Printer Friendly

Mg-rich coatings: a new paradigm for Cr-free corrosion protection of Al aerospace alloys.

Environmentally acceptable alternative coatings to chromate pigments and pretreatments for the corrosion control of Al alloy 2024 T-3, commonly used in aircraft, were designed, formulated, and tested as primer coatings to provide protection using particulate Mg-rich pigmentation. The system was designed by analogy to pigmented Zn-rich primer coatings used for the protection of steel. In the current study, four coating polymer systems were examined as possible candidates as polymer matrices for Mg-rich cathodic protect coatings. Mg-rich primers were formulated with ~50-micron average particle size magnesium powder, near to the critical pigment volume concentration (CPVC) for this system. Top-coated scribed coatings systems have been subjected to Prohesion[TM] exposure in dilute Harrison's solution for up to 5000 hr. These coatings are the first nonchromated coatings to satisfy 3000 hr of such exposure and remain shiny and undamaged in the scribe area, only showing damage at about 4800 hr. The corrosion byproducts generated in the scribe areas during Prohesion exposure were examined by energy dispersive X-ray analysis (EDXA), and the local pH of the coating determined by the nature of the salt formed as a function of exposure conditions and time, did not cause Al corrosion.

Keywords: Corrosion testing, electrochemical, corrosion, corrosion protection, epoxy resins, polyurethanes, accelerated testing, aerospace, aircraft, CPVC prediction, chromate removal and replacements, pigmentation, epoxy, urethane, silanes, aluminum


Following by analogy the formulation of Zn-rich primer coatings for the protection of steel, the formulation of Mg-rich primer coatings for the protection of aluminum alloys has been examined in this laboratory. The concept was considered when it was realized that a stabilized particulate Mg powder was available which allowed active metal sacrificial protection of Al and its alloys. This work has as its motivation the protection of high strength aircraft Al alloys such as 2024 T-3 and 7075 T-6 without the use of chromate-based pretreatments or chromate pigments. This class of alloys, whose high strength is based on phase separated intermetallic compounds within the bulk alloy, has proved resistant to efforts to develop corrosion protective (pretreatment + coating) systems that do not contain any Cr, a metal whose compounds are notorious for their toxicity and carcinogenicity. (1) Further, any aircraft painting or depainting operation that utilizes Cr-based pretreatments generates large amounts of hazardous waste which can be handled properly only at great cost. (2)

The basic principles of Zn-rich primer coatings are as follows (3,4):

(1) Coatings are either organic or inorganic in nature (see reference 3 for further discussion).

(2) They are pigmented with particulate Zn, in either spherical or flake form.

(3) The pigment volume concentration (PVC) of the Zn pigment in the coating should equal or exceed the CPVC in order for the coating to properly provide sacrificial/cathodic protection to the underlying steel substrate. Under these conditions, the Zn particles are all in mutual contact as well as in electrical contact with the steel substrate.

(4) The mode of protection by these coatings is sacrificial as long as the Zn is electrically connected to the steel, as the Zn is more anodic (reactive) than Fe (major constituent of steel) in the electrochemical series. Then, the mixed Zn oxides formed in the sacrificial oxidation fill the damaged areas and also sometimes passivate the steel surface by their basic nature. There is some evidence that the electrical connectivity of the Zn particles carries over from the PVC = CPVC (circa 60-70% by volume) to PVC = Volume Percolation Threshold (5) for Zn (~30% by volume for spherical particles), so some sacrificial protection occurs over this range even while the Zn is being consumed by sacrificial oxidation. The percolation threshold for flake pigments may be different depending on particle alignment. (6)

(5) The organic or inorganic matrix of the coating must be stable under the basic environment created by the zinc oxide, hydroxide, etc., formed from the oxidation of Zn in the presence of electrolyte. It must also adhere well to the steel alloys and be stable in a corroding environment.

(6) These primer coatings should be topcoated to function properly and have a long field lifetime. When used properly, these primers provide almost as much protection to steel as galvanizing. With a topcoat, they provide both barrier and damage (sacrificial/cathodic) protection to steel substrates.


As work in this lab and others (7) has shown that satisfactory corrosion protection has not been available for Al 2024 T-3 without the use of chromates, many alternate options for protecting this alloy have been considered by this lab and others, including plasma polymer layers (8) and conducting polymers. (9,10) Because of the availability of particulate Mg appropriate for use as a pigment in coatings, (11) it was decided to examine the possibility of designing Mg-rich coatings that would protect Al 2024 T-3 in a manner analogous to Zn protecting iron alloys (steel). There were two confounding features of considering Mg for cathodic protection of Al versus Zn for cathodic protection of Fe. The first was that particulate Mg can be a fire hazard, but this concern was alleviated by the manner in which Mg pigment was delivered by Eckart.* Their particulate Mg has a thin oxide layer that stabilizes it against further oxidation. [Mg content at 96% and MgO (magnesium oxide) content at 4%.] The second concern was that the oxidation products of Mg, MgO, and its various hydroxides in hydrated form, would create such basic conditions that Al would undergo basic corrosion ([dagger]) as is indicated in its Pourbaix diagram. (12) Fe is mostly passive under basic conditions, so this is not a valid consideration for Zn over steel. As will be shown, the natural Mg oxidation products do not yield a pH high enough to corrode and dissolve Al, so this concern was also alleviated.

Further Considerations

The primary goal of this Mg-rich technology was the development of a new type of primer for Al alloys and a resultant coating system that is suitable for objects whose structural components are made up of Al 2024 T-3 or other corrosion prone aluminum alloys. The system needed to be easy to apply and repair, be compatible with present aircraft topcoats, (13) and eliminate all use of chromate pretreatments and chromate pigments altogether if used as part of the total coating system by providing cathodic protection to the alloy. In addition, cyanide and other toxic substances are also used in most methods for chromate pretreatments. The secondary goal in this study has been to choose or develop coating matrix polymers appropriate for use with Mg and its oxidation products with an environmentally acceptable solvent formulation, so that VOC regulations are met.


Open Circuit Potential and Electrochemical Impedance Spectroscopy (EIS) Studies

The electrochemical studies of Mg-rich primers (without topcoat) formulated in our laboratory were first carried out on the surface of the primed Al 2024 T-3 alloy immersed in 3% NaCl solution. (14) The corrosion potential, [E.sub.corr], or open circuit potential (OCP), for the coatings in contact with the alloy was monitored and the EIS spectra of three primer sets, as formulated in an epoxy-polyamide polymer matrix, were recorded as a function of time. The OCP is the mixed potential achieved when a corrosion reaction is occurring between the anode and cathode of the reaction system. (12) The data presented below are for three of these primers based on ~50 micron average particle size Mg powder at 43, 46, and 50% PVC. These data indicate that the most effective protection from just the primer is about 46% PVC, which was the estimated CPVC for this system.

Figure 1 gives OCP versus exposure time for Mg-rich primers, formulated at 43, 46, and 50% PVC in a polyamide/epoxy coating polymer exposed to 3% NaCl solution at pH ~6.2. Interpretation of the events is as follows. Initial OCP values for the three sets correspond to a potential for Mg metal, [E.sub.Mg] = -1.50 V to -1.60 [V.sub.SCE], and the primers appear to be acting like bare Mg. (16) Subsequently, over a 24-hr period, Mg and the Al alloy polarize to a mixed potential corresponding to the corrosion potential, [E.sub.corr], at which the Mg is still sacrificially protecting the Al 2024 whose [E.sub.2024] is -0.68 V versus saturated calomel electrode (SCE). The observed mixed potential for Mg and Al alloy in 3% NaCl was found to be about [E.sub.corr] = -0.90 V to -1.00 [V.sub.SCE]. OCP values extending beyond the initial 24-hr period varied according to primer PVC. The initial lower mixed potential value, [E.sub.corr] (Figure 1), for the 43% PVC sample is thought to be due to the lower effective active metal area as a result of higher polymer coverage at the Mg/Al alloy interface. Initially, the Mg-anode dominates the OCP. The gradual rise in OCP for the 43% PVC sample toward [E.sub.2024] = -0.68 m[V.sub.SCE] is assumed to be due to reactive consumption of the exposed Mg in this system and the disbonding of epoxy coating polymer from the cathode surface. The gradual decrease in OCP of the 43% PVC sample toward [E.sub.mix] = -0.90 V to -1.00 [V.sub.SCE] may be due to resistance polarization by the formation and packing of Mg oxides in the coating. The initial and continuous decrease in OCP of the 50% PVC sample is concluded to be due to a higher void volume in the primer as well as a higher cathode area at the primer alloy interface. The OCP of the 46% PVC sample quickly arrives at the [E.sub.mix] = -0.90 V to -1.00 [V.sub.SCE] value and remains constant for the duration of the test time period. Thus, it is surmised that the 46% PVC primer corresponds to the critical pigment volume concentration (CPVC) for the primer, suggesting that cathodic protection of the Al alloy due to Mg metal occurs most effectively at or near CPVC. Figure 2 shows the low frequency impedance modulus |Z| versus exposure time measured in 3% NaCl solution at pH = 6.2 on 43, 46, and 50% PVC Mg-rich primers. This figure demonstrates the effect of PVC at CPVC for Mg-rich primers. The |Z| values for the 46% PVC samples yielded higher values over the 28-day period suggesting proper formulation at or near the critical pigment volume concentration, which is required to ensure close packing of Mg pigment with minimum resistance from the polymer matrix of the system, but with polymer matrix content sufficient enough to ensure good substrate wetting and reasonable physical properties from the primer.



Trends in OCP data suggest three distinct periods that distinguish the evolution and effectiveness of cathodic protection in the Mg-rich primers as a function of exposure time. These are as follows:

PERIOD I -- Initial immersion day one, the "activation" period when the value of the corrosion potential shifts to a cathodic value -1.1V vs SCE, corresponding to the Mg metal/Al-2024 T-3 mixed potential in the electrolyte. Magnesium immediately begins to oxidize when it contacts the sodium-chloride solution and is "activated."

PERIOD II -- Once initially past the "activation" period, the cathodic protection mechanism reaches its peak due to a maximum in the ratio of magnesium-to-aluminum area ratio. This occurs around day 5~7 when the corrosion potential shifts to a more anodic value of about -0.9 V vs SCE; it is where a relative stabilization called the "transition" period occurs.

PERIOD III -- After the transition period, and up to day 21, the corrosion potential shifts out of the cathodic protection domain, and the potential fluctuates as the film's solution chemistry begins to change. At this time oxygen reduction begins to occur on the upper part of the film causing a local increase in pH that changes the corrosion products from magnesium hydoxychlorides to magnesium hydroxides, the same as at the interface.

Initial Accelerated Testing

In metal-rich coatings, the pigment volume concentration is high and close to critical PVC, in the region at which paint properties such as water permeability and cohesive strength change dramatically. Therefore, Prohesion[TM] cyclic exposure in dilute Harrison's solution with no topcoat allows easy access of acidic electrolyte, atmospheric oxygen, C[O.sub.2], and water to the coating's Mg anode. Topcoating the Mg-rich primer insulates or screens it from the primary cyclic effects of Prohesion, which in turn prevents direct observation of the processes that occur when the coating is scratched or chinked when the Al alloy is exposed to an acid rain environment. In order to better observe the occurrence of such processes, primed panels were directly exposed to dilute Harrison's solution, without topcoat, and monitored. The pH of dilute Harrison's solution is about ~4.5 which corresponds to the pH at which Mg metal readily forms salts with C[O.sub.2], S[O.sub.4.sup.2-], and O[H.sup.-]. The formation of these salts was observed to occur on the surface and at the interface of the primer over a given time interval that corresponds to three distinct events:



(1) EDXA spectra have revealed the formation of magnesium carbonate hydrates at the primer liquid/vapor interface, dypingite [[Mg.sub.5](C[O.sub.3])4 (OH)[.sub.2]*8[H.sub.2]O] and hydromagnesite [[Mg.sub.5] (C[O.sub.3])[.sub.4](OH)[.sub.2]*4[H.sub.2]O]. These salts have been observed to be present in the first 500 hr of exposure for all non top coated primed Mg-rich panels tested.

(2) For exposure times beyond 500 hr, brucite [Mg(OH)[.sub.2]] domains begin to form and subsequently extend throughout the bulk of the primer. During this time the aluminum alloy remains cathodically protected as scribed lines remain unblemished.

(3) For exposure times greater than 1300 hr, primer failure and film delamination correspond to the accumulation of hexahydrite [(MgS[O.sub.4])*6[H.sub.2]O] compound at the interface. Failure occurs (Figure 3B) when the Mg-metal and brucite structure have been depleted from the coating polymer matrix and sufficient hexahydrite salts have accumulated at the alloy interface. At this time the coating polymer ruptures and fragments from compressive forces exerted by hexahydrite structures.

The first 24 hr of exposure to salt fog solution with atmospheric C[O.sub.2], magnesium forms magnesium carbonate compound, [Mg.sub.5](OH)[.sub.5]*C[O.sub.3], at the surface, which is replaced by a more densely packed magnesium hydroxide, Mg(OH)[.sub.2], pseudo-hexagonal crystal structure. The rosette structure observed in the magnesium epoxy primer scanning electron microscopy (SEM) images shown in Figure 3A is consistent with the brucite magnesium hydroxide Mg(OH)[.sub.2] (acicular needle) crystal formed in Prohesion exposure. Further observations were made on the Mg-rich primers exposed to Prohesion cyclic salt fog with dilute Harrison's solution:

(1) A white oxide area [magnesium hydroxide (brucite)] formed over magnesium metal and energy dispersed X-ray analysis (EDXA) measurements indicated the presence of magnesium, oxygen, and aluminum with a minimum amount of carbon detected.

(2) In the scribed area with no epoxy matrix or magnesium metal originally present, EDXA spectra show carbon, oxygen, magnesium, and aluminum with the possible presence of dypingite (magnesium carbonate) structure over the exposed aluminum surface.

In summary, Table 1 gives the relative pH, the solubility product, and the water solubility for magnesium salts identified in EDXA spectra. It was observed that the salts generated during the first 1000 hr of exposure increased in local pH according to a stratification scheme from the coating/alloy surface interface toward the external surface of the coating.

The degradation process of Mg-rich coatings exposed to an acidic environment may be described as follows. The more acidic salt (i.e., hexahydrite) was identified at the alloy interface where local pH conditions are lower due to anodic polarization conditions. (17) The carbonate salt was found to develop on the top of the coating along with Mg(OH)[.sub.2], and both are identified as species that form at higher pH. In addition, the damaged/scribed areas did not degrade either the coating polymer or the alloy surface until after the depletion of Mg(OH)[.sub.2] and as the accumulation of hexahydrite salts occurred. According to Kramer, (18) aqueous magnesium hydroxide acts a pH buffer that does not exceed a pH = 10.5, even in the presence of excess Mg(OH)[.sub.2].

Conclusions of Preliminary Feasibility Studies

These very encouraging results were obtained from a simple Mg-rich coating based on an off-the-shelf (OTS) polymer system with no optimization efforts. They showed that the oxidation products of the Mg pigment in an exposure environment, fairly typical of what an actual system might see in field exposure, did not cause basic corrosion of the Al 2024 T-3 alloy. Further, the Mg-rich system did provide cathodic protection to the Al 2024 T-3, giving the system significant corrosion protection properties in a completely Cr-free system with no chromates in pretreatment or chromate pigments in primer. The research studies now proceeded to the improvement of the coating polymer system and additional formulation studies.


Coating Polymer Selection

Traditionally, two-pack zinc epoxy/polyamide polymer materials have been used for the cathodic protection of steel, as they result in crosslinked matrices with a good adhesion and resistance against alkalis, so that any alkaline reaction involving zinc does not affect the binder itself. (19) More recently, epoxy siloxane "hybrid" coatings have been reported (20) to represent a significant advancement compared to epoxy, epoxy acrylic, and polyurethane coatings. Hybrid polymeric matrices, for high performance primers, are designed as polymer composites or alloys that contain a polymer backbone with at least two types of reactive groups that can take part in crosslinking and network formation under at least two different mechanisms.


Silane Modified Multi-Layer/IPN Polymer Matrix

The design of an improved polymeric matrix for Mg-rich coatings involves an easy-to-prepare multilayer scheme that requires minimum preparation of the Al alloy surface and is derived from existing sol-gel technology. The reaction scheme entails initial application of an organo-silane (N-[beta]-(aminoethyl)-[gamma]-aminopropyltrimethoxysilane) with subsequent grafting of organic layers from the surface into the bulk by utilizing a novel silane modified crosslinker. (21) The coating scheme is similar to the "sol-gel" process, but involves a multilayer approach that utilizes an organo-silane substrate treatment from which a moisture-cure polyisocyanate is applied. This has been reported (22) to involve an initial reaction with water to form an unstable carbamic acid intermediate that spontaneously decarboxylates into an amine and carbon dioxide, as shown in Figures 4A-C. To complete the scheme, further bulk crosslinking reactions between epoxy, silanol, and isocyanate are proposed to occur from an aminated surface into the bulk by employing a bulk/surface crosslinker. In brief, the organo-silane modified surface was subsequently sprayed with a 20% solution of polyisocyanate in propylene carbonate, with one of two polyisocyanate prepolymers, (1) 1,6-hexamethylene diisocyanate homopolymer (HMDI) trimer or (2) 4,4'-methylenediphenylisocyanate (MDI) prepolymer, as depicted in Figures 4A-C. Uniform coverage of the wet surface was achieved at ~2 mils/50 microns using a wet film thickness gauge.



The bulk reaction, extending from the surface, occurs between (1) the polyisocyanate, in the primer formulation and the aminated surface (Figure 4C) and (2) the isocyanate and 7-phenyl-1-[4-(trimethyl-silyl)-butyl]-1,2,3,4-tetra-hydro-quinoxalin-6-ol crosslinker (Figure 5) which upon further hydrolysis forms both a polyurea and polysiloxane IPN structure. The silane modified epoxy (HMDI or MDI) hybrids results from a polymeric material consisting of polyurea, polyurethane (from polyisocyanate prepolymer in the presence Ar-OH and Mg/MgO), epoxy-amine, and organo-silane linkages.

Coating Formulations

The materials used in this study are summarized in Table 2.


Critical Pigment Volume Concentration Estimates for Mg-Rich Primers

The CPVC of a coating is a function of the random dense packing efficiency of the pigment plus adsorbed layer thickness ([delta]), which must be experimentally determined. This has been discussed extensively in the literature and a recent review considers new developments. (28) The procedure for obtaining CPVCs for these Mg-rich systems is described as follows. Two magnesium powders, Eckagranules[TM] PK31 with a mean particle size distribution (PSD) of 30 [micro]m, and PK51 with a mean PSD of 70 [micro]m, were used as received, (see Figure 6) and mixed at a 52%-PK31:48%-PK51 volume. A 52:48 volume mix of the two powders yielded a higher bulk density value than that of either powder alone. The CPVC of the primers was first estimated by obtaining a resin/powder rub-up value with Aerosil[R] R202 at 2% volume on total pigment. The final CPVC was calculated from PSDs for all three pigments, assuming spherical geometry, combined with their experimentally determined oil absorption values. Figure 7 shows calculated CPVCs from the ternary diagram for the three-pigment mixture. The volume fraction coordinates (PK31 = 0.51, PK51 = 0.47, and R-202 = 0.02) yields a theoretic CPVC value in the yellow region (@ PVC [less than or equal to] 0.475), which corroborates the experimentally surmised CPVC, ascertained from the EIS data shown in Figures 1 and 2. All CPVC calculations were performed using software described in reference 28.

Characterization of Mg-Rich Coatings Properties

PANEL AND FILM PREPARATION -- The primers from materials in Table 2 were applied to 6 in. X 3 in. Al 2024 T-3 Q-panels[TM], scrubbed with a Scotch Brite[TM] pad, rinsed, and degreased with EEP, then immersed in a 10% phosphoric acid solution for 60 sec and rinsed with distilled water. Al panels were surface modified according to the methodology described in literature. (29) Mg-rich coatings were applied with a touch-up spray gun, and were cured at 35[degrees]C for 14 days. Primed panels were subsequently topcoated with Extended Lifetime[TM] topcoat ELT[TM]*. The average film thickness (FT) ascertained from SEM and EDAX images revealed primer film thicknesses to be estimated at about 50 [+ or -] 20 microns with topcoat film thicknesses estimated at about 100 [+ or -] 40 microns (see SEM and EDAX images illustrated in Figures 8A-D).

MICROGRAPHS SEM AND EDAX -- Coated samples were assembled on aluminum mounts and coated with gold using a Technics Hummer II sputter coater. SEM and EDAX images were obtained using a JEOL JSM-6300 scanning electron micro-scope. X-ray information was obtained by a ThermoNoran EDX detector using a Vantage digital acquisition engine. Figures 8A-D show EDAX cross sections of the four 50% PVC Mg-rich primers, with pigmentary Mg X-ray fluorescence (XRF) counts in red, and Silicon XRF counts in blue, demonstrating the general alignment of Mg powder at the Al interface and pigment distribution in the polymer matrix, which is thought to be related to its dispersion in the coating's polymer matrix.

Testing of Mg-Rich Coatings

MECHANICAL PROPERTIES OF MG-RICH COATINGS -- Tensile properties were measured according to (ASTM D 2370-82) using an Instron[TM] model 5542 with Merlin (2) software. DMTA measurements were made with a Rheometrics model 3-E dynamic mechanical analyzer.

FLAMMABILITY TESTING OF MG-RICH COATINGS -- Six-inch strips were cut from top coated Mg-rich Al panels and subjected to a modified flammability test, referenced in document IPC-SM840B (International Printed Circuit). Also refer to test method and the U.L.-94V flammability specification using a Bunsen burner, with a tube length of four inches, I.D. 0.37 inches with methane gas at equivalent 1000 BTU/F[t.sup.3]. A propane torch with flame temperature 1120[degrees]C (or 2048[degrees]F) (30) was applied for 30 sec to the backside of the aluminum panel covered with coatings. This test was further modified by scribing an X over the face of each panel to directly expose magnesium metal in the coating to air/oxygen.



EXPOSURE TESTING -- Prohesion exposure was performed according to ASTM D 5894-96. Top coated Mg-rich panels were prepared by covering panel backside and the edges with 3M electroplater's tape and edges were then sealed with a 2K industrial epoxy from Aldrich. Topcoated panels were scribed through the surface of the coating with a carbide tip glass scribe where an X pattern was formed, thus exposing the Al surface.

ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY (EIS) -- The corrosion protection properties of primed panels were evaluated by EIS. The experimental set-up consisted of a three-electrode cell containing 40 ml of 3.0 wt% NaCl aqueous solution, open to air, held at room temperature ~22[degrees]C/72[degrees]F. A saturated calomel electrode (SCE) was used as the reference electrode and a stainless steel plate served as the counter-electrode. All measurements were performed at the open circuit potential of the system. EIS measurements were performed with a Gamry PC-4/300[TM] electrochemical measurement system with potentiostatgalvanostat. Impedance spectra were recorded with a frequency sweep from 0.01 Hz to 10 kHz, the amplitude of the signal perturbation was 10 mV (rms), and Gamry 3.1 Framework[TM] software was used to analyze the data in Bode plot form. These results are presented below.




Viscoelastic Properties of Coating Polymers

Table 3 gives the measured viscoelastic properties for five polymer systems: [T.sub.g], elastic storage modulus E' (minimum), and calculated crosslink density. The significant differences in reported glass transition temperatures are assumed to be related to the individual coatings' chemical properties at full cure. Crosslink density was calculated from (E'):(T = [T.sub.g] + 50[degrees]C) at which the material is in the rubbery state; where [v.sub.e] is the elastically effective crosslink density (31): [v.sub.e] = 3 E'/RT (T + [T.sub.g]).

The results shown in Table 3 also suggest that the difference in the chemical composition of crosslinks formed may lead to observed differences in [T.sub.g]s. According to Hale and Macosko, (32) changes in [T.sub.g] arise both from the disappearance of chain ends and the formation of chemical crosslinks that yield elastically effective chain density at higher levels of branching.

Mechanical Properties of Mg-Rich Coatings

Table 4 gives the measured tensile properties of coating polymer films. Tensile tests were conducted on coating polymer film strips with no visible voids. Mechanical properties, in Table 4 show an improvement in both of the hybrid system's tensile strength and tensile modulus properties over their parent materials. The tensile modulus is known to be a better indication of a film's mechanical properties, as its measurement is less defect dependent than the film's tensile strength. A high tensile modulus also suggests that the material is more elastic which implies a higher degree of cure or conversion.


Both of the hybrid silane modified epoxy-urea/urethane analogs show lower [T.sub.g]s than the parent materials with no significant difference in crosslink density, suggesting the presence of (-N-R-Si-O-Si-R-N-) bonds throughout the IPN matrix.


Flame retardant coatings describe coatings that delay ignition and hinder flame spread. The common test method for evaluating flammability is the Limiting Oxygen Index (LOI) test (ASTM D 2863), where a material is normally considered flammable if the LOI is less than 26. (33) These coatings were all coated with a fluorinated ELT[TM] topcoat that may have contributed in some measure to the coatings nonflammability (Figure 9A-D). The most often reported parameter associated with coating flammability is the material's LOI value. Epoxy/polyamine systems vary from a low of 24 to a high of 32 for silane modified ceramer epoxies (34) while fluorinated polyurethanes are rated up to 50. Another contributing factor to improved nonflammability is the presence of the isocyanurate linkage. HMDI has been reported (35) to possess an inherently higher thermal stability than that of other urethane linkages, such as MDI, as the latter is reported to dissociate at about 200[degrees]C. In general, flammability decreases as the proportion of isocyanurate trizine ring increases. (36-38)

Accelerated Exposure

Prohesion exposure in dilute Harrison solution, (N[H.sub.4])[.sub.2]S[O.sub.4] acid rain conditions, resulted in Mg-rich coatings with conventional binders maintaining clean scribes up to ~1000 hr. Those coatings formulated with hybrid binders also showed clean scribes up to 3000 hr and showed signs of failure only at 5000 hr. The integrity of the primer vehicle appears to be the main issue associated with improved corrosion control in these systems. All samples of Mg-rich primer coatings gave better performance in this exposure than the standard chromate-based system with similar topcoat, and as seen in Figure 10, the best performing of these Mg-rich systems after 4800 hr visibly performed better than a nonpigmented primer/ELT system (Figure 10E) after 1800 hr exposure.



Electrochemical Studies of Mg-Rich Coatings

EIS studies of the Mg-rich primer at 50% PVC (above CPVC) with topcoat, under conditions of high, neutral, and low pH were performed. An EIS test method was used that involved subjecting circularly scribed Mg-rich coatings to immersion varying the pH conditions in 3% NaCl solutions. This method was utilized to help differentiate among the various primer formulations developed in this work. Subjecting the system to acid (pH = 2.8), neutral and basic (pH = 12.0) 3% NaCl immersion under scribed conditions allowed comparison of the formulations in a wide range of exposure conditions. The Mg-rich Al 2024 T-3 panels were topcoated with Deft 99 GY-001 ELT, a very chemically resistant coating, scribed, and then subjected to continuous immersion. Three coating systems were evaluated as Mg-rich coatings: two commercially available products, and one hybrid silane modified epoxy-urea developed in this laboratory, (21) as follows:



(1) Moisture cure (MC-PUR) aromatic polyisocyanate, Desmodur[TM] E23A, polyurea. (39)

(2) Epoxy/polyamine consisting of Epon[TM] 828 with a Mannich base polyalkylamine curative Epicure[TM] 3251.

(3) Hybrid silane modified epoxy-urea consisting of Epon 1001 and Desmodur N3300 aliphatic polyisocyanate, the silane was Silquest A-1120 (N-[beta]-(aminoethyl)-[gamma]-aminopropyl trimethoxysilane) (see Table 2 formulation C).

Experimental Setup

Cylindrical electrode cells were mounted over samples with 1.0 cm diameter circular scribes cut though the coating exposing the Al 2024 T-3 surface. Cylinders were filled with electrolytes of the following compositions: (1) Basic at 3% weight NaCl adjusted to pH = 12.0 with NaOH; (2) Acidic at 3% weight NaCl adjusted to pH = 2.8 with HCl; and (3) Neutral at 3.0% weight NaCl at pH = 6.2. Impedance measurements were carried out over an 11-day time period and the pH was adjusted at each test interval. (See previous section on Electrochemical Impedance Spectroscopy for testing methodology). Unscribed topcoated films were also examined electrochemically to determine what is happening in undamaged coatings due to immersion in neutral 3% NaCl solution.


The visual results from the scribed exposure tests (Figures 11-12) indicated that under conditions of high and low pH the Mg-rich coating formulated with traditional coating polymers, i.e., MC-PUR and the epoxy/Schiff base have weakness at these pH extremes. At high pH = 12, samples (A) and (C) blistered after immersion exposure. At low pH = 2.8, immersion exposure caused the film to disbond and lift from the substrate. The amino-silane modified hybrid polymer matrix provides a much more pH resistant system in an Mg-rich coating in adhesion and reactivity than the more traditional polymers. Zhu and van Ooij (40) have noted that protonated amine (-N[H.sub.3.sup.+]-) from [gamma]-APS ([gamma]-aminopropyltriethoxysilane) in films on Al 2024 T-3 promotes the ingress of corrosive chloride ions into the film. No visible difference in scribed samples after 11 days of testing suggests conditions at pH = 6.2 give higher stability at the primer coating/interface.

Electrochemical Studies

The literature on metal-rich paints as studied by EIS provides good insight into the metal-rich coatings that have not been topcoated, but are exposed directly to the immersion electrolyte. (41,42) Essentially, what these authors have seen is the lifetime of the reactive metal in the coatings and the transition of cathodic protection behavior when sufficient metal is present to protect the substrate, to a more barrier-like system as metal oxides form in the originally porous metal rich primer and reduced the porosity to give a higher resistance film. Our examinations of scribed systems gave similar results, and are shown in Figures 13-15. We just show the low frequency impedance modulus, |Z|[.sub.0], versus time as this tracks the resistance of the coating system versus time quite well. (43) Impedance data from damaged/scribed systems is notoriously difficult to interpret, as the damaged area of the coating dominates the electrochemical behavior of the system when its surface area is large. The impedance measurement in this case is averaging over the entire surface area under study. (44) The low frequency |Z| values seen in Figures 13-16 indicate that conductive pathways are dominating the measurements in scribed areas. Comparison to the |Z|[.sub.0] data from undamaged neutral systems, as shown in Figures 16A-C demonstrates this further. In this case, all of the coatings are in excess of [10.sup.6] [ohm], some as high as [10.sup.10] [ohm], while all of the scribed systems exhibit |Z|[.sub.0] values below [10.sup.5] [ohm], some as low as ~150 [ohm].

Electrochemical Impedance Spectroscopy of Nonscribed Mg-Rich Coatings

Generally, it is reported (45) that after exposure, low frequency impedance values less than [10.sup.7] ohms/[cm.sup.2] result in coatings with a short service lifetime, and coatings with low frequency impedances greater than [10.sup.7] ohms/[cm.sup.2] result in longer lifetimes. Much more data analysis on the electrochemical data for these Mg-rich systems must be obtained, including estimates of diffusion effects from the Warburg tails of the data gathered from the damaged/scribed system, as well as estimates of water up-take from system capacitance calculations. This will be presented in a later article from this laboratory. All of the electrochemical data, especially that of Figure 16, corroborated that the coating system with the Mg-rich primer based on the hybrid polymers system performed the best.


Dynamic and Mechanical Property Results

Viscoelastic DMTA measurements of polymer films revealed that hybrid silane modified epoxy-urea/urethanes displayed lower [T.sub.g]s at equal crosslink densities to those of the parent materials suggesting the formation of bulk (-N-R-Si-O-Si-R-N-) bonds throughout the IPN matrix. In addition, the tensile properties of the silane modified epoxy-urea hybrids were better than their parent materials.

Flammability Results

Using UL94V test protocol, there was no observed difference in flammability with respect to PVC for any of the four systems tested. The two conventional coatings, polyurea (MC-PUR) and epoxy/polyamide disbonded, liquefied, and incinerated with subsequent rapid magnesium incineration. The hybrid-E23A, MDI, did not liquefy or disbond, but formed a limited amount of char, without incineration of the magnesium metal. The hybrid N3300, aliphatic, containing N-alkylisocyanaurate (46) did not char nor did the Mg incinerate. It may be surmised that this Mg-rich coating was more covalently bonded to the Al substrate through the (-Al-O-Si-) linkages that may have decreased the coating's flammability.

Prohesion Exposure Results

Results of ASTM D 5894-96 performed on the four coating systems with varying pigmentary Mg content from 43, 46, and 50% PVC, showed no clear trend in PVC as a function of exposure in dilute Harrison's solution. However, the 50% PVC samples in the hybrid formulations did perform slightly better. The silane modified epoxy-MDI hybrid system performed best reaching about 5000 hr before failure, while the aliphatic silane modified epoxy-HMDI hybrid coatings failed over a range of 3000~3400 hr. The two other conventional Mg-rich coating systems: polyurea (MC-PUR) and epoxy/polyamide, failed over a time period between 2000-2600 hr with no clear difference between the two conventional coating systems.

Electrochemical Impedance Spectroscopy Results

The EIS data for the damaged/scribed systems in Figures 13-15 seem to indicate that the Mg-rich primers are providing active, cathode protection of Al 2024 T-3, and that the mode of protection of these systems is similar to Zn-rich primers for steel. The undamaged coatings (see Figure 16) yielded EIS data in Bode format after three weeks of 3% NaCl immersion for Mg-rich primers plus topcoat that showed no clear trend in PVC of Mg in the systems. The performance of the hybrid polymers system exceeded the other samples among the four coating systems studied and maintained the highest |Z|[.sub.0] throughout the study. Further, the behavior of this coating system was almost purely capacitive with only one time constant throughout this exposure, indicating an excellent coating system. However, the low frequency impedance [Z] modulus in the 0.01-0.1 Hz range for the MC-PUR coatings was in the range of [10.sup.6] ohms/[cm.sup.2], which is two orders of magnitude below the average for the other systems. This behavior is thought to be due to poor wetting of the primer by the ELT topcoat as revealed in SEM profiles, similar to Figure 8, in which numerous air channels and voids were observed to have formed at the primer/topcoat interface. Aside from this, the other coating systems displayed impedances that correspond to long lifetimes.


This study has shown that these pigmentary Mg systems provide sacrificial cathodic protection of Al aerospace alloys in a mode analogous to that of zinc-rich systems. It was also found that coatings formulated with conventional binders and conventional HAP solvents performed poorly in contrast with a hybrid silane modified epoxy-urea IPNs derived from non-HAP solvent systems. Thus, it is crucial that this new paradigm for corrosion protection of Al by Mg-rich pigmentation be obtained in a coating polymer system that provides proper film properties for use as an aerospace primer coating.
Table 1 -- Magnesium Salts Solubility pH

Salt Designation Ksp

Mg(OH)[.sub.2] Brucite 7.1% [10.sup.-12] (a)
MgC[O.sub.3]5*[H.sub.2]O Magnesite 3.8% [10.sup.-6] (a)
MgS[O.sub.4]6*[H.sub.2]O Hexahydrite soluble

Salt [H.sub.2]O g/100 ml pH

Mg(OH)[.sub.2] 7.8% [10.sup.-4] 9.6 ~ 10.4 (b)
MgC[O.sub.3]5*[H.sub.2]O 0.002 (c) 8.0 ~ 8.8 (c)
MgS[O.sub.4]6*[H.sub.2]O 95 (c) 6 ~ 9 (c)

(a) CRC Handbook of Chemistry and Physics, 68th Ed., Boca Raton, FL,
(b) Kramer, D., "Magnesium, Its Alloys and Compounds," U.S. Geological
Survey Open-File Report 01-34.
(c) Seelig, B.D., "Salinity and Sodicity in North Dakota Soils," EB-57,
North Dakota Extension Service, May 2000.

Table 2 -- Magnesium-Rich Primer Materials

Formulation Materials

(A) Hyb-E23A Desmodur[TM] E23-A, Aerosil R202, Eckagranules[TM]
 PK51/31, Epon[TM] 1001 CX, Propylene Carbonate/EEP
(B) MC-PUR (27) Desmodur E23-A, Bentone[R] 34, Eckagranules PK51/31
 Anti Terra[R]U, Aromatic solvent Xylol
(C) Hyb-N3300 Desmodur NC-3300, Aerosil[TM] R202, Eckagranules
 PK51/31, Epon[TM] 1001CX, Propylene Carbonate/EEP
(D) Epoxy-Polyamide Epon 828, Epicure[R] 3115, Aerosil R202,
 Eckagranules PK51/31, AntiTerra[TM]U, Aromatic
 solvent Xylol

Table 3 -- Viscoelastic Properties of Polymer Films

 Crosslink Density
Polymer Film Tg ([degrees]C) (mol/[cm.sup.3]) E' (Pa) Minimum

N3300 (MC-PUR) 122 3.4 [10.sup.-3] 2.6 [10.sup.7]
E23A (MC-PUR) 159 5.8 [10.sup.-4] 6.6 [10.sup.5]
Epoxy-polyamide 65 2.1 [10.sup.-3] 2.0 [10.sup.7]
Hyb-N3300 96 1.3 [10.sup.-3] 1.3 [10.sup.7]
Hyb-E23-A 100 6.9 [10.sup.-4] 6.9 [10.sup.6]

Table 4 -- Mechanical Properties of Polymer Films

 Elongation Tensile Strength Tensile Modulus
Polymer Film at Break (%) (MPa) (MPa)

N3300 (MC-PUR) 8.0 [+ or -] 0.1 45 [+ or -] 7 1250 [+ or -] 90
E23A (MC-PUR) 5.0 [+ or -] 0.3 25 [+ or -] 6 825 [+ or -] 110
Epoxy-polyamide 18.0 [+ or -] 0.1 5 [+ or -] 0.9 150 [+ or -] 50
Hyb-N3300 6.0 [+ or -] 0.3 56 [+ or -] 9 1800 [+ or -] 50
Hyb-E23-A 5.5 [+ or -] 0.2 50 [+ or -] 5 1500 [+ or -] 50


This project was supported by the Air Force Office of Scientific Research, under Grant No. F49620-99-1-0283, Lt. Colonel P. Truelove, Ph.D./Program Manager. Special thanks to Dr. Klaus Greiwe, Eckart GmbH; Heather A. Nash, Research Assistant, Dept. of Polymers and Coatings; Jonathan Wegner, Research Technician, Dept. of Chemistry; Dr. Thomas Freeman, Director, Electron Microscopy Center; Scott A. Payne, Research Assistant, Dept. of Plant Pathology; and all at NDSU that provided advice and/or experimental assistance during the course of this work. Special acknowledgment and thanks to Dr. Joel A. Johnson of the Materials Laboratory of the Air Force Research Laboratory at WPAFB, OH for his help in performing all of the CPVC calculations used in this study.

Presented at the 81st Annual Meeting of the Federation of Societies for Coatings Technology, November 12-14, 2003, in Philadelphia, PA.

* Eckart GmbH, Kaiserstrasse 30, Furth D-90763, Germany.

([dagger]) The primer binder was a 2K epoxy-polyamide: Epon[R] 1001CX and Ancamide[R] 2353 at stoichiometric ratios.

* Supplied by Deft Coatings.

References and Literature Cited

(1) Janata, J., Baer, D., Bierwagen, G.P., Birnbaum, H., Buchheit, R., Davenport, A., Isaacs, H., Hedberg, F., Kendig, M., Mansfeld, F., Miller, B., Wieckowski, A., and Wilkes, J., "Issues Related to Chromium Replacement," presented at the 187th Meeting of The Electrochemical Society, May 21-26, Reno, NV, 1995.

(2) Twite, R.L. and Bierwagen, G.P., "Review of Alternatives to Chromate for Corrosion Protection of Aluminum Aerospace Alloys," Prog. Org. Coat., 33, 91-100 (1998).

(3) Hare, C., "Corrosion Control of Steel by Organic Coatings," in Uhlig's Corrosion Handbook, 2nd ed., Review, R.W. (Ed.), Ch. 55, John Wiley & Sons, New York, pp. 1023-1038, 2000.

(4) Felix, S., Barajas, R., Bastidas, J.M., Morcillo, M., and Feliu, S., "Study of Protections Mechanism of Zinc-Rich Paints by Electrochemical Impedance Spectroscopy," in Electrochemical Impedance Spectroscopy, ASTM STP 1188, Scully, J.R., Silverman, D.C., and Kendig M. (Eds.), Amer. Soc. Testing and Materials (ASTM), Philadelphia, PA, pp. 438-449, 1993.

(5) Bohm, S., Holness, R.J., McMurray, H.N., and Worsley, D.A., "Charge Percolation and Sacrificial Protection in Zinc-Rich Organic Coatings," Eurocorr 2000, Queen Mary and Westfield College, London, September 10-14, 2000.

(6) Kalendova, A. and Kuckacova, A., Macromol. Symp., 187, 377-386 (2002).

(7) Osborne, J.H., Blohowiak, K.Y., Taylor, S.R., Hunter, C., Bierwagen, G.P., Carslon, B., Bernard, D., and Donley, M.S., "Testing and Evalation of Non-Chromated Coating Systems for Aerospace Applications," Prog. Org. Coat., 41, 217-225 (2001).

(8) Yasuda, H.K., Reddy, C.M., Yu, Q.S., Deffeyes, J., Bierwagen, G.P., and He, L., "Effect of Scribing on Corrosion Test Results," Corrosion, 57, 29-34 (2001).

(9) He, J., Johnston Gelling, V., Tallman, D.E., Bierwagen, G.P., and Wallace, G.G., "Conducting Polymers and Corrosion III: A Scanning Vibrating Electrode Study of Poly(3-Octyl Pyrrole) on Steel and Aluminum," J. Electrochem. Soc., 147, 3667-3672 (2000).

(10) Tallman, D.E., Pae, Y., and Bierwagen, G.P., "Conducting Polymers and Corrosion 2: Polyaniline on Aluminum Alloys," Corrosion, 56, 401-410 (2000).

(11) Information from Eckart GmBH, Kaiserstrasse 30, Furth D-90763, Germany, and technical references therein.

(12) Jones, D.A., Principles and Prevention of Corrosion, 2nd ed., Ch. 2, Prentice-Hall, Upper Saddle River, NJ, 1996.

(13) Bierwagen, G.P. and Tallman, D.E., "Choice and Measurement of Crucial Aerospace Coating System Properties," Prog. Org. Coat., 41, 201-217 (2001).

(14) Balbyshev, V.N., Bierwagen, G.P., and Berg, R.L., "Electrochemical Studies of Vinyl Ester Coatings for Fuel Tanks," in American Chemical Society, ACS Symposium Series 689, Organic Coatings for Corrosion Control, Ch. 23, Bierwagen, G.P. (Ed.), p. 292, 1998.

(15) Ellingson, L., "Corrosion Studies of the Protection of Aluminum Alloys and Outdoor Bronze," Masters Thesis, North Dakota State University, June, 2001.

(16) Song, G., Atrens, A., St. John, D., Nairin, J., and Li, Y., "The Electrical Corrosion of Pure Magnesium in 1N NaCl," Corros. Sci., 39, 855-875 (1997).

(17) Bunker, B.C., Nelson, G.C., Zavaldil, K.R., Barbor, J.C., Wall, F.D., Sullivan, J.P., Windlisch, C.F. Jr., Englehart, M.H., and Baer, D.R., J. Phys. Chem. B, Vol. 106, pp. 4705-4713 (2002).

(18) Kramer, D., "Magnesium, Its Alloys and Compounds, U.S. Geological Survey Open-File Report 01-341.

(19) Van Vliet, C.H., Prog. Org. Coat., 34, 220-226 (1998).

(20) Keijman, J.M., "High Solids Coatings: Experience in Europe and USA," Paper 40, Proc. PCE Conference, "Protecting Industrial and Marine Structures with Coatings," The Hague, The Netherlands, March 1997.

(21) Nanna, M.E., Ph.D. Dissertation, "Mg-Rich Coating Systems for AA 2024--T3 with Silane Modified Epoxy-Urea Primers," North Dakota State University, Polymers and Coatings Department.

(22) Ni, H., Nash, H.A., Worden, J.G., and Soucek, M.D., J. Polymer Sci.: Part A: Polymer Chem., 40, 1677-1688 (2002).

(23) Kohli, P. and Blanchard, G.J., Langmuir, 16, 4655-4661 (2000).

(24) Vien, D.L., Colthup, N.B., Fateley, W.G., and Grasselli, J.G., The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, London, 1991.

(25) Ege, S.N., Organic Chemistry, Heath, D.C. and Co., Lexington, 1989.

(26) Aresta, M. and Quaranta, E., Tetrahedron, 48, 1515 (1992).

(27) Adapted from, Bayer Desmodure[TM] E-23A Moisture Cure Urethane, Zinc Rich Primer, Formulation #294-35.

(28) Bierwagen, G.P., Fishman, R.S., Storsved, T., and Johnson, J., "Recent Studies of Particle Packing in Organic Coatings," Prog. Org. Coat., 35, 1-10 (1999).

(29) Walker, P., "Organo Silanes as Adhesion Promoters for Organic Coatings," JOURNAL OF COATINGS TECHNOLOGY, 52, No. 670, 49 (1980).

(30) Bhargava, A. and Griffin, G.J., J. Fire Sciences, 17, 188-208, (May/June 1999).

(31) Hill, L.W., Prog. Org. Coat., 31, 235-243 (1997).

(32) Hale, A., Macosko, C.W., and Bair, H.E., Macromolecules, 24, 2610 (1991).

(33) Tewarson, A. and Pion, R.F., Combust. Flame, 26, 85-103 (1976).

(34) Chiang, C-L., Wang, F-Y., Ma, C-C.M., and Chang, H-R., Polym. Degrad. Stab., 77, 273-278 (2002).

(35) Reymore, H.E., Lockwood, R.J., and Ulrich H., J. Cell. Plast., 14, 332 (1979).

(36) Dick, C., Dominguez-Rosado, E., Eling, B., Liggat, J.J., Lindsay, C.I., and Martin, S.C., Polymer, 42, 913-923 (2001).

(37) Fabris, H.J., "Thermal and Oxidative Stability of Urethanes," Advances in Urethane Science and Technology, Vol. 4, pp. 891-911 (1976).

(38) Singh, L.W. and Mollica, J.C., Rubber Age, 98, 12 (1996).

(39) Schwindt, J., "Moisture-Cured, Polyurethane-Based Surface Tolerant Coatings: An Economical Alternative for Corrosion Control," Leverkusen, Germany, Bayer Co., Formulation #294-35 (2001).

(40) Zhu, V., and van Ooij, W.J., J. Adhesion Sci. Tech., Vol. 19, 1235-1260 (2002).

(41) Vilche, J.R., Bucharsky, E.C., and Guidice, C.A., "Application of EIS and SEM to Evaluate the Influence of Pigment Shape and Content in ZRP Formulations on the Corrosion Prevention of Naval Steel," Corros. Sci., 44, 1287-1309 (2002).

(42) Abreu, C.M., Izquierdo, M., Keddam, M., Novoa, X.R., and Takenouti, H., "Electrochemical Behaviour of An-rich Epoxy Paints in 3% NaCl Solution," Electrochim. Acta, 41, 2405-2415 (1996).

(43) Bierwagen, G.P., Tallman, D., Li, J., and He, L., "EIS Studies of Coated Metals in Accelerated Exposure," Prog. Org. Coat., 46, 148-157 (2003).

(44) Phillipe, L.V.S., Walter, G., and Lyon, S.B., "Invesigating Local Degradation of Organic Coatings," J. Electrochem. Soc., 150 (2003) in press.

(45) Bierwagen, G.P., Jeffcoat, C.S., Li, J., Balbyshev, S., Tallman, D.E., and Mills, D.J., Prog. Org. Coat., 29, 21-20 (1996).

(46) Levchek, L.V., Levcheck, G.F., Balabanovich, I.A., Gambino, G., and Coasta, L., Polym. Degr. and Stab., 54, 217-222 (1996).

Michael E. Nanna and Gordon P. Bierwagen--North Dakota State University*

*Dept. of Polymers & Coatings, Fargo, ND 58105.
COPYRIGHT 2004 Federation of Societies for Coatings Technology
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:First Place 2003 Roon Award Competition Paper
Author:Bierwagen, Gordon P.
Publication:JCT Research
Date:Apr 1, 2004
Previous Article:Telegraphing.
Next Article:Modes and mechanisms for the degradation of fusion-bonded epoxy-coated steel in a marine concrete environment.

Related Articles
Major FSCT awards presented at Annual Meeting.
FSCT Roon Award Competition entries invited; technical awards to be given at ICE 2004.
2003 Joseph J. Mattiello Memorial Lecture.
FSCT to offer a two-part virtual conference on Critical Pigment Volume Concentration.
Teaching and research in Polymers & Coatings at North Dakota State University.
Call for papers issued for Western Coatings Symposium, papers and posters awards competitions announced.
And the winners are ...
Zinc-nickel alloy coatings electrodeposited by pulse current and their corrosion behavior.
ICE 2006--Poster Sessions.
Gordon P. Bierwagen, of NDSU, to deliver 2007 FSCT Mattiello Memorial Lecture.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters