In June, EPA published a Federal Register notice that included Significant New Use Rules (SNURs) for two carbon nanotubes (as well as 21 other chemicals).  That notice certainly got the attention of lawyers in town (see here, here and here).  The nanotube SNURs would require anyone planning to produce or process either of the two substances to notify EPA if the person intended not to comply with the (rather limited) risk management conditions specified by EPA.  Well, as reported yesterday by Sara Goodman of E&E News, EPA is now withdrawing the SNURs, at least temporarily.   

The withdrawal notice is posted here.  On one level, the withdrawal is based on a technicality.  EPA had issued the SNURs as part of what is called a "direct final rule," a mechanism EPA can and typically does use as a short-cut to get around having to go through lengthy full notice-and-comment rulemaking (see p. 31299 of this 1989 Federal Register notice).  Such a rule applies immediately upon issuance – unless someone files, within 30 days, a notice of intent to submit "adverse or critical comments."

Surprise, surprise, someone did just that.  One of those DC lawyers, James Votaw of the firm Wilmer, Hale, had the notice hand-delivered to EPA just 2 days before the deadline, "on behalf of one or more clients."  (Hey, this is a regulation, after all, so it's just begging to be challenged.  Besides, things are kinda slow in DC during the dog days of August.) 

Votaw's notice of intent is posted in an EPA docket you can access here.

What's EPA's next move?

Based on EPA's withdrawal notice, EPA "intends to publish in the Federal Register, under separate notice and comment rulemaking procedures, proposed SNURs" for the two nanotubes.  It would then presumably proceed to issue a final rule containing the SNURs at some point.

It's unclear from this whether EPA intends now to proceed via full notice-and-comment rulemaking, or to reissue the SNURs via an "interim final rule" (again, see p. 31299 of this 1989 Federal Register notice).  The latter takes effect on the date of publication, even as public comment is taken and considered.  As long as EPA promulgates a final rule within 180 days, the SNUR remains in effect during the interim.

The "interim final rule" approach may well be more advisable in this case.  Otherwise, some devious DC lawyer could advise his or her clients that, if they hurry, they could engage in the very activity for which EPA intended to require notification – but without ever having to tell EPA – simply by starting the activity identified as a "significant new use" in the proposed rule before the rule goes final.  Such a client could be a downstream processor or another manufacturer of the nanotubes, as the SNURs would apply to both.

This loophole speaks more generally to the major limitations facing EPA in trying to use SNURs to regulate new chemicals or new uses of existing chemicals.  By definition, any activity already ongoing at the time a SNUR is issued cannot be considered a "new use" and hence is beyond the reach of a SNUR – it can only require notification for activities not already occurring.

Still with me?

A red herring

What Sara Goodman reports as Mr. Votaw's main problem with the nanotube SNURs is that EPA did not sufficiently identify the specific carbon nanotubes to which they apply.  Indeed, the SNURs themselves refer only to a "generic" name for each nanomaterial. 

Well, that's standard practice under the Toxic Substances Control Act (TSCA).  The specific identities of the nanotubes are claimed as confidential business information (CBI) by the original submitters of the corresponding pre-manufacturing notifications (PMNs), so EPA is barred from revealing them publicly and must use a generic name.  (EPA's use of the generic name apparently confused enough folks that EPA felt obliged to email around a clarification that the SNURs do indeed apply only to the very specific nanotubes made by the companies that submitted the original PMNs.  Makers of any other nanotubes would still need to file their own PMNs.)

Mr. Votaw and his clients surely know this.  They also surely know that, if his clients really want to know whether the SNURs in question apply to a nanotube they intend to produce, there's an app for that:  They are to submit to EPA what's called a "Bona Fide Intent to Manufacture or Import Notice."  EPA then determines whether the specific nanotube the client proposes to make is or is not the same as that to which the SNUR applies, and informs them of the determination.

All of that so as not to reveal the confidential identity of the nanotubes that started all this.  (In case you're wondering, as ordinary folk, you and I can't file a bona fide request, so there's no way for the public to participate in this ritual.)

The bigger problem

Notwithstanding the unfounded basis for Mr. Votaw's concern, in light of the standard remedy EPA has provided, the fact remains that EPA lacks a nomenclature system that can distinguish between different nanotubes (or any other classes of nanomaterials, for that matter).  For now, EPA has indicated it will identify nanotubes based on what company produced them (among other factors).  So it may consider even two apparently identical nanotubes made by different companies to be different. 

That means that future SNUR development and any evaluation of potential risks will be done on a case-by-case basis for each nanotube.  That's good on one level, in that their properties may well be quite specific; bad on another level, in that any SNUR will apply only to a very specific nanotube and hence does not provide a viable avenue to require notification or to regulate nanotubes more generally.

This episode also vividly illustrates how cumbersome chemical regulation is under TSCA:

• Even to require notification via a SNUR, EPA must go through a rulemaking – and if anyone objects, a notice-and-comment rulemaking – for each and every case.
• A SNUR cannot require notification by any companies who maintain they are complying with the conditions of the SNUR, naturally raising compliance questions.
• A SNUR cannot reach any activities associated with a chemical that are already underway, because by definition they are not a "new use."
• A SNUR does not regulate a chemical's production or use; it only requires notification of EPA and provides an opportunity for an EPA review.  Any regulation would require EPA to demonstrate "unreasonable risk" and promulgate a separate rule under TSCA's Section 5(f) (for a new chemical) or Section 6 (for an existing chemical), a task that has proven virtually impossible in practice.
• In the absence of a SNUR, any company can produce and use any chemical on the TSCA Inventory under any conditions it chooses, without having to notify EPA it is doing so.

Solutions

This convoluted situation is one of many aspects of current chemical regulatory practices in the U.S. that is motivating calls for major reform of TSCA.  EDF is working within the Safer Chemicals, Healthy Families coalition to achieve this aim. 

Among the changes we propose that are relevant in the current context are:

• Requiring that all producers of a chemical (including a nanomaterial) – whether new or existing – identify themselves to EPA and provide basic safety information.
• Requiring that any significant change in a company's production or use of a chemical automatically trigger both EPA notification and an update safety review.
• Extending the definition of specific chemical identity to include physical as well as chemical characteristics of a substance, to ensure EPA can distinguish among nanomaterials based on more than just their underlying chemical structures.
• Limiting the ability of companies to claim a chemical's identity to be confidential in association with any information regarding that chemical's safety.

In earlier posts (here and here), I discussed a notice EPA had received in July of 2008 from BASF reporting toxic effects at very low doses of a carbon nanotube (CNT) observed in a 90-day rat inhalation study.  In that notice, BASF had declared the specific identity of its CNT to be confidential business information, hence denying that information to the public.  Now, in a setting more to its liking, it appears the company has decided to reveal the identity after all.

The original notice was submitted by BASF as required under Section 8(e) of the Toxic Substances Control Act (TSCA).  That provision requires a company that makes a chemical to notify EPA within 30 days if it obtains new information that "reasonably supports the conclusion that such substance or mixture presents a substantial risk of injury to health or the environment."

Identity hidden

BASF complied – but in doing so, claimed the identity of the nanomaterial in question to be confidential business information (CBI), requesting that it be identified only generically as "Carbon Nano Tube."  No way, therefore, for the public to tell whether it was single- or multi-walled or much of anything else about its actual identity or structure.

Identity revealed

Now, just last week BASF published a paper in the journal Toxicological Sciences that is almost certainly based on the same study.  (The abstract is publicly available, but the full paper requires a subscription.)  All of the details – the doses, number and type of animals, exposure conditions, and observed effects – match those reported in the Section 8(e) notice.

But in this setting BASF chose to fully describe the identity of its CNT, disclosing for the first time that it is multi-walled and describing other structural features in detail.

As I reported before, the study found this nanomaterial to be highly toxic, causing lung inflammation and granulomas at doses 200-fold lower than the high-concern level identified under EPA and international standards.  That makes it at least an order of magnitude more toxic than crystalline silica and – as BASF itself describes in the published paper – also at least 10-fold more toxic than nano-structured carbon black.

A major effort has been mounted in the nanotechnology community to demand that researchers fully identify and characterize their nanomaterials when publishing papers in the peer-reviewed literature, and that journals accept only such papers.  See, for example, the "Minimum Information for Nanomaterial Characterization Initiative, or Characterization Matters for short.

That likely explains why BASF provides such a full identification and characterization in its recent paper.

So why, then, did BASF claim the identity of its nanomaterial to be confidential when submitting the same study to EPA?

The answer is simple:  Because it could.

Right-to-know under TSCA:  Health impacts - yes; chemical identity - no

EPA routinely posts notices it receives under Section 8(e) in which the chemical identity, and often the submitter's identity, are claimed confidential and hence not publicly disclosed.  It's hard to imagine a less useful – and more frustrating – public disclosure system:  Some chemical is found to cause serious health or environmental effects, serious enough for TSCA to require immediate submission to EPA.  As a member of the public, you get to see just how bad the effects are – but are left to guess just what chemical causes the effects!

The ultimate insult is that, under TSCA, EPA actually should be prohibiting companies from declaring confidential the identity of a chemical that is the subject of a submitted health and safety study. 

Consider:

• TSCA itself does not preclude – and hence requires – disclosure of health and safety studies; see TSCA Section 2613(b).
• EPA regulations clearly define chemical identity to be an integral part of a health and safety study; see the definition of a health and safety study at 40 CFR §716.3 and §720.3(k).
• EPA regulations then do provide certain conditions under which a company may assert a confidentiality claim for the identity of a new chemical even when associated with a health and safety study. However, the regulations further state that EPA will deny such a claim unless the claimant demonstrates that "the specific chemical identity is not necessary to interpret a health and safety study." See 40 CFR §720.90(c)(3).

How could anyone reasonably argue that knowing the specific identity of a carbon nanotube is not critical to interpreting a health and safety study?  Given, for example, recent findings that multi-walled CNTs of certain lengths, but apparently not single-walled CNTs of any length, behave like asbestos, knowing a nanomaterial's identity is central to interpreting health and safety data.

Yet EPA has taken no apparent action to preclude such claims for nanomaterials.  Indeed, four of the eight TSCA Section 8(e) notices EPA has received for nanomaterials that I discussed in my earlier post had masked the materials' identities.

One possible reason why EPA hasn't acted?  Lack of EPA resources to review and challenge the thousands of CBI claims asserted annually by industry. 

Given the ease with which TSCA allows CBI claims to be asserted, there is every reason to expect that many such claims wouldn't pass muster if actually examined.  A 1992 EPA study identified extensive problems with respect to the extent of inappropriate CBI claims; see pp. 32-33 of this 2005 report from the Government Accountability Office.

EPA can, of course, challenge CBI designations on a case-by-case basis, but it rarely does so because of the extensive resources required.  In the absence of a successful challenge by EPA, the information must be held as confidential.

This is a nano problem that requires a macro solution:  Fundamental reform of the CBI provisions of TSCA.  That's one of ten essential elements in TSCA reform that I discuss at length in this recent paper I published in Environmental Law Reporter.

Some months ago, my colleague John Balbus posted here about studies finding that when multi-walled carbon nanotubes (MWCNTs) are injected into the abdominal cavities of mice, they induce inflammation and mesothelioma-like reactions similar to those caused by asbestos.  He appropriately cautioned that – among other critical questions – these studies had not demonstrated that inhaled MWCNTs could actually move out of the lung and into the tissues where asbestos gives rise to its effects.  Well, that particular dot now appears to have been connected.

We learned about the new findings via a blog post by Liz Borkowski at The Pump Handle.  She noted a disturbing item on the NIOSH blog posted by Vince Castranova and his colleagues late last week, in which they are seeking to share more broadly results they first presented at the recently-concluded Society of Toxicology meeting in Baltimore.

The NIOSH researchers reported new data showing for the first time that MWCNTs can migrate intact from the alveoli out of the lungs of mice and into the pleura, the tissue surrounding the lungs.  And it is in the pleura (as well as the abdominal cavity) where asbestos induces its signature form of cancer, mesothelioma.

In this case, the MWCNTs were introduced into the lungs using pharyngeal aspiration, a procedure by which mice are induced to inhale a droplet of liquid in which the MWCNTs are suspended.  While this procedure is thought to mimic direct inhalation, the NIOSH researchers note this and other limitations of the study, and caution that the results are preliminary and have not yet been peer-reviewed.

They also note that it's possible that the mice used in the study are unique and may not accurately portray what would happen in people, say, workers exposed to MWCNTs.  And, as my colleague pointed out in his earlier post, whether sufficient material could or would be suspended in the air to result in inhalation exposure also remains an open question.

Nonetheless, these new findings strongly suggest that, like asbestos, MWCNTs behave as stable fibers capable of penetrating and migrating through the lung.  And together with the earlier studies showing that introducing MWCNTs into the tissues surrounding the lung induces mesothelioma-like reactions, it's fair to say the alert level on MWCNTs just went up significantly.

It had to happen sooner or later. After several years spent by the UK and US governments conceptualizing, vetting, proposing, again vetting, developing, yet again vetting, and finally launching and reporting on their voluntary reporting programs for engineered nanoscale materials – only to have them largely spurned by the intended targets – other governments observing all this have decided that mandatory approaches are needed.

The UK was the first to try the voluntary approach.  Its Voluntary Reporting Scheme was launched in September 2006 and ran for two years.  It closed out in September 2008 after attracting only 11 submissions. 

It took the US another 16 months to follow suit with its voluntary Nanoscale Materials Stewardship Program (NMSP), launched in January 2008 and also intended to run for at least two years.  EPA's interim report, which evaluated the submissions it received from 21 companies in the first seven months, was released last month.  That count has increased now to 29 companies, more than the UK program attracted but still a small fraction of the companies engaged in nanomaterial activities in the US (I blogged about the report's findings earlier).

With many more companies and nanomaterials in or about to enter commerce than have been captured by these voluntary efforts, and the inability of government to judge how complete or selective the data it is receiving actually are, it appears other governments are opting for mandatory reporting.

First comes France, which, as reported by SafeNano, proposed new legislative amendments in early January that would require that "those who manufacture, import or place on the market nanoparticulate substances periodically report to the administrative authorities the identity, quantity and uses of these substances.  In addition, a further note declares that information on identity and use of substances should be made available to the public, except if doing so would be potentially damaging to national defense."

Then comes California (which is not at all unaccustomed to acting as if it were a national government!).  Its Department of Toxic Substances Control (DTSC) posted a notice on its website and sent a letter on January 22 to several dozen companies, universities and research facilities that produce in or import into the state carbon nanotubes (CNTs).  DTSC's "information call-in" for CNTs requires submission, within one year, of data regarding "analytical test methods [for sampling, detection and measurement of CNTs in workplaces and the environment], fate and transport in the environment, and other relevant information."

And now it appears Canada is poised to issue in mid-February a Ministerial Notice requiring mandatory reporting of information relating to nanoscale materials.  Canada has been planning to issue such a notice for some time, based on information it submitted for inclusion in a June 2008 Tour de Table document (see Canada section) recently made public.  That document summarizes nanotechnology safety-related activities of various OECD countries participating in the Working Party on Manufactured Nanomaterials (WPMN) of the Organisation for Economic Cooperation and Development (OECD). 

As reported in the Daily Environment Report on January 26 (subscription required), a spokeswoman for Environment Canada said companies would have to report "basic information on the quantity of the substance that is manufactured or imported, details about the use of the substance, and any data on physical and chemical properties, toxicity data currently available to respondents, and information on the procedures, policies and technological solutions that have been put in place to prevent or minimize releases of the substance to the environment and exposure to individuals."  The reporting requirement will apply to nanomaterials produced or imported in amounts exceeding one kilogram in 2008, and the requested information is to be submitted within four months.

With even EPA having hinted in its interim report on the NMSP that it will now look seriously into developing mandatory reporting and testing rules, it appears that the "voluntary phase" of government policy toward nanomaterials may have run its course.

In some nanotechnology circles, it is almost a mantra that, once released to the environment, nanoparticles will inevitably aggregate or agglomerate into larger masses and thereby lose their nanoscale-related properties and, by implication at least, any associated risks.

But can we count on nanoparticles released to the environment to self-regulate their own risk so conveniently?

For example, the National Nanotechnology Initiative prominently features on its website an article it commissioned, titled "Understanding Risk Assessment of Engineered Nanomaterials:  How can we know what is a risk and what is not?"  In a section designed to lead readers to question published studies that suggest nanomaterials might pose risks, the article says: "In solution or in air, it's quite difficult to keep nanomaterials separate, as they tend to clump in larger aggregates or agglomerates."  This is a point the author of this rather short article felt compelled to repeat twice more.

An FAQ issued by Germany's Federal Institute for Risk Assessment states that "nanoparticles tend to aggregate into larger unions which are generally larger than 100 nm. The toxic effects of nanoparticles linked to their small size and higher reactivity are then no longer relevant."

And a recent post on the blog of the nanotechnology practice group at Porter and Wright asserts that nanoparticles "have been shown to have fewer potential adverse health effects when they occur in cluster form (aggregates and/or agglomerates).  In the 'good news' department, scientists studying aerosol dispersion of nanoparticles have found they tend to cling together when dispersed into the environment." 

I won't even begin to try to lay out here how much more complex and unpredictable than this nanoparticle aggregation and environmental fate and transport are in the real-world.  Instead, let me just cite two excellent papers that do so:  see Maynard and Weisner et al.

But I do want to briefly discuss and cite some recent studies supporting three reasons why we can't count on nanoparticles released to the environment to self-regulate their own risk so conveniently: 

1.  Some nanomaterials can be stabilized as nanoscale particles in solution under environmental conditions.  A number of studies have found that carbon-based nanomaterials – despite their inherently very low water solubility – can be "solubilized," that is, can enter and remain in stable suspensions upon interaction with water or with other common, naturally occurring substances.  The latest study, authored by Salonen et al. and published in the journal Small, finds that C₇₀ fullerenes can form "stable, homogeneous suspensions" in water through interaction with ubiquitous phenolic acids that are present in and released from virtually all plant matter.  It appears that individual C₇₀ fullerenes first become coated with the phenolic acid, and then form small, loose clusters with diameters on the order of a few nanometers.  This study merited a "spotlight" on the Nanowerk website.

Earlier work has found similar behavior:  Fortner et al. identified the formation of stable suspensions of "nanocrystals" of C₆₀ fullerenes in water – nanoscale (25-500 nm diameter) aggregates they call "nano-C₆₀" that have entirely shed the extreme hydrophobicity of the individual fullerenes.  Hyung et al. found that multi-walled carbon nanotubes could be stabilized as individual particles through interaction with natural organic matter found in river waters; the natural material actually worked better than commonly used surfactants selected to serve that same solubilizing function.

2.  For performance reasons, nanoparticles are being actively engineered not to clump.  For most nanomaterial applications, optimal performance depends on minimizing any disordered clumping or even maximizing dispersal, so that the properties of individual nanoparticles or highly ordered nanostructures can fully exert themselves.  For this reason, researchers are working overtime to coat, cap, chemically modify or otherwise force nanoparticles not to aggregate or agglomerate.  See, for example, Yang et al.'s use of special capping agents to prevent aggregation of platinum nanoparticles, and Nadagouda and Varma's similar work with silver and palladium nanoparticles.  Similar efforts have been mounted to chemically modify ceramic nanoparticles to ensure dispersal.

So even to the extent that native or current forms of nanoparticles do readily clump or retain their hydrophobicity, any assumption that engineered nanomaterials entering commerce and the environment will inevitably do so is wholly unwarranted.

3.  Even agglomerated or clumped nanoparticles can be toxic.  The assumption that aggregated nanoparticles lose all of their nanoscale properties or become benign is also unwarranted.  Maynard and Kuempel have amply demonstrated that even large aggregates on individual nanoparticles typically retain many of their nanostructural features and properties.  But what about toxicity?

Fortner et al. found that their fullerene nanocrystals exhibited antimicrobial activity, suppressing bacterial growth and respiration.  In addition to confirming nano-C₆₀'s antibacterial activity, a recent paper by Lyon and Alvarez cited a number of studies demonstrating that the formation of these nanoscale aggregates in water yields a material with high toxicity to aquatic invertebrates, fish and the cells of higher organisms.  The aggregates have also been shown to enter and accumulate in those cells and to adhere to lipids.

Finally, Salonen et al. showed that their phenolic acid-coated C₇₀ clusters could readily translocate across the membranes of human cells in culture and enter the membrane surrounding the cell nucleus.  Moreover, they induced the contraction and ultimate death of those cells – ironically, apparently by aggregating into micro-sized particles through interaction with the cell membranes.

Once again, we find that nanomaterials' actual behavior confounds conventional wisdom and, when approaching their toxicology, forces us to question or abandon our assumptions and biases.

Can nanoparticles get into our drinking water and if so, what's the harm?

Nanoparticles are being used in cosmetics and other personal care products with increasing frequency.  Carbon fullerenes, also known as buckyballs, have recently been touted as imparting age-defying antioxidant benefits when added to skin cream.  And there are some studies that seem to support these claims.  But even if such claimed benefits turn out to be true, this is by no means the end of the story. 

Skin creams are eventually washed off and down the drain.  Once they enter a sewage treatment system, the fate of nanoparticles is largely unknown.  Depending on their physical and chemical properties, engineered nanoparticles may wind up in the sludge – the solids – or they could spill over into the wastewater discharge, and wind up in lakes, rivers and streams.  If the latter occurs, they could end up in our drinking water.

This is not wild speculation.  One of the most hotly debated issues in environmental science is the frequent detection of chemicals from pharmaceuticals and personal care products in our drinking water.

If fullerenes act as antioxidants, then what's the big deal if they get into our drinking water — wouldn't they be beneficial? The very properties that allow antioxidants to scavenge the reactive oxygen species (ROS) that can damage cells also allow these compounds to liberate them, by cycling between uptake and release.  Under certain circumstances, antioxidants are more likely to release ROS than to take them up.  More research is needed to understand under what conditions so-called antioxidants might be likely to contribute to rather than prevent cell damage.

A new study published in Environmental Health Perspectives evaluated the effect of a single oral dose of fullerenes and single-walled carbon nanotubes (SWCNTs) on the extent of oxidative damage to DNA in the liver, colon and lungs of rats.  The authors were duplicating a study that demonstrated oxidative DNA damage in these organs following a single oral dose of the constituents in diesel exhaust.  DNA damage is important, because it is often a precursor to cancer.

In the new study, rats fed either fullerenes or SWCNTs exhibited increased levels of DNA damage in the liver and lungs, sometimes even at quite a low dose.  Neither nanoparticle had a significant effect on the colon at the doses tested, however.

So what does all this mean?  In my view, since we already know that chemicals in personal care products can wind up in drinking water, we should take more of a look-before-we-leap approach to evaluating new chemicals, including engineered nanomaterials, for use in these products.  This need also extends, of course, to the thousands of ingredients already present in personal care products that have not been adequately tested for safety.

Among the questions we need answered up front:  what is the expected fate and transport of such nanoparticles in the environment, who or what might be exposed during this process (including fish, bacteria, and any other critters), and if they wind up in our drinking water, what are the potential health effects?  The study discussed in this post is but one tiny piece of this puzzle.  (Of course, oral exposure to nanoparticles may occur through routes other than drinking water contamination, such as through direct ingestion of nanomaterials in foods or via the mucociliary escalator that removes inhaled particles from the lung to the digestive tract.)

Clearly we need regulators to be asking more of these questions – and requiring companies to provide the answers.  As my colleague Richard Denison pointed out in two of his recent posts, however, this doesn't appear to be happening with new chemical notifications being processed by EPA.  Instead, EPA seems to regard an absence of data as grounds for concluding an absence of risk.  And EPA gives the potential impacts at the downstream end of nanomaterials' lifecycles very short shrift.

My last post identified two Section 8(e) "substantial risk" notices pertaining to carbon nanotubes, one submitted by BASF, the other by Arkema.  I have in my files one additional Section 8(e) notice for a single-walled carbon nanotube (SWCNT), submitted by DuPont.  With three Section 8(e) notices submitted for different rat pulmonary toxicity studies on carbon nanotubes, it's interesting to compare their results.

The DuPont Section 8(e) notice was submitted to EPA on April 10, 2003 (#8EHQ-0403-15319), but appears not to be available anywhere online.  It reports the results of an intratracheal instillation study in rats, which was among the first to identify the formation of lung granulomas.  The full study was published in Toxicological Sciences in 2004.

Here are a few details of each of the three studies reported in the Section 8(e) notices:

DuPont, both in its Section 8(e) notice and in the published paper, questioned the physiological significance of its finding of granulomas, suggesting it was likely an artifact of the method of administration.  Indeed, the instillation procedure introduced clumps or ropes, rather than individual CNTs, and some heavily dosed animals died of apparent suffocation.  DuPont based its conclusion on a variety of factors, including the apparent lack of dose-response and the absence of other signs of lung toxicity.  It called for inhalation toxicity studies to be conducted to resolve whether the effects it observed were real.

What do inhaled carbon nanotubes do to the lungs?

Via the more recent Section 8(e) notices, we now appear to have confirmation of the ability of CNTs to produce lung granulomas when inhaled.  While there are differences in the nature of the CNT material administered and other study details, both Arkema and BASF report the dose-dependent formation of granulomas. 

They also report numerous other signs of lung toxicity, including increased lung weights.  Arkema established 0.1 milligrams per cubic meter (mg/m³) as its short-term no-effect level, while BASF's longer-term study found granulomas at that same dose – the lowest tested – and hence could not establish a no-effect level.

This level of toxicity exhibited in a 90-day repeated dose inhalation study is very high.  To judge toxicity, EPA uses internationally accepted toxicity criteria set forth under the Globally Harmonized System (GHS) of Classification and Labeling.  The GHS "high-concern" classification criterion for a 90-day inhalation toxicity study involving exposure to a dust is a lowest-observed adverse effect level (LOAEL) of less than 0.02 milligrams per liter (mg/L) per day (see Appendix 2, Table 5 of this EPA document).

In the BASF study, the LOAEL was the lowest dose tested, 0.1 mg/m³ = 0.0001 mg/L, which is 200 times lower than the GHS' "high-concern" cutoff value!!

For further context, consider the LOAEL values found in rat subchronic and chronic inhalation studies of another particulate material, respirable crystalline silica.  LOAELs ranged from 1.0 mg/m³ in a 24-month inhalation study to 2.0 mg/m³ in 6-month studies (see Table 18 of this document).  That means the CNTs studied by BASF are at least an order of magnitude more toxic than silica.

Also note that inhaled silica is much more toxic to humans than to rats:  the human LOAEL for inhaled silica for chronic lung disease is far lower than the rat LOAEL, on the order of 0.02-0.05 mg/m³ (see Table 16 of this document).  So if humans are also more susceptible to the effects of CNTs than are rats, the human toxicity level could be far lower than what BASF has observed in rats.

Remember that other recent studies suggest certain MWCNTs can behave biologically rather like asbestos.  Together with these inhalation studies, it's no wonder EPA is requiring inhalation toxicity studies and workplace inhalation exposure controls for CNT producers, as I discussed in a previous post. 

But these findings beg the question:  Will that be enough?

[UPDATE:  Thanks to Andrew Maynard's blog post today, I have learned of a new published study by Shvedova et al. that found that administration of SWCNTs by inhalation was actually more effective than administration by pharyngeal aspiration in causing inflammatory responses and other signs of lung toxicity in the lungs of mice.]

Section 8(e) of the Toxic Substances Control Act (TSCA) requires any company that manufactures, imports, processes or distributes chemicals in the U.S. to notify EPA within 30 days if it obtains new information that "reasonably supports the conclusion that such substance or mixture presents a substantial risk of injury to health or the environment."  Are there Section 8(e) notices for nanomaterials?

EPA posts "sanitized" versions of Section 8(e) notices on its website, stripped of any information deemed confidential by the submitter – which, as allowed under TSCA, frequently includes the identity of the substance, the submitter or both.

So how many of these "substantial risk" notices have been received for nanomaterials?

EPA provides no ability to search for such documents on its Section 8(e) notice website, so we had to do examine each of EPA's month-by-month listings.  These postings extend from September 2008 back to January 2004 – excluding, for some strange reason, the second half of 2004.  (A note that has been on the site for a long time now optimistically states:  "Attention: The 2000-2004 submissions will be posted on this page in the near future.")

Our search found eight Section 8(e) notices that were identified as pertaining to nanomaterials (though there may be others that claimed the nanomaterial identity confidential).  Here they are:

In my next post, I'll look more closely at the Section 8(e) notices EPA has received for carbon nanotubes.

The manufacture of carbon nanotubes (CNTs) is a very complicated business.  Different production processes leave behind different kinds of metal catalysts, which yield differences in physical and chemical – as well as toxicological – properties of the CNTs. 

Removal of metals can alter the surface and other properties.  These properties affect both the performance and the potential toxicity or other biological activity of the CNTs.

While a given material's intended use often dictates the means of manufacturing and subsequent processing, the challenge is to integrate safety considerations up front rather than as an afterthought to be considered only at the end of the design process.  But to do so, we need to develop a much better understanding of how properties and structural features correlate with safety risks. 

A pair of interesting papers just published in the journal Chemical Research in Toxicology examines the relationship between these physicochemical properties and potential hazard. 

In these papers, a group of European researchers created and characterized five kinds of CNTs, and then explored the question of what features are the most important in determining hazard:  residual metals, the presence of oxygenated functional groups, or surface defects.  As measures of hazard, the authors examined the ability of each CNT to scavenge free radicals (a protective function since free radicals can damage cell membranes), cause lung inflammation in rats, and damage DNA.  The results are intriguing:

• The presence of oxygenated functional groups on the surface of the CNTs appears to be most closely associated with genotoxicity. 
• Lung inflammation increased with the number of surface defects or oxy groups, and also with increasing metal concentration. 
• Somewhat counter-intuitively, CNTs that lacked both residual metals and surface defects were the only ones not effective as scavengers of free radicals.  Introduction of surface defects into these CNTs enhanced their ability to scavenge free radicals, supporting the hypothesis that such scavenging is associated with surface defects. 

It's more complicated than I can describe here and it's likely that multiple features contribute to most or all of the observed effects.  But these papers begin the difficult task of teasing out which are the most important properties in determining hazard potential. 

The findings inch us closer to the day when designers of nanomaterials can incorporate specific properties that meet the dual goals of performance and low toxicity. 

Since my first post concerning EPA's Consent Order, I've been reflecting further on the management conditions it imposes – or, more accurately, on what conditions it doesn't impose.  The Order's only such conditions address potential worker exposure.  What about the rest of the nanomaterial's lifecycle?

The need to consider the full lifecycle and the full range of potential release and exposure pathways is a basic tenet of sound and responsible management of nanotechnology.  That's the backbone of the EDF-DuPont Nano Risk Framework, and it's also a key principle in EPA's own Nanotechnology White Paper and Nanotechnology Research Strategy.

Yet the Consent Order lacks conditions to address any potential releases or exposures beyond requiring gloves and other personal protective equipment for workers handling the nanomaterial.

Remember, the Consent Order notes that no test data were included in the producer's premanufacture notification (PMN).  On what basis, then, has EPA concluded that no other potential risks exist?  What about potential releases:

• from the manufacturing facility to the ambient air or water?
• from disposal or other management of wastes?
• from downstream transport, storage or processing?
• from post-use management (e.g., aging, weathering, repair, recycling) and disposal of products (electronics, polymer composites) containing the nanomaterial?

Nothing in the Consent Order addresses these questions – not, for example, a requirement to test products for potential releases, not even provisions to require reporting of waste management information or measurement or monitoring of releases.

Now, it may be that EPA has somehow managed to fully evaluate these and related questions and has determined that all of these risks are – and will remain, no matter what quantity of the nanomaterial is produced and used in the future – negligible.  If so, it should disclose how and on what basis it did so. 

EPA's failure to make public the decision framework it uses to evaluate new chemical submissions for nanomaterials – something we and other stakeholders have been requesting for some time – is a major impediment to building public trust in its process.

EPA's only bite at the apple

Some might argue that all this is premature and that EPA should wait until manufacture and use of this nanomaterial has ramped up even to consider such questions.  But here's the problem with that:  Under TSCA, the PMN review and conditions imposed through the Consent Order are essentially EPA's only bite at the apple.

Once Swan Chemical commences manufacture of its multiwalled carbon nanotubes (MWCNTs), they will be listed on the TSCA Inventory and will no longer be a "new" chemical.  At that point, anyone may manufacture and use the substance without even having to notify EPA.  And no matter what the quantity of the nanomaterial being produced and used, no further review by EPA would be triggered.

It should be noted that EPA is likely developing a Significant New Use Rule (SNUR) to accompany this Consent Order, and there are hints of that in the Order itself.  But all that a SNUR will do is to extend the same conditions that apply to the submitter of the PMN to other producers; it would require them to notify EPA only if they don't comply with these conditions.

Should a concern later develop about some other type of release or exposure not addressed by the Order and SNUR, EPA's only recourse would be to seek to use its authority under TSCA Section 6 to regulate the MWCNTs as an "existing" chemical – something EPA was unable to do, ironically, even for asbestos.  Yet that would be the only way that EPA could impose further conditions on production, processing, use, distribution or disposal of this nanomaterial.

Absent reform of TSCA to provide EPA with greater authority to regulate "existing" chemicals, EPA's new chemical review is, practically speaking, the only chance to ensure that potential risks across a new substance's full lifecycle are addressed.  An examination of EPA's Consent Order suggests that this opportunity has been lost for this nanomaterial.