Methanolic Corrosion of Aluminum

Inquiry

I understand that there was some issue with CableCURE^®/XL fluid when deployed in Germany in aluminum stranded cables. How do Novinium’s Ultrinium 732 or 733 fluids interact with aluminum stranded cables?

Response

It is true that about 1% of cables treated in Germany in 2002 with CableCURE/XL fluid failed because of the methanolic corrosion of aluminum. Even though the incidence of failures was quite low, the failure mode was dramatic as illustrated nearby. See the image labeled “German Cable Failure” taken from Bertini, “Failures in Silicone-treated German Cables Due to an unusual Aluminum-Methanol Reaction,” (ICC October 29, 2002). In the illustration the grey material between the strands is aluminum methoxylate, the profuse formation of which caused the insulation to bulge. The bulge is described as similar to when a snake eats a rat. An analgous phenomenon occurs in low voltage cables in the presence of water. In both cases, the bulging occurs because the aluminum hydroxide or aluminum methoxylate has a very low density, or put another way, takes up a great deal of volume. Because rejuvenation is utilized to improve the reliability of power distribution cables, even a 1% induced failure rate is unacceptable.

Novinium was founded in 2003, so we enjoyed the benefit of hindsight into the methanolic corrosion of aluminum and hence we addressed this issue in all of our Ultrinium and Perficio technologies. Novinium has not had a single incident of methanolic aluminum corrosion.

To demonstrate why Novinium technology avoids methanolic corrosion, it is useful to understand the mechanism of methanolic corrosion. One compound and one element are required for methanolic corrosion to occur … the compound is methanol; the element is aluminum.

Compound

CableCURE/XL fluid, Perficio 011 fluid, and Ultrinium 732 fluid all include methoxy silanes, which react with encountered water and produce methanol as a by-product. Ultrinium 733 fluid and CableCURE/DMDB do not produce methanol as by-products, instead these fluids produce larger, less chemically reactive higher boiling point alcohols, namely 2-ethylhexanol and n-butanol.

The reaction of methanol with native aluminum (methoxylation) proceeds at a rate proportional to the concentration of the methanol. The concentration of methanol in the strands of a treated cable is influenced by four factors:

1. The amount of water that is present in the strands and the strand-shield. Less water means less methanol; more water yields more methanol.

2. The stoichiometry of the silane water reaction. Stoichiometry is chemist-speak for the ratios at which materials react. For the CableCURE/XL fluid and Perficio 011 fluids, which utilize the same monomeric silane, the maximum possible methanol concentration is about 25% by weight. For Ultrinium 732 fluid the maximum is about 20% by weight. All other things being equal, Ultrinium would enjoy about a 20% lower methoxylation rate because of the superior stoichiometry.

3. The rate at which methanol diffuses from the strand area out of the cable. The diffusion of methanol is quite fast, so the risk of methoxylation decreases rapidly for all technologies. Higher temperature accelerates the diffusion and dissipation of methanol.

4. The use of non-methanol-based alkoxysilanes reduces methanol concentration beyond the 20% stoichiometric advantage described in factor 2 above. In a patented process (U.S. patent 7,611,748 and its foreign equivalents) Novinium adjusts the formulation with more and more non-methanol-based Ultrinium 733 fluid as the anticipated temperature of the treated cable rises.

Element

Except for copper stranded cables that are immune to methanolic corrosion, at first blush it appears obvious that elemental aluminum is available in an aluminum stranded cable, but it is not. As soon as aluminum strands are drawn and laid into a strand bundle on the factory floor, the outside layer of aluminum reacts with oxygen to form aluminum oxide (Al₂O₃). Aluminum oxide forms a dense barrier that protects the underlying native aluminum metal. This aluminum oxide layer is called a patina and it protects the underlying aluminum from further corrosion.

Patina

If you take a piece of aluminum and scrape off the patina with a knife, you will see bright and shiny native aluminum underneath. In the presence of oxygen, the patina begins to reform immediately. The shiny surface will soon return to its dull grey appearance. Of course, in a power cable there are no knives scraping off the protective patina, so how did the CableCURE/XL fluid penetrate the patina? One problem with CableCURE/XL fluid and CableCURE/DMDB is the use of a condensation catalyst called titanium (IV) isopropoxide.  Also known as tetraisopropyltitanate, we will call it TIP. Over the course of Novinium’s research we learned that TIP facilitates the degradation of the patina. Novinium does not use TIP in its Perficio or Ultrinium formulations. Novinium uses a patented catalyst (U.S. Patent 7,700,871 and its foreign equivalents) that does not suffer the same problem.

A second way that the patina can be damaged is bubble nucleation or boiling. Bubbles form in microscopic cracks in the patina and their rapid expansion and sudden disappearance mechanically perturb the patina. In the discussion above we learned that Ultrinium 732 fluids enjoy about 20% less stoichiometric methanol and hence the boiling point of the mixture is higher. Put another way, it takes a greater temperature escalation for Ultrinium to produce bubble nucleation than for CableCURE/XL and Perficio fluids. The patented silanes (U.S. Patents 7,658,808 and 8,101,034 and their foreign equivalents) included in Ultrinium fluids by Novinium and our partners enjoy improved stoichiometry, mitigating methanolic corrosion. CableCURE/XL fluid is particularly egregious in this dimension, because it includes an ingredient called trimethylmethoxysilane (TMMS) that has a boiling point even lower than methanol. To mitigate the aggressive bubble nucleation of 2002 vintage CableCURE/XL fluid UTILX Corporation reduced the concentration of TMMS in CableCURE/XL by a factor of between 3 and 6. This problem with TMMS is well documented by U.S. Patent Application 2009/0114882 and its international equivalent WO 2006/119196. Besides attacking the patina the TMMS creates a fire and explosion hazard. Novinium does not use TMMS in Ultrinium or Perficio fluids.

In addition to mitigating the causes of patina damage, Novinium utilizes a patina stabilizer from BASF^®, called Tinuvin^® 123 hindered amine light stabilizer. In experiments undertaken at Novinium, Tinuvin 123 outperformed all other patina stabilizers by at least a factor of two. Tinuvin 123 has other beneficial performance attributes to extend cable life and is included in Ultrinium and Perficio fluids and its use is protected by U.S. Patents 7,658,808 and 8,101,034 and their foreign equivalents.  In a patented process (U.S. patent 7,611,748 and its foreign equivalents) Novinium increases the supply of Tinuvin 123 by increasing Ultrinium 212 fluid as the anticipated temperature of the treated cable rises.

Summary 

Novinium substantially reduces the methanol concentration using proprietary silanes, does not use low boiling and highly flammable TMMS demonstrated to cause bubble nucleation even at moderate temperatures, eliminates a patina attacking catalyst utilized in the offending formulations, and adds a patina stabilizing compound to all but prevent methanolic corrosion of aluminum in its Ultrinium formulations. Perficio technology includes the improved catalyst and patina stabilization, and does not use low boiling TMMS. Perficio suffers from a higher methanol concentration than Ultrinium technology. Perficio technology should not be utilized in high temperature aluminum-conductor applications.

Dielectric I

Dear Wisest Webbed one,

I am working on my comments/concerns for the Safety Section for the C30 Draft Guide [Editors note: C30 is a group within the IEEE/PES/ICC working towards the writing of a “Draft Guide for Rehabilitation and Rejuvenation of Extruded Dielectric Cable Rated 2.5 kV through 46 kV.” While that Draft is being crafted, interested readers may wish to review a Novinium published early version undertaken at the behest of the group’s chair, which is available at www.novinium.com/Standards.aspx.], and have a few questions. Just because the equipment is all or mostly plastic materials, these questions still need to be explored and discussed.

1. What is the insulation rating of the fluid(s) that you use for injecting?
2. What is the insulation rating of the hose(s) you use from your canisters to the injection point?
3. What is the insulation rating of the canisters themselves?
4. What is the insulation rating for the combination elbow/canister?
5. Do you have the test data for these pieces of equipment? Will you share the test data with the group?
6. Have you looked at the electrical separation distance from your canisters to the cover or live bushing?
7. Have you looked at the electrical separation distance from your canisters to ground wires in the equipment?
8. Have you looked at the electrical separation distances from your canister connected to one phase and to the other phases in a three phase installation?

Your froguidance would be appreciated.

Alabama detailed draft

Dear Alabama-

I don’t know about you, but the fact that the first three letters in dielectric spell “die” makes me want to take extra measure to assure the safety of all who use dielectric enhancement technology. The heart of your eight questions can be summarized thusly: To what extent does the introduction of injection equipment into energized devices impact the safe operation of medium voltage circuits? Before I answer that question generally and some of your more specific questions, it is important to recognize that there are three distinct methods of dielectric enhancement fluid injection and two conductor realms in which those methods are deployed. Ordered from oldest to most advanced, the three methods are:

    UPR ♦ Unsustained Pressure Rejuvenation,

    iUPR ♦ improved Unsustained Pressure Rejuvenation, and

    SPR ♦ Sustained Pressure Rejuvenation.

To learn more about the details of these rejuvenation methods, check out my June 18, 2010 post titled “How to Inject.” Novinium is synonymous with safety, so we use the two most advanced processes almost exclusively. In the very unusual cases where Novinium leaves injection devices attached to energized devices for more than a couple of days, equipment designed specifically for that case is deployed. Because it is a rare case, I refer interested readers to my April 15, 2001 post, “Soaking II: Safety First” for more details.

In this series of posts, I focus on the cases most important to circuit owners. In the table nearby the implementation of the three injection methods are compared for the two conductor cases, namely small conductor and large conductor. Small conductors include stranded conductors, generally with 19 or fewer strands. Note that for large conductors, cables are injected de-energized whichever injection method is utilized, and hence your questions are moot for these large conductors. Further, SPR is applied over 99% of the time to deenergized cables, and hence the questions are again largely moot. The only firm in the world that can deliver iUPR and SPR is Novinium, because we own the intellectual property on the methods and chemistry that make those processes possible. For most injection work we undertake, we operate in the green … my favorite color.

When we do inject into energized devices we mitigate the risks about which you are inquiring by limiting the period that injection equipment is connected to energized components. In the red portion of the table, older technology requires extensive soak periods spanning several months. At first glance you might assume that reducing the period of time by sixty-fold would reduce the exposure to the risk by sixty-fold, but you would be underestimating the impact. In the iUPR process, fluid flows in only one direction. A feed bottle with a positive pressure, typically about 20 psig, is attached to one cable end and a vacuum bottle with a negative pressure of around -10 psig is attached to the other cable end. Fluid flows from high to low pressure. Both the feed tank and the vacuum tank are removed concurrently. Thus on the feed side, fluid with a high dielectric strength flows through tubing made of a high dielectric polymer and there is a negligible probability of substantive current flow. I’ll provide some data in my second post, Dielectric II.  With iUPR the only substantive design issue is on the vacuum or outlet side, there is no way to be certain what will come out of the outlet, so we have to assume that it will be water. In practice, water seldom comes out of 7-strand or 19-strand cables (To learn why, click here.), but from a design perspective this is the worst-case assumption.

UPR has the same worst-case on the outlet end, but it also suffers from a more insidious issue on the inlet end. I examined this issue in some detail in my April 15, 2011 post, “Soaking II: Safety First.” You should read that post, but in summary, UPR suffers a risk that conductive contaminants will render the feed fluid conductive. The worst-case assumption on the inlet side is about the same as the on the vacuum side, but the feed is connected for 60 days or more. This risk is recognized by the practitioners of UPR, because they explain the risk and a potential solution to mitigate the risk in U.S. Patent 7,704,087 of April 27, 2010. Amazingly after exposing the risk those practitioners have not implemented their mitigation strategy. I usually speak only of technical matters, but you don’t have to have a law degree to recognize that if you know about a safety problem, but do nothing to mitigate the issue – your liability undoubtedly increases. Risk managers should take this knowledge into account as they make their choice between the various injection processes. If your firm believes in safety as the most important criteria, don’t utilize UPR.

In my next post (Dielectric II), I will provide some data to show why the feed end of an iUPR injection is not a safety issue. In my third post in the series (Dielectric III), I will discuss the design issues of the vacuum tank designed for iUPR and the features that make iUPR the second safest injection approach. Finally, in Dielectric IV, I will address the equipment separation issues you raise in your questions, 6 through 8. For a thorough description of all of the rejuvenation dimensions of safety check out the 89-page treatise, “A Comparison of Rejuvenation Hazards & Compatibility.”

Practicing safe rejuvenation,

T. B. Frog

DMDB Doubts

In my December 29, 2010 post at …

Crazy-Competitor-Claims

Wonderer in the Wilderness inquired …

Question 5. A new fluid, DMDB, has been introduced.  Will this improve injection performance on my URD cables?

In that first post I provided an abbreviated answer.  We learned from the abbreviated answer that DMDB is not appropriate for URD cables in particular, because of two inherent inefficiencies.  One inefficiency is by design; the other … well it’s not by design.  I illustrate the first problem nearby.  We call this property of the fluid, “stoichiometric efficiency.”  (Pronounced stoyk-E-O-meh-tric)

Stoichiometry defines the quantitative relationships that exist between the reactants and products in chemical reactions.  When any of the six monomers in the figure nearby react with water they form desirable products and some undesirable by-products.  The percentage of desirable products compared to the total is the stoichiometric efficiency.  The stoichiometric efficiency can be calculated knowing only the chemical composition.  In the graph titled “Hydrolyzate Concentration in Condensate,” this math has been performed for all alkoxysilanes of commercial significance as a function of the number of alkoxy carbon atoms, from one to twelve.  The six globally significant alkoxysilanes are each illustrated on the figure and their positions within the hydrolyzate concentration continuum are pinpointed.  The table nearby defines the acronyms and provides commentary on each monomer.

Methanol, a one-carbon alcohol, is an undesirable by-product of the first three fluids, which are generally deployed in small diameter URD cables.  At very high operating temperatures (i.e. conductor temperatures above 55°C), methanol can corrode aluminum.  Fortunately, the methanol generally diffuses out of the system very quickly and small diameter cables do not routinely experience 55°C conductor temperature.  As a consequence the risk of methanolic corrosion is quite low in applications where these fluids are properly deployed.  All Novinium fluid formulations include Tinuvin^® 123, which stabilizes the patina on aluminum strands and further reduces the risk of methanolic corrosion.  Patina (pronounced pa-TEE-na), is the natural corrosion resistant coating that forms on metals such as aluminum.

To address methanolic corrosion in larger conductor cables CableCURE^®/DMBD fluid was introduced.  DMDB’s undesirable by-product is butanol, a four-carbon alcohol.  The good news is that butanol is unlikely to cause aluminum to corrode; the bad news is that it comes at the price of stoichiometric efficiency.  In the figure nearby, I have circled in blue the portions of the silane monomers, which yield undesirable by-products.  For the DMDBS monomer, about two-thirds of the molecule provides no benefit to the cable.  Because of its low stoichiometric efficiency, you won’t find this frog suggesting that it be used for small diameter cables – especially when there are much better solutions in wide commercial application.

How about larger conductor cables?  Does DMDB do the trick with those?

In the 1980’s the guys at Du Pont discovered that alcohols with 6 to 24 carbon atoms are “tree growth inhibitor[s] capable of imparting at least a thousand-fold increase in electrical endurance as measured by an accelerated test procedure.” (See U.S. Patent 4,206,260.)  In the figure nearby my green laser is pointing to a lightly-green-shaded region of the graph that falls within the Du Pont discovery.  Instead of undesirable by-products, these longer-carbon-chain alcohols are superb dielectric enhancement fluids.  In other words, with Novinium’s Ultrinium™ 733 fluids there are no undesirable by-products, and hence the stoichiometric efficiency is 100%.  Like the 4-carbon alcohol by-product of DMDBS, the 8-carbon alcohol will not corrode the aluminum.  High performance – no compromises!

Finally, to get the total efficiency, it is necessary to consider other efficiencies including catalytic efficiency.  In my January 3, 2011 post, Catalytic Considerations – Component I, I shed some frog wisdom on that subject.  There is also a technical paper in the Novinium library titled Considerations for Injecting Cable with High Conductor Temperature, which provides even more detail.  The bottom line is DMDBS is definitely not appropriate for small conductor URD cables.  For large conductor cables the best choice is Ultrinium™ 733 fluid which enjoys 100% stoichiometric efficiency and a much higher catalytic efficiency.

Always state-of-the-art,

Thermonuclear Bull Frog

Catalytic Considerations – Component I

In my December 29, 2010 post at …

Crazy-Competitor-Claims

Wonderer in the Wilderness inquired …

1. How can Novinium get the same cable life extension without a soak period?  It would seem to me that Novinium puts less fluid into the cable than one would get with a soak period.

In the previous post, I provided an abbreviated answer; in this post I will provide a more comprehensive answer.  We learned from the abbreviated answer that when Novinium founders conceived of the first generation of treatment fluid over two decades ago, we failed to check the relative diffusion rates of the phenylmethyldimethoxysilane (PMDMS) monomer and the condensation catalyst we chose to provide long life.  In the figure nearby, I show diffusion data that demonstrate that the monomer diffuses about six times faster than the catalyst.  I also show two-dimensional scale representations of these two molecules.   From what we learned in Size Does Matter, one can see that the larger and less flexible titanium(IV) isopropoxide (TIP) molecule would diffuse slower than the PMDMS and it does.

This diffusion mismatch was a tragic mistake, because the monomer does not condense to a longer molecule in the absence of a catalyst.  As a consequence, a substantial portion of the PMDMS diffuses out of the cable shortly after injection without adding to the long-term reliability of the cable.  In fact, about 39% diffuses out prematurely.  This premature exudation is caused by catalytic inefficiency.  We cleverly fixed that problem with U.S. Patent 7,700,871, and I will explain the elegance of that solution in a subsequent post in this series …

Catalytic Considerations – Component II

It isn’t just how much fluid is delivered to the cable strands.  Even more important than the fluid quantity injected are:

1.    The amount of fluid that persists in the cable insulation over its post-injection life, and

2.    The capabilities of the molecules that are delivered to …

a.    interfere with the growth of water trees,

b.    interrupt the conversion of water trees to electrical trees, and

c.    disrupt the inception of partial discharges.

However, it is true that when using Unsustained Pressure Rejuvenation (UPR), more fluid can be delivered into the cable strands with longer soak periods.  Novinium is not dead-set against soak periods.  We employ soaking for special cases.  For example, we sometimes employ a soak period on submarine cables with constrained geometries.  A soak period, however, to compensate for a catalyst error is unforgiveable.  Soak periods compromise safety and operational efficiency and should only be utilized where technical or economic considerations preclude the use of sustained pressure rejuvenation or for unusual cases such as the aforementioned constrained geometry submarine cable.  The figure nearby provides a summary of laboratory measurements of the amount of PMDMS fluid that is supplied during a typical 60-day soak period for 1/0 AWG and No. 2 AWG cables.  Each experiment was performed in triplicate.  The amount of PMDMS provided during the soak period is about the same as the amount of fluid lost from catalytic inefficiency.  Perficio™ 011 fluid uses a patented catalyst system that enjoys a 2% catalytic inefficiency compared to the 39% inefficiency suffered by the TIP catalyst.  Therefore, Perficio fluid delivers about the same amount of active ingredient in the cable insulation without a soak as the older technology delivers with a 60-day soak period.

An even better option is to use Ultrinium™ 732 fluid, which not only uses the same state-of-the-art catalyst used by Perficio, but also includes five other ingredient types (both water reactive and not water reactive), which all increase cable life well beyond that possible with the venerable PMDMS fluid utilized in the Perficio formulation.  In future posts, I will examine each of these ingredients in more detail to shed light on how they function to extend cable life.  The table nearby will be updated to provides the links to the post for each of the five ingredient types.

In the mean time check out Cable Rejuvenation Mechanisms: An Update from the 2010 CIGRÉ Canada Conference on Power Systems.

Smarter each day,

Thermonuclear

Fan the Strands

Dear Greatest Amphibian,

I have a question regarding a statement I heard one of your colleagues make at an ICC meeting recently.  Would you comment on the criticality that wire brushing of the conductor has when installing a connector?

Kindest regards,

JA at Xcel-lent

Dear JA-

You are indeed Xcel-lent because you are not afraid to ask the tough questions.  And I am Xub-erent, because I love to dispel myths.  When connectors are qualified to ANSI C119.4 do you think the manufacturers use old corroded conductors?  If you answered yes, stop reading here.  If you answered no, read on. More...