this great chemist, who enjoyed considerable prestige, insisting until the end of his days on the constancy of an element’s valence acted as a brake on the discovery and classification of compounds like coordination compounds. Faced with the discovery of more and more compounds where the constancy of element valences seemed to waver, Kekulé provided alternative explanations to enforce his belief in the constancy of chemical valence. It’s important to present these solutions to give students a correct picture of science where successes are as important as failures. Before reviewing Kekulé’s attempts to maintain the constancy of chemical valence in various compounds, let’s see how he defined the atomicity of an element in his 1864 article in the journal Zeitschrift für Chemie und Pharmacie,
What Dalton’s atomic theory doesn’t explain is the question of why atoms of different elements prefer to combine in certain proportions rather than others. I thought I could clarify the whole matter of these facts through what I called the atomicity of elements. The theory of atomicity is therefore a modification that I believed I could add to Dalton’s theory, and you can see that, according to my view, atomicity is a fundamental property of the atom, as constant and unchangeable as the atomic weight itself. It would be like using the term in a completely different sense than what I attributed to it when I proposed it, if one were to assume that atomicity could vary and that the same substance could sometimes work with one atomicity and sometimes with another. That would be confusing the concept of atomicity with that of equivalence. No one doubts anymore that the same substance, and even elementary substances, are capable of reacting with different equivalents. The equivalent can vary, but atomicity cannot.
It’s important to clearly understand the distinction that Kekulé made between the atomicity of an element (or chemical valence) and its equivalence, or rather equivalences. Atomicity expressed the property of a certain element to bond, and it had a unique value for that element. The equivalence of an element in a specific compound indicated the hydrogen atoms (atomicity one) that could be substituted. This was something that could be determined experimentally. If in the compound PCl3 the phosphorus atom had an atomicity equal to three, in the compound PCl5 the phosphorus atom kept its atomicity but showed an equivalence equal to five (five chlorine atoms were equivalent to five hydrogen atoms). The maximum saturation capacity (that is, the maximum equivalent) could thus exceed atomicity. Kekulé defended the invariability of an element’s atomicity or chemical valence until his death (1896). To resolve the apparent paradox of the uniqueness of atomicity and the variability of equivalence, he proposed different explanations for categories of compounds in which elements showed variable chemical valence, showing us the ingenuity of the solutions proposed beyond their correctness. Kekulé’s attempt to differentiate atomicity and equivalence sparked a lively debate, which ultimately led to the definitive overcoming of the invariability of valence, paving the way for the understanding of coordination compounds, especially through Werner’s work. The novel point in the article in Zeitschrift für Chemie und Pharmacie is the introduction of the concept of molecular compounds:
The compounds in which all the elements are held together by the mutual affinities of the atoms could be called atomic compounds (‘atomic combination’). These are the true chemical molecules and the only ones that can exist in a vapor state. Besides these atomic compounds, we need to distinguish a second category of compounds, which I want to call ‘molecular compounds.’ The existence and formation of these compounds can be explained through the following considerations. Attraction must be noticeable even between atoms belonging to different molecules. This attraction causes the molecules to come closer and align, a phenomenon that always precedes actual chemical decomposition. It can also happen (particularly in cases where double decomposition is made impossible by the very nature of the atoms) that the reaction stops after this approach, so to speak, that the two molecules merge, forming a group with a certain stability, although lower, compared to atomic compounds. This explains why these molecular compounds do not form vapors, but decompose under the influence of heat, while the molecules from which they originated regenerate. Among the connections of this type, I list the following: triatomic elements: PCl3, Cl2; NH3, HCl, etc. Diatomic elements: SeCl2, Cl2; TeBr2, Br2 and the other connections indicated by Naquet, etc. Monoatomic elements: JCl, Cl2, etc.
The formation of compounds with higher chemical valence is explained by the formation of molecular associations. For example, in PCl3 and PCl5, the P atom keeps its atomicity or valence at three, but the equivalence changes in PCl5 because a molecular association forms between PCl3 and Cl2. In opposition to Naquet, he shows how you end up with paradoxical conclusions if you assume atomicity is the maximum saturation capacity of an atom:
Instead of choosing among the different possible values the one that explained things best, that is, in the simplest and most complete way all the combinations, people believed they could define atomicity as the largest equivalent or the maximum saturation capacity. From this arose the need to consider elements, which I had assumed to be triatomic, as pentatomic, like N, P, As, Sb, Bi. A consequence of these same ideas led Mr. Naquet to designate elements ϴ, S, Te, Se, which had so far been considered diatomic, as tetratomic. The same idea should also lead to seeing iodine as triatomic and consequently other elements now considered monoatomic, like chlorine and bromine. If the existence of compounds like NH4Cl, PCl5, etc., shows that nitrogen and phosphorus are pentatomic; if the substances cited by monsieur Naquet (SCl4, SeCl4, TeCl4, TeBr4, TeJ4) establish the tetratomic nature of sulfur, selenium, and tellurium, then one must conclude in the same way (and no one would dispute the correctness of the conclusion) that the existence of a compound JCl3 shows us that iodine is triatomic. Such reasoning cannot be fought with arguments; it disproves itself. In fact, it is enough to assume the triatomic nature of iodine and consider the compounds PJ3 and TeJ4 to be convinced that phosphorus is nonatomic and the atomicity of tellurium is 12. And since chlorine also clearly has the same atomicity as iodine, that is, it is also triatomic, the existence of the compound JCl3 shows that iodine is no longer triatomic but nonatomic, and so on.
In the same issue of the journal, the chemist Alfred Joseph Naquet (1834-1916) countered Kekulé with no less than two articles. Naquet is a figure of great human and scientific interest; among other things, he was a professor of chemistry at the University of Palermo, where he worked alongside Stanislao Cannizzaro from 1863 to 1866. I thought it would be useful to attach the two translated articles because they offer a precise and calm critique of Kekulé’s position. In the first article, On the Atomicity of Elements, Naquet specifically criticizes the molecular compounds introduced by Kekulé to justify the variation of atomic valence in their compounds. For example, we have PCl5, which is represented as PCl3·Cl2, meaning a molecular compound made up of two atomic compounds, held together by a weak attraction compared to that found in atomic compounds. This would explain the tendency of the molecular compound to dissociate when one tries to bring it to a gaseous state. Naquet does not deny the formation of molecular compounds, as happens for example with water of crystallization, but he does not accept it for the compounds cited as examples by Kekulé:
I gladly grant Mr. Kekulé the existence of molecular aggregates; the water of crystallization leaves no doubt about that. But I believe Kekulé is absolutely wrong in assuming that all these compounds [referring to compounds of P, N, Se, Te] are molecular and cannot exist in the gaseous state, and from this error, in my opinion, comes the incomplete definition he gives of atomicity. It is well known that even well-characterized chemical molecules decompose at sufficiently high temperatures. So we shouldn’t be surprised if we see some unstable compounds that cannot even withstand the temperature that would bring them into the gaseous state. This fact can be explained with known laws, so there’s no need to resort to a new hypothesis. If it were otherwise, and if, as Kekulé hypothesizes, the property of not being able to exist in the gaseous state indicated that a compound is just a molecular aggregate, then most of the compounds known so far would have to be considered as not arising from atomic bonding. I do believe, however, that the property of volatilizing is not what allows a distinction between molecular and atomic combinations.
According to Kekulé, when double decomposition reactions between two atomic compounds are not possible (for Kekulé, double decomposition is the fundamental process of chemical reactivity in which two molecules exchange parts to form new stable chemical products), that is, when double exchange reactions between the atoms of the two compounds do not occur, associations can form between the two compounds called molecular, held together by weak forces where, therefore, there are no bonds between the atoms governed by valence rules. According to Naquet, even the associations considered molecular by Kekulé can participate in actual chemical reactions, forming new compounds whose atoms retain their valence.
However, phosphorus pentachloride, ammonium chloride, and the pentachlorides of selenium and tellurium can undergo double decomposition and thus form other compounds belonging to the same degree of combination; they are therefore truly atomistic combinations, showing that nitrogen and its related compounds are pentatomic, and oxygen and its related compounds are tetratomic.
In the second article ‘On the Atomicity of Oxygen, Sulfur, Selenium, and Tellurium,’ Naquet openly advocates for a variable valence of the elements. After criticizing, as we saw in his previous article, the solution proposed by Kekulé for molecular compounds, he supports the variable valence of compounds of oxygen, sulfur, and tellurium. These compounds would have a maximum valence of four, including oxygen. The reasoning that leads him to support valence four for oxygen is interesting. Although no compounds are known in which oxygen shows valence four, only two, he still considers that oxygen has many similarities with sulfur and if it does not show valence four, it is because, even though it potentially has this capacity, it does not express it due to the lack of radicals capable of forming compounds with oxygen at maximum valence.
A rough comparison will make it easier to understand my thinking. Let’s imagine, for example, that atoms have small hook-shaped appendages attached, which serve to latch onto corresponding appendages of other atoms and thus allow the formation of compounds: it’s clear that the number of hooks attached to an atom would represent the absolute atomicity of that atom. If these hooks don’t have the ability to bond indiscriminately with the appendages of other bodies, it’s understandable how the actual or relative atomicity of a radical can sometimes be lower than its absolute or real atomicity. Simply following the ideas expressed by this obvious hypothesis makes it clear how I came to suppose that the apparent atomicity of a body doesn’t always match the real one, and how I also consider it reasonable that, in certain cases, the analogies of a body [with sulfur and tellurium in the case of oxygen] allow its actual saturation capacity to be determined, even if this is greater than the apparent saturation capacity.
Equally firm and balanced is another criticism that, in the same journal, is directed at Kekulé’s position by another prominent chemist, Charles-Adolphe Wurtz (1817-1884). Wurtz’s contribution to this interesting debate is also included in the attached translation. Wurtz’s criticism was not limited to the concept of valence but also contributed to the graphical representation of molecules. He was a great admirer of Kekulé, with whom he organized the first International Chemistry Congress in Karlsruhe in 1860. Both were strong advocates of atomic theory. What separated them was the meaning they attributed to valence. In his article, he critiques the nature of those molecular compounds attributed to Kekulé, while still recognizing hydrated salts as molecular associations. His stance is to support a variable valence. He provides an interesting example showing that valence is variable and that the compound is not molecular. He considers the reaction between iodine trichloride and silver acetate.
If iodine trichloride is a JCI,Cl2 molecular compound, how can it form the following compound when it reacts with 3 molecules of silver acetate?
If two chlorine atoms were simply added to the JCI compound [as Kekulé claimed], it would seem that on the first collision they would detach, and on 2 molecules of silver acetate they would act like 2 free chlorine atoms. What do we actually see instead? The two chlorine atoms are exactly the same as the third, they release 2 C2H3O2 residues, and the 3 acetyl residues remain bonded through the triatomic iodine atom, indivisible.
It lists the platinum compounds where the metal shows valence two and four. It’s remarkable that among the examples it includes Magnus’ salt, showing its structure.
Pt(NH3)2Cl2
The formula is the representation of the salt according to the Type theory [see card 2], with Pt having a valence of two. While in the following salt it assigns a valence of four to platinum (tetrachlorodiamminoplatinum)
PtCl4(NH3)2
It’s notable that among the examples it includes the first platinum coordination compound synthesized by Heinrich Gustav Magnus in 1830. This isn’t surprising because in these compounds metals like platinum showed a highly variable valence. Another example is given with the potassium hexachloroplatinate(IV) [K2(PtCl6)] where PtCl4 can still react with two molecules of KCl. He is aware that he is dealing with a reality that can’t be confined to the rigid framework of fixed valences.
Also, I say: what we need to pay attention to in chemical combinations is the role that elements, particularly polyatomic ones [which show variable valences], play in holding the different parts of a molecule together. To explain the structure of these often so complicated ‘buildings’ [coordination compounds], I didn’t focus on the virtual combining capacity of elements, with any absolute or ideal atomicity, but on the combining capacity that each element actually shows. In a given combination, this capacity manifests itself in a certain way, which I need to understand; it manifests differently in other combinations. So, the atomicity of elements varies depending on the combinations. It grows according to certain proportions, as Dalton’s law shows, and also changes depending on the nature of the compounds.
Coordination compounds will put to the test the fixed valence of an element supported by Kekulé. We’ll see in another post how Kekulé will tackle this additional challenge.
