13414
Langmuir 2008, 24, 13414-13419
Adsorption of Sodium Dodecyl Sulfate and Sodium Dodecyl Phosphate on Aluminum, Studied by QCM-D, XPS, and AAS Philip M. Karlsson,* Anders E. C. Palmqvist, and Krister Holmberg Department of Chemical and Biological Engineering, Applied Surface Chemistry, Chalmers UniVersity of Technology, SE-412 96 Go¨teborg, Sweden ReceiVed July 27, 2008. ReVised Manuscript ReceiVed August 28, 2008 The adsorption of two anionic surfactants, sodium dodecyl sulfate (SDS) and sodium dodecyl phosphate (SDP), at surfaces of aluminum and aluminum oxide has been studied by means of atomic absorption spectrometry (AAS), X-ray photoelectron spectroscopy (XPS), and quartz crystal microbalance with dissipation monitoring (QCM-D). It was shown that more SDP than SDS binds to the surface and that SDP prevents dissolution of aluminum in water whereas SDS does not. This was not obvious, since the adsorption isotherms of the two surfactants to aluminum pigment powder are quite similar, as shown in an earlier work. The decreased aluminum dissolution with SDP compared to SDS was explained by the formation of a more compact protective layer with less permeability on the aluminum surface with SDP than with SDS. This is explained by differences in complexing ability between the surfactants and the aluminum pigment surface. While SDP is expected to form an inner-sphere complex with aluminum, leading to a lower accessibility of aluminum sites to water, SDS is likely to form a weaker outer-sphere complex.
Introduction Aluminum pigment flakes is a product widely used in applications such as silver inks and protective coatings.1 Waterborne paints and printing inks are developed as a response to the increasing demands of authorities and consumers to replace solvent-borne coatings with more environmentally friendly alternatives.2 Replacing organic solvents with water for formulations containing aluminum pigment flakes is not trivial, however. Aluminum reacts with water and forms aluminum hydroxide and hydrogen gas under the slightly alkaline conditions of most waterborne paints, according to reaction 1.
2Al + 6H2O f 2Al(OH)3(s) + 3H2(g)
(1)
This ability to reduce water and generate hydrogen gas is a desirable property of aluminum flakes in their applications as an additive in explosives and autoclaved aerated concrete.1 In waterborne paint and printing ink formulations, however, reaction 1 will lead to loss of the metallic appearance of the formulation and the evolution of hydrogen gas may cause dangerous pressure buildup in the paint containers. Hence, the surface of the pigment particles has to be protected from water. There are several procedures in use to achieve this, and much of the work done in the area has been summarized in a review.3 In a recent study, different surfactants were evaluated as inhibiting agents with promising results.4 In an aqueous suspension of charged particles, surfactants with a charge opposite that of the surface are presumed to form a double layer on the particles at a concentration just below the critical micelle concentration (CMC) of the system. The middle apolar domain of the double layer should ideally prevent the water from reaching the aluminum surface, thus retarding its oxidation. We have previously shown that there is * To whom correspondence should be addressed. Telephone: +46317725616. E-mail:
[email protected]. (1) Wheeler, I., Metallic Pigments in Polymers; Rapra Technology Limited: Shrewsbury, 1999. (2) Niemann, J. Prog. Org. Coat. 1992, 21, 189–203. (3) Karlsson, P.; Palmqvist, A. E. C.; Holmberg, K. AdV. Colloid Interface Sci. 2006, 128-130, 121–134. (4) Karlsson, P. M.; Baeza, A.; Palmqvist, A. E. C.; Holmberg, K. Corros. Sci., 2008, 50, 2282–2287.
a big difference in inhibiting effect between the structurally similar sodium dodecyl sulfate (SDS) and sodium dodecyl phosphate (SDP). Aluminum reacted within 1 day when SDS was used as the reaction inhibitor, whereas aluminum protected with SDP was stable for several months.4 In an attempt to explain the pronounced difference in inhibiting capacity between SDS and SDP, the adsorption isotherms were determined by means of a colorimetric method. The isotherms were found to be virtually identical for SDS and SDP, however. Thus, the large difference in inhibition capacity between the two anionic surfactants cannot simply be attributed to differences in adsorption behavior. The adsorption of surfactants to solid hydrophilic surfaces in water is influenced by interaction between the headgroup of the surfactant and specific sites at the surface. Here, we study the interaction between the two surfactants and the aluminum surface in more detail in order to better understand the differences in their aluminum inhibiting properties.
Experimental Methods Materials. Aluminum pigment flakes (with particle size distribution values d(10) ) 9.9, d(50) ) 29.0, and d(90) ) 80.1 µm) were kindly provided as a dry powder by Carlfors Bruk AB. Heptane (99%), 1-dodecanol (g98%), SDS (g99%), dichloromethane (>99.9%), and N,N-bis(2-hydroxyethyl)glycine (bicine, g99%) were all from Aldrich, and sodium hydroxide (NaOH, g99%) was from Merck. Poly(phosphoric acid) (PPA, reagent grade) from Akzo Nobel AB and fatty alcohol ethoxylate (Sabopal LM11, C10-16E11, 85% in water) from Univar AB were both gifts. All chemicals were used as received without further purification. Surfactant Synthesis and Characterization. The acid form of SDP, that is, the monododecyl phosphoric acid ester, was synthesized by reaction of PPA (30 g) with 1-dodecanol (40 g) with heptane (100 mL) as solvent in a three-neck flask at 80 °C while stirring. The PPA was kept in an oven at 80 °C and added to the mixture in small portions during a period of 30 min. The mixture was left reacting at 80 °C for 1 h and was then allowed to cool to room temperature. The crude product was precipitated, filtered, and recrystallized twice in heptane. SDP was prepared by neutralization of the acid with 1 M NaOH solution prior to use. Mass spectroscopic analysis of dodecyl phosphate was performed with a VG 7070E magnetic sector instrument (VG Analytical/
10.1021/la802198s CCC: $40.75 2008 American Chemical Society Published on Web 11/04/2008
Adsorption of SDS and SDP on Aluminum
Langmuir, Vol. 24, No. 23, 2008 13415
Micromass, Manchester, U.K.). Fast atom bombardment was achieved with a Xe gun at 8 kV, glycerol as matrix, and an acceleration voltage of 5 kV. An electric signal from a coil placed in the magnetic field was used for mass calibration. The magnet scan rate was typically from 10 to 850 m/z in 4 s. Prior to neutralization, the dodecyl ester of the phosphorous acid was analyzed by 1H NMR and 31P NMR on a 400 MHz JEOL spectrometer using CDCl3 as solvent. The CMC of SDS and SDP was determined at pH 8.4 at both 25 and 40 °C employing tensiometry, utilizing a Sigma 70 tensiometer equipped with a du Nou¨y ring. Quartz Crystal Microbalance with Dissipation Monitoring (QCM-D). A QCM-D instrument from Q-Sense AB was used to measure the amount of surfactant adsorbed at a model surface of a 100 nm thick aluminum oxide layer sputtered onto a gold-coated quartz crystal. The use of this model surface was necessary to utilize the QCM-D technique, which exploits the piezoelectric properties of a gold-coated quartz crystal, 14 mm in diameter. The fundamental frequency of the crystal is around 5 MHz, and the third overtone is around 15 MHz. The oscillating frequency of the crystal decreases upon adsorption of solutes at its surface. The surfactant concentrations used were chosen relative to their CMC. Surfactant solutions containing 10 mM SDS and 30 mM SDP in Millipore water were prepared, and their pH was set to 8.4. The surfactant solutions were degassed for 20 min prior to use. The alumina modified QCM-D crystals were cleaned twice in the following way: UV/ozone treatment for 10 min, sonication for 5 min in dichloromethane, rinsing with Millipore water, drying with pure nitrogen gas, and a final UV/ ozone treatment for 10 min. The crystals were mounted in the instrument cell just before use. The experiments were performed at 25 °C, since higher temperatures negatively affected the performance of the instrument. Degassed water was let into the cell, and the baseline frequency signal was measured. A surfactant solution thermostatted to 25 °C was injected into the cell, and the decrease in frequency and increase in dissipation caused by the adsorption of the surfactant were recorded for the third, fifth, and seventh overtone. When no further decrease in frequency was detected, a rinsing cycle was started by injecting Millipore water (25 °C) into the cell. The Sauerbrey relation5,6 was used to convert a difference in frequency (∆f) to a difference in mass (∆m), according to the following expression:
∆mS ) -
∆fC ) δLFL n
(2)
where C is the mass sensitivity constant 17.7 ng/(cm2 Hz) and n is the overtone number (1, 3, 5, or 7). The viscoelastic properties of adsorbed surface layers can be determined by means of the dissipation factor (D), which is the damping of the crystal and consists of the dissipated energy and the stored energy according to eq 3.
D)
Edissipated 2πEstored
(3)
in the modeling. QCM-D measurements were performed and modeled repeatedly for each surfactant, and a maximum difference of 5% between measured and modeled values was allowed. More information on the modeling principle can be found in an article by Voinova et al.7 X-ray Photoelectron Spectroscopy (XPS). XPS was employed to analyze the surface composition of the aluminum powder after exposure to surfactant solutions. The aluminum pigment powder was washed twice in dichloromethane for 10 min and then suction filtered to remove organic contaminants from the surface before exposure to the surfactant solution. The surfactant concentrations used were chosen relative to their CMC. Solutions containing 10 mM SDS and 30 mM SDP were prepared with Millipore water and pH adjusted to 8.4 with NaOH. An aqueous slurry of aluminum pigment powder was prepared and used as reference. The mixtures were stirred at 40 °C for 3 h and subsequently suction filtered. Based on results from previous work, this duration of mixing was considered sufficient for complete adsorption.4 A portion of the aluminum pigment powder was collected, and the rest of the powder was washed with Millipore water for 10 min at 40 °C. A few drops of Sabopal LM 11 were added as wetting agent during the washing. The washed powders were filtered and collected. A small amount of the powder was placed on a piece of tape with adhesive on both sides for analysis with XPS. The measurements were performed on a Quantum 2000 scanning XPS microprobe from Physical Electronics. An AlK (1486.6 eV) X-ray source was used, and the beam spot size was 100 µm in diameter. The analyzed sample area was 500 × 500 µm2, and the takeoff angle was 45° with respect to the sample surface. The information depth was approximately 4-5 nm, and the observed elements were given as atomic percentage of the surface composition. Atomic Absorption Spectrometry (AAS). AAS was used to detect dissolved aluminum in solution. Samples of the aluminum pigment powder were washed twice in dichloromethane for 10 min and suction filtered to remove organic contaminants from the surface. A 10 mM bicine pH buffer solution was prepared using Millipore water. Three mixtures, containing 3 g of aluminum powder and 100 mL of bicine solution each, were prepared and stirred at 40 °C. After 1, 2, and 3 h stirring, the mixtures were allowed to sediment for a few minutes and about 20 mL of the clear solution was then collected with a filter-equipped syringe (0.2 µm, cellulose acetate). The solution was mixed with a CsCl/HNO3 solution to finally contain 2 wt % CsCl and 0.1 wt % HNO3 and then used as reference. Solutions containing aluminum powder and SDS or SDP were prepared accordingly. For the SDS mixtures, a concentration of 20 mM was used in each case and stirring was varied between 1 and 3 h. The three SDP mixtures contained SDP in concentrations of 5, 20, and 40 mM and were all terminated after 3 h. Standard solutions with 5, 25, 50, 100, and 250 ppm aluminum were prepared from a 1000 ppm aluminum stock solution. The standard solutions also contained 2% CsCl and 0.1% HNO3. The measurements were performed on a Varian spectrAA-30 flame atomic absorption spectrometer (red plume), λ ) 396.1 nm, lamp current ) 10 mA, with N2O as oxidizing agent and C2H2 as fuel.
Results and Discussion
The Sauerbrey relation is only valid for small D values, that is, for rigid films, and not for large D values where the film is more viscoelastic. If the D values are large compared to ∆f (D > 10∆f), the film thickness has to be modeled. Modeling was done using the Voigt model, which approximates the adsorbed layer with a spring and a dashpot connected in parallel. In this model, the thickness of the adsorbed layer, δL, is obtained and can be used to calculate the adsorbed mass, ∆mV. The density of the layer, FL, has to be estimated, and the ∆f and D values need to be measured for at least two overtone numbers. An AccuPyc 1330 pycnometer from Micromeritics using helium gas was employed to determine the density of the pure surfactant powders. The densities were determined to 1.08 and 1.12 g/cm3 for SDS and SDP, respectively, and these values were used
Surfactant Synthesis and Characterization. SDP was synthesized according to a method that mainly produces monoesters of phosphorous acid in order to enable a proper comparison with SDS, which is exclusively single chained. For SDP, absence of di- and triester was confirmed by mass spectrometry and further demonstrated by 31P NMR, which only gave one signal from P corresponding to a phosphorous acid monoester. 1H NMR showed traces of 1-dodecanol. The CMC values of SDS and SDP at pH 8.4 were 8 and 25 mM, respectively, both at 25 and 40 °C. The CMC for SDS corresponded well with values found by others.8 CMC values
(5) Sauerbrey, G. Z. Phys. 1959, 155, 206–222. (6) Hook, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796–5804.
(7) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391–396. (8) Flockhart, B. D. J. Colloid Sci. 1961, 16, 484–492.
13416 Langmuir, Vol. 24, No. 23, 2008
for SDP under the conditions used could not be found in the literature. Normally, the amount of surfactant adsorbed at a hydrophilic surface reaches a plateau at the CMC,9 and this was also observed in an earlier work for both SDS and SDP adsorbing at aluminum flakes. The plateau value for both surfactants, 10 µmol/m2, was reached around the CMC.4 The concentrations of SDS and SDP used in the present study were selected with respect to their CMC values. The pH of the surfactant solutions was set to 8.4, which is a typical value for waterborne paint formulations. Aluminum exposed to air10 or water11 has been found to be covered with aluminum oxide of bayerite type. The point of zero charge of bayerite in water is 9.1,12 and the pigment surface, thus, carries a net positive charge at pH 8.4. Although ζ-potential measurements of pure aluminum powder should be interpreted with care, due to the risk of corrosion reactions, the ζ-potential has been measured by Cai et al.13 and found to be positive in the tested pH interval of 4-8.5. The two anionic surfactants, SDS and SDP, are consequently likely to interact with the aluminum surface by means of attractive electrostatic forces, although these should be weak under the conditions used because of the weak cationic character of the surface. Surfactant at the Surface Studied by QCM-D. We have previously shown that SDP is much more efficient than SDS in preventing aluminum pigments from reacting with water.4 This was not obvious a priori, since the isotherms for adsorption of two anionic surfactants at the aluminum pigment powder were about the same. Thus, the differences in inhibition capacity seem not to be due to differences in driving force for the surfactants to adsorb at the aluminum oxide surface. However, it is conceivable that the difference in inhibition effect of SDS and SDP may be explained by a difference in the mode of binding of the surfactant polar headgroup to the alumina surface, leading to different arrangements at the surface. As a tool to further investigate the adsorption of the surfactants to alumina, QCM-D was used. QCM-D provides a quantitative and direct measure of adsorption at solid surfaces exposed to a solution, and in this work an aluminum oxide surface was used. The QCM-D technique is relatively new and so far not widely used for studying surfactant adsorption.14-17 To the best of our knowledge, the method has not been used before for investigating surfactants at alumina. Representative QCM-D curves for SDS and SDP are displayed in panels a and b in Figure 1, respectively. The data shown constitute the change in frequency (∆f) relative to the baseline frequency at three different overtones, n ) 3, 5, and 7, normalized by the overtone number. It can be seen that adsorption of SDP gave rise to a considerably larger decrease in frequency than adsorption of SDS. Furthermore, the character of the curves differed for the two surfactants. An instant adsorption can be seen in both cases, but whereas the adsorption of SDS rapidly reached a constant value, the SDP curve had not reached its (9) Scamehorn, J. F.; Schechter, R. S.; Wade, W. H. J. Colloid Interface Sci. 1982, 85, 463–478. (10) Phambu, N. Mater. Lett. 2003, 57, 2907–2913. (11) Alwitt, R. S. In Oxides and Oxide Films; Diggle, J. W., Vijh, A. K., Eds.; Marcel Dekker: New York, 1976; Vol. 4, pp 169-254. (12) Hiemstra, T.; Yong, H.; Van Riemsdijk, W. H. Langmuir 1999, 15, 5942– 5955. (13) Cai, K.; Ode, M.; Murakami, H. Colloids Surf., A 2006, 284-285, 458– 463. (14) Caruso, F.; Serizawa, T.; Furlong, D. N.; Okahata, Y. Langmuir 1995, 11, 1546–1552. (15) Caruso, F.; Rinia, H. A.; Furlong, D. N. Langmuir 1996, 12, 2145–2152. (16) Boschkova, K.; Feiler, A.; Kronberg, B.; Stalgren, J. J. R. Langmuir 2002, 18, 7930–7935. (17) Knag, M.; Sjoeblom, J.; Oye, G.; Gulbrandsen, E. Colloids Surf., A 2004, 250, 269–278.
Karlsson et al.
Figure 1. Change in frequency (∆f) of an aluminum oxide coated crystal upon adsorption of (a) SDS and (b) SDP at the crystal at 25 °C, illustrated as the third, fifth, and seventh overtone number, measured with QCMD. The frequencies have been normalized by the overtone number. Surfactant solution is injected at 1 and the adsorption is terminated after a few rinsing cycles, 2, 3, and 4.
plateau after 25 min. The rinsing procedure gave a similar appearance of the curves for the two surfactants, with a complete or almost complete removal of the adsorbed surfactant from the surface within two rinsing cycles. A long-term measurement in which the QCM-D crystal is exposed to the SDP solution for 2 h is shown in Figure 2. This measurement is performed in order to allow for a proper comparison with the XPS results presented below. It can be seen that also in this experiment the surfactant is completely removed from the surface upon rinsing. Graphs showing the change in dissipation (∆D) for SDS and SDP can be seen in panels a and b in Figure 3, respectively. For both surfactants, the dissipation, that is, the damping, rapidly reached a constant value upon injection of the surfactant, suggesting that the adsorbed layer is fully developed within a few seconds. For both surfactants, ∆D decreased to zero after rinsing.
Adsorption of SDS and SDP on Aluminum
Langmuir, Vol. 24, No. 23, 2008 13417
Figure 2. Change in frequency (∆f) of an aluminum-coated quartz crystal upon adsorption of SDP at the crystal at 25 °C, illustrated as the third, fifth, and seventh overtone number, measured with QCM-D. The frequencies have been normalized with the overtone number. Surfactant solution is injected at 1 and the adsorption is terminated after a few rinsing cycles, 2.
The adsorbed amounts (∆mS) expressed as ng/cm2 and µmol/ m2, as obtained by the Sauerbrey relation, are summarized in Table 1 for both surfactants and for the third, fifth, and seventh overtones. The sodium counterions were included in the surfactant molar mass. From Table 1, it can be seen that the adsorbed amount of SDP, as determined by the Sauerbrey relation, is 2-3 times higher than that of SDS. It can also be noticed that the adsorbed mass is higher for the lower overtones than for the higher overtones, which suggests that more of the adsorbed layer is included for the lower overtones than for the higher overtones. The Sauerbrey relation, however, assumes rigid films and hence underestimates the adsorbed mass due to coordinating water molecules that are also included in the QCM-D results. The Sauerbrey relation is only valid for adsorbed layers with low viscoelasticity, and to evaluate its validity here the ∆D values for both surfactants at the three overtones, normalized by the change in frequency, were calculated. The values are presented in the table. The ∆D/ ∆f ratio is a measure of the viscoelasticity of the film and it can be seen that SDP exhibits a somewhat higher ∆D/∆f ratio, showing that this film has a higher viscoelasticity than that formed from SDS. In both cases, the viscoelasticity was high and the data therefore had to be modeled. The modeled results of the adsorbed amounts (∆mV) are also presented in Table 1. Based on the results from the Sauerbrey relation, as well as from modeling of the data, one may conclude that SDP adsorbs considerably more than SDS at a model surface of aluminum oxide. To provide an absolute measure of surfactant adsorption excluding the mass contribution of water, the results have to be combined with other analysis methods, such as surface plasmon resonance14,18 or reflectometry.19 Surfactant at the Surface Studied by XPS. The excellent inhibition capacity of SDP compared to SDS may be related to a difference in bonding strength of the surfactants to the aluminum (18) Hedin, J.; Lofroth, J.-E.; Nyden, M. Langmuir 2007, 23, 6148–6155. (19) Dijt, J. C.; Cohen Stuart, M. A.; Fleer, G. J. AdV. Colloid Interface Sci. 1994, 50, 79–101.
Figure 3. Change in dissipation (∆D) of an aluminum oxide-coated quartz crystal upon adsorption of (a) SDS and (b) SDP at the crystal at 25 °C, illustrated as the third, fifth, and seventh overtone number, measured with QCM-D. The values have been normalized with the overtone number. Surfactant solution is injected at 1. Points 2 and 3 indicate rinsing cycles.
pigment surface. A semiquantitative measure of the amount of SDS and SDP remaining at the surface after filtration and washing of the aluminum pigment powder was obtained by XPS. The XPS results are summarized in Table 2 as atomic percentage of all detected elements. It can be seen that a few percent of phosphorus was detected on the surface of the aluminum pigment flakes exposed to the SDP solution. The amount of carbon was also relatively high, as it should be if an alkylphosphate is present on the surface. The amount of phosphorus did not decrease upon washing, which indicates that the SDP is strongly bound to the surface, in agreement with earlier results that phosphate esters bind strongly to aluminum oxide surfaces.4,20,21 (20) Jeon, J. S.; Sperline, R. P.; Raghavan, S.; Brent Hiskey, J. Colloids Surf., A 1996, 111, 29–38. (21) Maege, I.; Jaehne, E.; Henke, A.; Adler, H. J. P.; Bram, C.; Jung, C.; Stratmann, M. Macromol. Symp. 1998, 126, 7–24.
13418 Langmuir, Vol. 24, No. 23, 2008
Karlsson et al.
Table 1. Adsorption of SDS and SDP to an Aluminum Oxide Surface, Measured by QCM-Da SDS ∆mS ∆mV
SDP
n
ng/cm2
µmol/m2
ng/cm2
µmol/m2
3 5 7
131 84 81 142
5.0 2.9 2.8 5.0
303 252 230 568
10 8.0 7.0 18
∆Dn)3/∆fn)3 [106/Hz] ∆Dn)5/∆fn)5 [106/Hz] ∆Dn)7/∆fn)7 [106/Hz]
0.16 0.18 0.18
0.25 0.24 0.22
a ∆mS is the adsorbed amount of surfactant calculated by the Sauerbrey relation for the third, fifth, and seventh overtone. ∆mV is the adsorbed amount of surfactant calculated with the Voigt model, and ∆D/∆f is the dissipation for the third, fifth and seventh overtone, normalized with the frequency.
Table 2. XPS Results Giving the Atomic Composition of the Aluminum Pigment Surfacea reference SDP SDP, washed SDS SDS, washed a
C
O
Al
Na
P
S
11.1 58.6 58.4 17.8 18.0
64.4 27.2 27.7 62.7 61.0
24.5 7.6 8.5 19.5 21.0
0 3.1 1.7 0 0
0 3.5 3.7 0 0
0 0 0 0 0
P is related to the adsorbed amount of SDP, and S to SDS.
In addition, the filtered powder was hydrophobic when exposed to pure water, also indicating an adsorbed layer of SDP.21 In contrast, no sulfur could be found on the aluminum powder that had been exposed to SDS and filtered. This indicates that already the first filtration step was enough to remove this surfactant from the surface. The amount of carbon present in the samples exposed to SDS before and after washing was practically constant and much lower than that of the SDP-treated sample, also in support of less SDS at the surface. Furthermore, the filtered aluminum powder exposed to SDS was completely wetted by water, supporting that no or only a small amount of hydrophobizing surfactant was present on the surface. As revealed by the XPS results in Table 2, SDP binds stronger than SDS to the surface of the aluminum pigment. This has also been suggested by others who have shown that PO43- ions adsorb selectively from a mixture of SO42- and PO43– at aluminum oxide,22 and the same was observed for a mixture of organic esters of phosphorous acid and sulfuric acid.23 He et al. have stated that PO43- and SO42- ions form different types of complexes at the aqueous surface of γ-Al2O3.24 They claimed that whereas PO43- ions form strong inner-sphere complexes, SO42- ions only associate with aluminum as an outersphere complex. We here postulate that the same type of interactions occur for the monoalkyl esters, that is, SDP and SDS, and that this difference in binding at the surface of aluminum pigment is the cause of the difference in inhibition efficiency. Figure 4 illustrates an inner-sphere complex for SDP and an outer-sphere complex for SDS. An additional indication of this difference in organization at the surface is the different shapes of the QCM-D adsorption curves with a more gradual adsorption of SDP compared to SDS, suggesting a reorganization of SDP following an initial adsorption. The formation of different kinds of complexes for SDP and SDS with the aluminum surface may to some extent be explained by the different pKa values of the (22) Tanada, S.; Kabayama, M.; Kawasaki, N.; Sakiyama, T.; Nakamura, T.; Araki, M.; Tamura, T. J. Colloid Interface Sci. 2003, 257, 135–140. (23) Coletti-Previero, M. A.; Pugniere, M.; Mattras, H.; Nicolas, J. C.; Previero, A. Biosci. Rep. 1986, 6, 477–483. (24) He, L. M.; Zelazny, L. W.; Baligar, V. C.; Ritchey, K. D.; Martens, D. C. Soil Sci. Soc. Am. J. 1997, 61, 784–793.
Figure 4. Illustration of an outer-sphere complex between a sulfate ester and the alumina surface (left) and of an inner-sphere complex between a phosphate ester and the alumina surface (right).
two surfactants, which have been reported to be 3.3 for SDS and 7.2 for the second proton of SDP.25,26 However, as shown in the QCM-D measurements illustrated in Figures 1 and 2, both SDP and SDS were almost completely removed from the aluminum oxide surface within two rinsing cycles, and this is the case also for the 2 h QCM-D experiment with SDP. These results are somewhat contradictory to the XPS results above, where SDP was found to bind more strongly than SDS to the aluminum pigment flakes. It seems that even if an inner-sphere complex is formed between SDP and the aluminum oxide surface, this kind of binding is still not sufficiently strong to withstand the extensive rinsing that takes place in the QCM-D experiment. Another possible reason for the divergent results from the XPS measurements and from the QCM-D runs is that in the former the actual aluminum pigment flakes are used while a model surface of sputtered aluminum oxide is used in the QCM-D measurements. There may be structural differences between a native oxide surface formed on metallic aluminum and a sputtered aluminum oxide surface, and this difference may have an impact on the formation of complexes with SDP and SDS. Dissolution of Aluminum. The use of XPS analysis or QCM-D as a monitoring technique provides measures of the adsorbed amount of SDS and SDP at the aluminum oxide surface. The results obtained may be correlated to inhibition efficiency, as discussed above. Another way to measure the effect of the surfacebound surfactant is to determine the concentration of aluminum dissolved in solution. Proper protection through formation of inner-sphere complexes should result in a low concentration of aluminum in solution. Atomic absorption spectrometry (AAS) is a well-known technique to quantitatively determine the concentration of metal ions in solution.27 It was here used to study the effect of SDS and SDP on the dissolution of aluminum. The solutions were prepared with the buffer bicine, which has a pKa of 8.46,28 making it suitable for our purpose, as the adsorption studies were made at pH 8.4. The measured concentration of aluminum in solution is shown in Figure 5. It can be seen that the concentration of dissolved aluminum in the presence of different SDP concentrations is much lower than those of the reference samples taken after 1, 2, and 3 h. This clearly shows that SDP protects the aluminum from dissolution. SDS adsorbed at the surface of the aluminum pigment, on the other hand, seems to be completely ineffective, since the concentration of dissolved aluminum in the presence of SDS was even somewhat higher than that in the reference samples. Apparently, adsorbed SDP, but not adsorbed SDS, prevents leakage of aluminum from the pigments, again (25) Moosavi, A. A.; Gharanfoli, M.; Nazari, K.; Shamsipur, M.; Chamani, J.; Hemmateenejad, B.; Alavi, M.; Shokrollahi, A.; Habibi-Rezaei, M.; Sorenson, C.; Sheibani, N. Colloids Surf., B 2005, 43, 150–157. (26) Nakayama, K.; Tari, I.; Sakai, M.; Murata, Y.; Sugihara, G. J. Oleo Sci. 2004, 53, 247–265. (27) Welz, B.; Sperling, M. Atomic Absorption Spectrometry, 3rd ed.; Wiley: Weinheim, 1999. (28) Beynon, R. J.; Easterby, J. S. Buffer Solutions; Oxford University Press: Oxford, 1996; p 79.
Adsorption of SDS and SDP on Aluminum
Langmuir, Vol. 24, No. 23, 2008 13419
Conclusion
Figure 5. Concentration of dissolved aluminum in different mixtures measured with AAS. References are samples without surfactant measured after 1, 2, and 3 h exposure. For the SDS samples, exposure times were 1, 2, and 3 h, and for the SDP samples an exposure time of 3 h was used. SDP concentrations were 5, 20, and 40 mM.
indicating a distinct difference in the binding to the alumina surface of the two structurally similar anionic surfactants. It seems likely that the palisade layer of long alkyl chains more effectively prevents contact between water and the aluminum pigment surface when bound via an inner-sphere complex than via an outersphere complex. One may note that addition of SDS gave an increase in aluminum in the solution. Ridley et al. have shown that sulfate ions promote dissolution of gibbsite by formation of soluble aluminum-sulfate complexes.29 SDS may act in a similar way. (29) Ridley, M. K.; Wesolowski, D. J.; Palmer, D. A.; Kettler, R. M. Geochim. Cosmochim. Acta 1999, 63, 459–472.
Three analytical methods have been used to explain the pronounced difference between the structurally similar anionic surfactants SDP and SDS in their ability to protect aluminum pigments from reacting with water. The QCM-D measurements showed that considerably more SDP than SDS adsorbed at the model surface of sputtered aluminum oxide. However, the experiments also showed that both surfactants desorbed during the rinsing procedure. Thus, the QCM-D experiments did not give a conclusive explanation of the distinct differences in inhibition efficiency that we have previously reported.6 The XPS results were very different from those of the QCM-D experiments. SDP, but not SDS, remained at the surface after a simple filtration step. These results were in accordance with the results from the tests of dissolution of aluminum. Very little aluminum was dissolved from pigments that had been treated with SDP, whereas SDS-treated pigments gave even higher amount of dissolved aluminum than untreated pigments. Thus, the results from both the XPS analyses and the dissolution tests can be interpreted as being in line with the previous experience of SDP being a good and SDS being a poor aluminum pigment inhibitor. It is likely that the divergent result obtained with the QCM-D technique is at least partly because the model surface used in these experiments is different in character from the real aluminum pigment surface. We postulate that the good inhibiting effect seen with SDP (and also with other alkyl phosphates, as we have previously demonstrated) is due to the formation of inner-sphere complexes with surface aluminum. Acknowledgment. The Knowledge Foundation and Carlfors Bruk AB are acknowledged for financial support of this work. We are grateful to Professor Willis Forsling for valuable discussions about complex formation at the aluminum surface. Lic. Eng. Johannes Bogren is thanked for help with the AAS measurements, Mrs. Anne Wendel for the XPS measurements, and Mr. Daniel Cederkrantz for the surfactant density measurements. LA802198S
