The process of writing this piece was unusual for me. Generally speaking I am mostly concerned with the sound of a piece and the emotional reaction that it generates. For this piece, since the compositional process was almost algorithmic, with each dimension of music mapped to a particular dimension of the chemical synthesis, the process and primary concerns were much different. For this piece, the process was mostly centered around precompositional decisions surrounding what musical features correspond to what chemical features. The primary concern, then, was simply realizing those decisions as accurately as possible, while still attempting to retain some element of playability for the musicians.

In this post, I wanted to provide a bit more insight into the precompositional decisions that formed this piece. As I mention in the printed program note, the piece is inspired by the structural changes that occur in a molecule during a chemical reaction. So the idea was to have a musical structure that slowly changed and developed over the course of the piece until the “product structure” was reached at the end, in this case, a musical structure that corresponds to D-luciferin.

In this piece, pitch corresponds to molecular structure as determined by hydrogen nmr when possible, and carbon nmr or parent mass spectrometry when necessary (for example, since phosphorous oxychloride lacks hydrogen for h-nmr, and also lacks carbon for c-nmr, the mass spectrum was used to determine the pitch structure).

Below are shown the two h-nmr spectrum for the reactants from the first movement:  p-anisidine and ethyl oxalate respectively. Generally speaking the c-nmr and mass spectrometry data tend to look more or less the same, so these are representative.

p-anisidine h-nmr spectrum

Ethyl oxalate h-nmr spectrum

In order to map these spectra to a pitch collection, I actually just held up an image of a keyboard up to these spectra and marked where the spectrum peaks aligned with the keyboard, rounded to the nearest quarter tone. This is shown below.

p-anisidine h-nmr spectrum mapped to the keyboard

ethyl oxalate h-nmr spectrum mapped to keyboard

These two reactants are combined to form the first product in this synthesis. The h-nmr and associated keyboard mapping of that product is pictured below.

h-nmr of XII

keyboard mapping of XII

Having now determined the pitch structures that represented the two reactants in the first movement, the next step was to determine how these structures should shift and change over the course of the movement to form the first product. This was done through simple interpolation, which I did the old-school way with colored pencils. As you can see in the image below, the p-anisidine pitch structure is shown on the left in orange, the ethyl oxalate pitch structure is shown on the left in purple, and the product is shown in black on the right. Each pitch on the left moves by quarter-tone steps to the closest pitch on the right.

Pitch interpolation of the first movement

Obviously this process is simply repeated through all eight movements until the end product is reached.


Deciding how to approach rhythm in this piece was a challenge. The solution that I arrived at was to use the molecular structure of the solvents used in the various reactions of the synthesis. I think this makes a kind of sense: rhythm and pitch can be thought of as separate (non-interactive) elements in music. One could say that a rhythm is imposed onto a pitch structure. Similarly, the solvents used in these reactions don’t directly contribute to the structural changes that occur throughout the synthesis. The reactants exist within the solvents.

Further, rhythms are made up of individual values of varying sizes. This is also true of molecules, where the composite molecule is made up of individual elements of varying atomic sizes. So, carbon has a molecular weight of 12, and hydrogen has a molecular weight of 1. If a sixteenth note is used to represent hydrogen, then the value that represents carbon would be a dotted half note (a dotted half note is 12 sixteenth notes). This is ultimately how I mapped solvent molecular structures to rhythmic values.

So, in the case of methanol, which is a solvent used in the fourth, fifth, and eighth movements, three sixteenth notes and a dotted half note correspond to the methyl group on the left side of the molecule, and a whole note and sixteenth note correspond to the oxygen-hydrogen bond on the right side of the molecule. This is shown below.


In cases where a solvent contains several different isomers, as in xylene, the three different isomers were each mapped separately and assigned to instruments to create a distribution that represented the distribution of each isomer within the given solvent.

Finally, there are several instances where no solvent was used to dissolve a reactant, or at least none mentioned in the experimental procedure I followed. In these cases, I allowed the performers to improvise a non-period rhythm. The solvents and associated rhythms are shown in the image below.

Solvents and associated rhythms

Other dimensions generally had simper mappings related to the larger scope physical elements of the synthesis. The tempo of each movement maps relatively simply to the temperature of a given portion of the synthesis. The only work that was really done was to find a reasonable maximum and minimum tempo to generate an effective range for tempo across the piece.

The length of each movement is proportionally relative to the length of that step in the synthesis. So, if the overall synthesis takes 10 hours (it’s actually much longer), and one step of it took one hour, that constitutes about 10% of the overall synthesis time, so the corresponding section in the music would constitute about 10% of the length of the piece.

Dynamics are mapped to the volume of the reaction. So, if the size of a reaction is large, the music is played loudly, and small equates to a softer dynamic. This also seemed appropriate since “loudness” can sort of be thought of the “size” of a particular sound. Again, the only process involved in this mapping was what working out the size of a particular reaction step, and finding a dynamic range that seemed reasonable for the music.

Lastly, if a recrystallization step occurred as part of a reaction within the synthesis, the players use extended techniques to brighten their tone, and pause between movements.