Summary: Clearly defining the structure of fracking chemicals is foundational to understanding their chemical properties and hazard profile. We summarize the basic structural classes of FracFocus materials by using SciFinder’s “Substance Classes,” such as “Small Molecule” and “Polymer.” However, we find that more than 20% of FracFocus materials are not classifiable with any structure type, which are used in almost all fracking jobs, making it especially difficult to understand health and environmental hazards.

The materials reported in FracFocus are a motley crew. Some are very well-defined and well-understood by their structure; others are defined by the process that generates them or their structure is only loosely known. The SciFinder¹ chemical database tool provides one or more Substance Classes for all materials with a CASRN. These classes help us understand not only the types of materials they are but sometimes also how well defined they are.

Currently, there are about 1,400 distinct, CAS-registered chemicals in FracFocus. One first cut to understanding them is to understand the meaning of the substances classes to which they belong.

“Element” class

Members of the “Elements” class are the simplest pure substances and cannot be broken down further by chemical means². These are the entities you find on a periodic table.

This is not a large substance class in FracFocus (less than 2% of reported chemicals³). Nitrogen (used as a foamed gas) is the most used member of this class. Some elements reported, such as Iridium and Antimony, were used as tracers and appear rarely.

“Small Molecule” class

The CAS Substance Class “Small Molecule” is probably what you think of when you hear “chemical.” These compounds – a combination of two or more elements – are bonded covalently⁴. All molecules are of identical structure⁵. Thus, a quantity of something from this class is a homogenous collection of identical particles. This is the biggest CAS Substance Class and represents about one third of FracFocus substances. Examples include:

The health and environmental affects of these chemicals are often better understood than other substance classes. The members of this large substance class play a diverse set of roles in the fracking process, and are involved in just about every stage and function.

“Salt And Compound With” class

This CAS Substance Class, that goes by the name “Salt And Compound With,” refers to substances formed by the combination of different chemical entities, usually through ionic or other non-covalent interaction. These entities or components, taken individually, have separate CASRN. Typically, at least one of the components is from the “small molecule” substance class. The components are often in some defined ratio. For example, the compound “Ammonium acetate” (CASRN: 631-61-8) is composed of “acetic acid” (64-19-7) and the “ammonium ion” (14798-03-9), in the ratio of 1:1.

This substance class represents about 20% of FracFocus reported chemicals. Other examples include:

While FracFocus does not indicate individual components, it can be important to know them. Non-covalent bonds⁷, which are typical in this substance class, hold the components together as long as this substance is isolated. However, the components are likely to dissociate in a mixture, especially in solution or under conditions like increased temperature, dilution, pressure, or presence of other compounds. All of these are almost certain conditions in a fracking job.

Therefore, when assessing the hazardous profile of a mixture containing materials in this substance class, it is crucial to consider individual components, not just the complex substance itself. This is because the components, when examined independently, often have characteristics unlike the reported substance. For instance, the hazard profile for the substance shown above, 38172-91-7, is characterized by the EPA’s ChemInformatics⁸ as mostly unknown, whereas its components are better understood and quite hazardous:

“Polymer” class

The FracFocus materials that are classified by the “Polymer” CAS Substance Class are quite diverse. While the most commonly used material (Guar Gum CASRN 9000-30-0) is from a natural source, the vast majority of polymers in FracFocus are synthesized. Among all FracFocus reported materials, almost 20% are polymers and represent over 570,000 records.

Like the “Salt and Compound With” materials, polymers are a complex of components but their form is different. Polymers are large molecules made up of repeating structural units, called monomers. And instead of a loose collection of components, monomers are connected by a covalent bond, typically making the polymer a very stable molecule.

Those monomers may be composed of a single “small molecule” (for example: Polyvinylidene chloride CASRN: 9002-85-1) or by many (2-Propenoic acid, polymer with 4-(1,1-dimethylethyl)phenol, formaldehyde, 2,5-furandione, 2-methyloxirane, 4-nonylphenol and oxirane CASRN: 129828-31-5; 7 components in each monomer). Furthermore, individual molecules of a polymer often differ in how many of monomers are linked together, ranging from just a handful to millions.

Some other examples include⁹:

Polymers may also have other non-structural Substances Classes, as discussed below.

“Mineral” and “Tabular Inorganic” classes

In the two Substance Classes named “Mineral” and “Tabular inorganic” we find 25 FracFocus materials (< 2%). The most common proppant, sand (or more specifically, Quartz (SiO2) CASRN 14808-60-7), is the most dominant member of these classes from FracFocus (over 250,000 records). Besides that obvious player, we also find these materials:

Ulexite	1319-33-1	“Mineral”; 12,000 records
Talc (magnesium silicate hydrate)	14807-96-6	“Mineral”; 6,400 records
Cristobalite (crystalline silica)	14464-46-1	“Mineral”; 5,700 records
Biotite	1302-27-8	“Mineral” and “Tabular Inorganic”; 4,100 records
Goethite (Fe(OH)O)	1310-14-1	“Mineral”; 3,800 records
Attapulgite (hydrated aluminum-magnesium silicate (palygorskite))	12174-11-7	“Mineral”; 2,900 records
Illite	12173-60-3	“Mineral” and “Tabular Inorganic”; 2,900 records
Boron potassium oxide (B4K2O7)	1332-77-0	“Tabular Inorganic”; 2,900 records
Kyanite (Al2O(SiO4)) (aluminum silicate)	1302-76-7	“Mineral” and “Tabular Inorganic”; 2,300 records

“Coordination Compound”

The CAS Substance Class “Coordination Compound” refers to a unique set of compounds with a central metal atom bonded to nonmetallic ions or molecules (ligands). These compounds often serve as catalysts of chemical reactions. There are 13 representatives in FracFocus in fewer than 2,000 records. Some examples:

“Ring Parent”

The “Ring Parent” substance class refers to cyclic (ring-forming) compounds that serve as the foundation for a whole family of related substances. There are 10 such materials in FracFocus and are simultaneously classified as “Small Molecule.” A few examples are:

“Mixture”

The formal “Mixture” substance class is rarely found in FracFocus (only 2 materials). By definition, the class encompasses substances that are made up of two or more different components physically mixed together, without being chemically bonded.

Other structural classes

SciFinder assigns other structural classes to materials that have not been reported in FracFocus. They are:

Alloy	A mixture of metals or a metal and another element (e.g., bronze, steel).
Nucleic Acid Sequence	A sequence of nucleotides that make up DNA or RNA, carrying genetic information.
Protein/Peptide Sequence	A sequence of amino acids that make up a protein or peptide, with various biological functions.
Radical Ion	An atom or molecule with an unpaired electron, often highly reactive.
Subatomic Particle	A particle smaller than an atom, such as a proton, neutron, or electron.
Isotopically Labeled Substance	A substance containing one or more atoms replaced with isotopes, used in research and medical applications.
Stereoisomer	Compounds with the same molecular formula and connectivity but different spatial arrangements of atoms, leading to varying properties.

Distribution of structural classes

Across all reported materials in FracFocus, 345 (about 23%) are not assigned a structural class by the Chemical Abstract Service. These materials are assigned other Substance Classes (“Non-structural Classes” below) that indicate the reasons; mostly it is because the materials are so loosely or incompletely defined that the components cannot be pinned down. The Toxics Substance Control Act has a category that fits many of these materials: “Unknown, Variable, Complex reaction products or Biological material” (UVCB). Understanding the hazard profile of these materials is a thorny problem¹⁰.

More than 20% of materials in FracFocus¹¹ are not assigned to any of the structural classes above.

---

Non-structural classes

Besides indicating the type of chemical structure a material has, a Substance Class can also indicate other information that the Chemical Abstract Service deems important. We highlight some of those here, especially those that reflect on the complexity or “unknown” aspects of the material.

“Manual Registration”

The “Manual Registration” Substance Class indicates that the material was registered manually by experts at CAS. This usually happens when a substance has unique characteristics or complexities that make it difficult to categorize automatically. That might be due to unusual bonding or complex structures that defy typical rules of classification. It is also used to label materials with limited information.

More than 900,000 FracFocus records involve this class.

The number of “Manual Registration”-identified CASRN found along with structural classes:

Small Molecule	Element	Salt and Compound With	Polymer	Mineral or Tabular Inorganic	NOT CLASSIFIED
0	0	0	97	4	302

“Generic Registration”

The “Generic Registration” Substance Class indicates materials with poorly defined structures or variable compositions that don’t fit the usual criteria for a standard CASRN. An example would be mixtures or polymers with ill-defined structures.

The number of “Generic Registration”-identified CASRN found along with structural classes:

Small Molecule	Element	Salt and Compound With	Polymer	Mineral or Tabular Inorganic	NOT CLASSIFIED
0	0	0	35	0	62

“Incompletely Defined”

This class encompasses substances where the chemical structure or composition is not fully known or cannot be precisely defined. This lack of complete information can arise from:

• Unknown connectivity: The exact arrangement of atoms and bonds in the molecule might not be fully known. This could be due to limitations in analytical techniques or the complexity of the substance.
• Variable composition: The substance might be a mixture of compounds with varying proportions or have a structure that cannot be represented by a single, well-defined molecular formula.
• Unknown substituents or positions: The substance might have known core structure but with unknown or variable substituents or attachment points.

A simple example is “Xylenes” (CASRN: 1330-20-7). The relative position of the two methyl groups (CH₃) on the ring can be one of three places (indicated by the grayed area on the ring).

The number of “Incompletely Defined”-identified CASRN found along with structural classes:

Small Molecule	Element	Salt and Compound With	Polymer	Mineral or Tabular Inorganic	NOT CLASSIFIED
0	0	0	9	0	37

“Registered Concept”

The class “Registered Concept” doesn’t represent a specific type of substance with a defined structure. Instead, it acts as a broader category for concepts, ideas, and groupings of substances. As such, this class indicates that a material is not well defined – it may contain a number of different kinds of components that satisfy the “concept” but nevertheless are unique and, often, not known or are variable in some respect. Refinery products commonly have this classification as they are defined not by what is in the material, but by the process by which they were manufactured¹². It is common to find the TSCA designation of “UVCB” in this substance class.

Some common FracFocus materials in this substance class:

Distillates (petroleum), hydrotreated light	64742-47-8	170,000 records, the biggest 10% used at least 60,000 pounds.
Alcohols, C12-16, ethoxylated	68551-12-2	49,000 records, the biggest 10% used at least 5,500 pounds.
C12-16-Alkylbenzyl-dimethylammonium chlorides	68424-85-1	49,000 records, the biggest 10% used at least 4,000 pounds.
Solvent naphtha, petroleum, heavy arom.	64742-94-5	26,000 records, the biggest 10% used at least 9,000 pounds.
Tar bases, quinoline derivatives, benzyl chloride-quaternized	72480-70-7	17,000 records, the biggest 10% used at least 275 pounds.
Bentonite	1302-78-9	3,800 records, the biggest 10% used at least 15,000 pounds.

The number of “Registered Concept”-identified CASRN found along with structural classes:

Small Molecule	Element	Salt and Compound With	Polymer	Mineral or Tabular Inorganic	NOT CLASSIFIED
0	0	0	47	0	207

“General Derivative”

The Substance Class “General Derivative” is used when a substance is derived from a known compound but has some modifications or variations that make it distinct. Almost all of 27 FracFocus materials with this classifications are also classified as “Polymer.” Some examples are:

Edifas B	9004-32-4	4,900 records and the biggest 10% were at least 3,600 pounds.
Ethoxylated, propoxylated trimethylolpropane	52624-57-4	1,600 records and the biggest 10% were at least 870 pounds.
Carboxymethyl hydroxypropyl guar	68130-15-4	400 records and the biggest 10% were at least 165,000 pounds.

Small Molecule	Element	Salt and Compound With	Polymer	Mineral or Tabular Inorganic	NOT CLASSIFIED
0	0	0	25	0	2

---

Other indicators of status

“Unspecified” molecular formula

Additional evidence that a material is not well defined is the term “Unspecified” in the “Molecular Formula” field of SciFinder (not the Substance Class). Many well-defined materials are variable for some parameter, such as the number of monomers in a polymer.

However, the term “Unspecified” for the entire molecular formula is a signal of significant lack of information about the material’s chemical makeup. Some reasons we’ve found for a completely “Unspecified” formula:

• Truly unknown composition: The material might be a complex mixture or a poorly characterized material where the individual components and their amounts are not known. This is may be the case for many natural products, materials extracted from plants, animals, or microorganisms with complex and variable compositions. It may also be the case for industrial mixtures; such as formulations or byproducts with numerous components that haven’t been fully analyzed.
• Confidential information: In some cases, the molecular formula might be withheld intentionally to protect proprietary information or trade secrets. Unfortunately, the confidential status is not usually available in the SciFinder information, so we cannot tell.

In both cases, our ability to understand important characteristics such as the hazard profile is blocked.

The “Unspecified” label is pretty common among FracFocus materials. There are almost 350 materials with the entire molecular formula given by “Unspecified.” Another 40 more materials have at least one component that is labeled as “unspecified.” When they have been given a structural class, it is almost always “Polymer.”

Conclusion

In the context of FracFocus, the Substance Class of SciFinder gives us two important bits of information: first, it gives us the structural nature of the material (small molecule, polymer, etc.) when it can be sufficiently described. Some structural classes suggest that we need to look deeper for hazard information (such as examining the components of materials in the “Salt and Compound With” class as well as the reported CASRN).

Secondly, the Substance Class flags for us when a material’s composition is ambiguous enough that experts at CAS cannot assign a structure. This latter condition happens for 20% of the materials reported in FracFocus, many of which are commonly used. This translates into over 95% of disclosures reporting materials that are not structurally classifiable. The health and environmental impacts of these materials are harder to characterize, and in many cases, are essentially unstudied.

---

Title image by B. Mansfield (2025); cropped by author. The molecular images are courtesy of EPA’s CompTox or CAS’s SciFinder.

1. SciFinder is a widely-used, authoritative, online chemistry tool, produced by the Chemistry Abstract Service. It is a pay subscription service. CAS also publishes some of the SciFinder material at the free service, Common Chemistry. ↩︎
2. However, this substance class can include compounds of a single element, such as ozone which is a covalently-bonded molecule of three oxygen atoms. ↩︎
3. The statistics on this page about the frequency of FracFocus materials and records were calculated from a FracFocus data set downloaded in December 2024. ↩︎
4. A covalent bond involves the sharing of electrons to form electron pairs between atoms. This sharing forms a balance of the attractive and repulsive forces between atoms, typically forming a stable, strong bond compared to other types of bonds. ↩︎
5. While this class is composed mostly of compounds covalently bonded, it also includes some simpler salts (ionically bonded) when they are 1) a single cation and anion, 2) no solvent molecules are in the crystal structure and 3) they are relatively low molecular weight. An example is table salt (NaCl). ↩︎
6. Across all FracFocus disclosures, the use of any given chemical can range from less than one pound to over a hundred thousand pounds, so looking at the biggest 10% just indicates the upper end of that range. ↩︎
7. Such as ionic bonds and hydrogen bonds. ↩︎
8. See this Open-FF page for an explanation of the graphic. ↩︎
9. In the molecule image, the brackets with the subscript “n” indicates the position and number of the monomers ↩︎
10. See: The Next Frontier of Environmental Unknowns: Substances of Unknown or Variable Composition, Complex Reaction Products, or Biological Materials (UVCBs). Lai et al., 2022 ↩︎
11. 345 individual CASRN, as of Jan 2024. ↩︎
12. For example, for the common material 64742-94-5, the SciFinder notes indicate the refinery parameters: “A complex combination of hydrocarbons obtained from distillation of aromatic streams. It consists predominantly of aromatic hydrocarbons having carbon numbers predominantly in the range of C₉ through C₁₆ and boiling in the range of approximately 165°C to 290°C (330°F to 554°F).” ↩︎