Application of XPS and Solution Chemistry Analyses to Investigate

Oct 31, 2011 - Surface Analysis Laboratory, Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg,. Virginia 24061-...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/est

Application of XPS and Solution Chemistry Analyses to Investigate Soluble Manganese Removal by MnOx(s)-Coated Media Jose M. Cerrato,†,|| William R. Knocke,*,† Michael F. Hochella, Jr.,‡ Andrea M. Dietrich,† Andrew Jones,† and Thomas F. Cromer§ †

The Charles E. Via Jr. Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0246, United States ‡ Center for NanoBioEarth, Department of Geosciences, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0420, United States § Surface Analysis Laboratory, Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0001, United States

bS Supporting Information ABSTRACT: X-ray photoelectron spectroscopy (XPS) was applied to investigate Mn(II) removal by MnOx(s)-coated media under experimental conditions similar to the engineered environment of drinking water treatment plants in the absence and presence of chlorine. Macroscopic and spectroscopic results suggest that Mn(II) removal at pH 6.3 and pH 7.2 in the absence of chlorine was mainly due to adsorption onto the MnOx(s) surface coating, while removal in the presence of chlorine was due to a combination of initial surface adsorption followed by subsequent surface-catalyzed oxidation. However, Mn(III) was identified by XPS analyses of the Mn 3p photoline for experiments performed in the absence of chlorine at pH 6.3 and pH 7.2, suggesting that surface-catalyzed Mn oxidation also occurred at these conditions. Results obtained at pH 8.2 at 8 and 0.5 mg 3 L1 dissolved oxygen in the absence of chlorine suggest that Mn(II) removal was mainly due to initial adsorption followed by surface-catalyzed oxidation. XPS analyses suggest that Mn(IV) was the predominant species in experiments operated in the presence of chlorine. This study confirms that the use of chlorine combined with the catalytic action of MnOx(s) oxides is effective for Mn(II) removal from drinking water filtration systems.

’ INTRODUCTION Soluble manganese (Mn) removal through the catalytic action of MnOx(s)-coated media surfaces combined with the application of free chlorine (e.g., HOCl, OCl) is a very effective water treatment technique.14 Several decades of research have revealed critical aspects of this technique: (a) efficient soluble Mn removal occurred after Mn-oxide coatings naturally developed in filter media through a process referred to as “filter aging”4 resulting in an apparent “natural greensand effect” with the presence of various oxidants;1,5 (b) effective soluble Mn(II) removal was accomplished on MnOx(s)-coated media in the absence of an oxidant because sorption capacity increases with increasing pH or surface MnOx(s) concentration or both;1,2 (c) the application of free chlorine together with MnOx(s)-coated media resulted in effective Mn removal at pH 6 or above, resulting in continuous “regeneration” of media surface Mn adsorption sites over time;1,2 (d) filter backwashing may have altered the physical and chemical properties of MnOx(s)-coated media.3,6 Although the research cited above has provided invaluable information about key factors influencing soluble Mn removal by MnOx(s)-coated media, the specific mechanisms that affect this r 2011 American Chemical Society

treatment technique are still unknown. Most studies to date have obtained kinetic information from solution chemistry experimental data; there is limited microscopic and spectroscopic evidence that can provide a more specific perspective about possible reaction mechanisms applicable to the context of drinking water treatment. Extensive biogeochemistry research performed in natural environmental systems has focused on biotic and abiotic heterogeneous oxidation reactions on MnOx(s) surfaces, reporting the presence of Mn(III), but its role is still unclear.713 More recently, researchers have attempted to link macroscopic and microscopic analyses to better understand the role of MnOx(s)-coated media in drinking water treatment.6,1417 Few investigators have characterized the mineral phase and oxidation state of MnOx(s)-coated media surfaces. X-ray diffraction (XRD), a surface analysis technique appropriate for crystalline phases, has also been used to study the mineral phases in Received: September 16, 2011 Accepted: October 31, 2011 Revised: October 27, 2011 Published: October 31, 2011 10068

dx.doi.org/10.1021/es203262n | Environ. Sci. Technol. 2011, 45, 10068–10074

Environmental Science & Technology MnOx(s)-coated media.6,16 However, the applicability of the results obtained by XRD is limited due to the amorphous and heterogeneous nature of MnOx(s)-coated media typically used in drinking water treatment.6,15,16 X-ray photoelectron spectroscopy (XPS) is a well established surface analysis technique that has been used to determine Mn oxidation states in minerals.6,7,1820 Mn(IV) and Mn(III, IV) oxidation states have been identified in MnOx(s)-coated media using XPS through the determination of the Mn 3s multiplet splitting and the shape and position of the Mn 3p photopeak.6,20 Other investigators used the XPS Mn 2p3/2 peak and detected a mixed Mn(III) and Mn(IV) oxidation state in the surface of Mnoxide coated sand21 and in the biological products resulting from biofiltration.22 However, literature shows that determination of Mn oxidation states using the XPS Mn 2p3/2 peak is not a practical approach as the resulting photopeak shifts from different oxidation states are so small that it is difficult to detect their separation.7,23 The objective of this investigation was to better understand the mechanisms by which soluble Mn-removal by MnOx(s)coated anthracite media is accomplished in the absence and presence of free chlorine. The novel aspect of this study was to couple carefully controlled mass balance measurements (obtained during Mn adsorption and subsequent elution studies) with independent measurements of Mn oxidation states on the media coating using XPS Mn 3p analyses. Integration of XPS and solution chemistry analyses allow for more definitive assessments of the relative role of surface adsorption versus surface-catalyzed Mn oxidation in the overall removal of soluble Mn observed in drinking water treatment plants that employ MnOx(s)-coated filter media.

’ EXPERIMENTAL SECTION Sample Site. MnOx(s)-coated anthracite media were collected in 500 mL autoclaved polyethylene bottles from the filter of a drinking water treatment facility in North Carolina. This conventional water treatment facility utilizes a ferric chloride coagulant and its filtration system consists of dual-media filter beds of anthracite and silica sand. Chlorine is applied just prior to the filtration step for water disinfection and MnOx(s)-coated media regeneration; the chlorine residual in the filter effluent typically ranged from 3 to 6 mg 3 L1. Laboratory Column Experiments. Column tests were performed to evaluate soluble Mn(II) uptake under experimental conditions representative of drinking water treatment plants in the absence and presence of chlorine. A previous investigation showed that Mn-oxidizing and -reducing microorganisms are present in filtration media and could potentially affect the fate of Mn in drinking water treatment plants.24 Particles of MnOx(s)coated anthracite media obtained from the North Carolina facility were soaked in a 100 mg 3 L1 chlorine solution for 48 h prior to being placed in the experimental setup. This was done in order to achieve MnOx(s) coating regeneration on the media and inactivate microorganisms present on the media to reduce the possible effect of microbiologically influenced Mnredox processes during the course of the experiments. Glass columns of 0.95 cm internal diameter were packed with regenerated MnOx(s)-coated anthracite media. Filter column media depths were only 7.62 cm due to the high Mn(II) removal capacity of the MnOx(s) coatings obtained from North Carolina.2,15 Using a manganous chloride stock, an influent soluble Mn(II) concentration

ARTICLE

of 0.3 mg 3 L1 was dosed to columns operated at pH 6.3 and pH 7.2 in the absence and presence of free chlorine under oxic conditions (8 mg 3 L1 dissolved oxygen (DO)). The pH range of 6.37.2 was chosen for this study because it is characteristic of surface water coagulation treatment plants. Another set of experiments was performed using influent Mn(II) concentrations of 0.3 mg 3 L1 (also added as manganous chloride) to feed columns operated at pH 8.2 in the absence of chlorine with DO concentrations of 8 mg 3 L1 and 0.5 mg 3 L1. Nitrogen gas was bubbled through feed solution to maintain 0.5 mg 3 L1 DO levels through the system for this specific condition. A “control” column was operated with no influent Mn and no chlorine addition at pH 6.3 to assess soluble Mn(II) release from the MnOx(s)-coated anthracite media columns due to chemically or microbiologically catalyzed reduction. Solutions consisting of 1 M NaOH and 1 M HCl were used to adjust the feedwater in the column experiments to achieve the desired pH values. The feedwater used for these experiments consisted of deionized water with 0.5 meq 3 L1 as CaCO3 hardness and 1 meq 3 L1 as CaCO3 alkalinity as a source for background ions. Stocks of CaCl2 and NaHCO3 were used as the source of calcium and bicarbonate. Chlorine concentrations for the experiments were targeted at 3 mg 3 L1 dosed as sodium hypochlorite at an application point located close enough to the columns so that the target reaction time with Mn(II) before reaching the columns was approximately 5 s. This reaction time would have been insufficient to promote homogeneous oxidation of Mn(II) with chlorine prior to reaching the MnOx(s)-coated anthracite media.5 Hydraulic loading rates targeted at 2500 L 3 day1 3 m2 were applied to the columns using peristaltic pumps to control the flow. Effluent Mn was monitored for 48 h for the columns operated with chlorine and 72 h for the columns operated with no chlorine. These monitoring times were selected based on the results obtained from previous studies which indicated that (a) 48 h is more than adequate time to evaluate soluble Mn uptake on an MnOx(s) surface in the presence of chlorine due to the rapid regeneration of the MnOx(s) media surface Mn adsorption sites, and (b) 72 h is a reasonable time for the MnOx(s) media to reach sufficient exhaustion due to adsorptive capacity saturation when free chlorine is not present.2,25 After this column-loading period, all columns were eluted with a mild acid solution (pH 5.0) consisting of deionized water (H2SO4 or NaOH were added for pH adjustment) to assess the fraction of Mn(II) that was adsorbed (unoxidized) during column operation. As stated in a previous study,2 it was assumed that operation at pH 5 would promote the release of adsorbed unoxidized Mn(II) due to competition with H+ for adsorption sites26 without promoting reduction of MnOx(s) to Mn(II). Effluent Mn concentrations were measured using a Perkin-Elmer 5100 flame atomic absorption spectrophotometer (Waltham, MA, USA) and a Hach DR/2400 spectrophotometer (preprogrammed method for low range concentrations: 0.0060.700 mg 3 L1 Mn). Chlorine was measured using a Hach pocket colorimeter II test kit (DPD  total chlorine method). A check was performed using the leucoberbelin method27,28 to assess whether any oxidized Mn was present in the elution from the columns using a procedure described by Cerrato et al.24 The leucoberbelin method can distinguish between Mn(II) and oxidation states Mn(III) or higher.28,29 Elemental Extractions. A previously developed method2 was used to determine the extractable elemental content on the filtration media samples as described in another article.20 Solution samples were filtered through a 0.45-μm membrane and 10069

dx.doi.org/10.1021/es203262n |Environ. Sci. Technol. 2011, 45, 10068–10074

Environmental Science & Technology

ARTICLE

Table 1. Peak Position and Fit Parameters for the Gaussian Lorentzian Asymmetric Algorithm Curve-Fit of the XPS Mn 3p Photo-Peaks Obtained for the Mn(II), Mn(III), and Mn(IV) Reference Standards standard

peak position (eV)

fwhma

% Gaussb

T.L.c

T.S.c

χ2d

Mn (II)

47.99

1.91

88

73.3

0.49

1.07

Mn(III)

48.43

2.07

96

64.1

0.64

0.44

Mn(IV)

49.07

2.35

85

76.1

0.36

0.98

a

fwhm = Full width at half-maximum. b % Gauss = Percent of the pure Gaussian line shape. c T.L. and T.S. = Parameters describing the tail of the asymmetric GaussianLorentzian XPS peak. d χ2 = Chi-squared goodness of fit parameter.

metal concentrations were analyzed using an inductively coupled plasma mass spectrophotometer ICP-MS according to Standard Method 3125-B.30 XPS Analysis. The oxidation state of Mn was identified using XPS by analysis of the position and shape of the Mn 3p photoline as described by Cerrato et al.20 XPS analyses were also performed in the Mn 3s core-level region. However, there was no clear evidence for the detection of Mn(III) by XPS through the determination of the Mn 3s multiplet splitting under the experimental conditions used for his study.31 Thus, XPS Mn 3p analyses were pursued instead as it was expected that the chances of detecting Mn(III) in the presence of Mn(IV) and Mn(II) could be greater in the Mn3p region (located ∼4554 eV) because it is closer to the valence band than the Mn 3s region (located ∼8094 eV) and, thus, more sensitive to photopeak shifts. A Perkin-Elmer 5400 X-ray photoelectron spectrometer (Physical Electronic Industries, Inc.) consisting of an aluminum X-ray source and two-channel collector was used. The X-ray anode was operated at 12 keV and 250 W and the vacuum was maintained at or below 5  107 Torr. The pass energy for survey wide scans was 89.45 eV. The pass energy for narrow scans was 17.9 eV. Data were analyzed using AugerScan 3.12 (RBD Enterprises Inc.). The binding energy of the samples analyzed in the Mn 3p region was charge referenced to the gold Au 4f7/2 peak position at 84.0 eV; gold was vacuum deposited on all samples. Narrow scans of the Mn 3p photoline were first analyzed using Shirley background subtraction;18,32 the data were normalized for comparison with Mn(II), Mn(III), and Mn(IV) reference standards as shown by Cerrato et al.20 An asymmetrical GaussianLorentzian algorithm was then employed for curvefitting of MnOx(s) media column samples based on the careful determination of the shape and position of the Mn(II), Mn(III), and Mn(IV) reference standards. The peak position and fit parameters obtained for the GaussianLorentzian asymmetric algorithm curve-fit of the Mn 3p photopeaks obtained for the Mn(II), Mn(III), and Mn(IV) reference standards are presented in Table 1. A similar approach for the analysis of narrow scans was used in a previous study that treated the Fe 3p spectra as a single peak.33 Theoretical and experimental justification for the single broad asymmetric line of the Mn 3p photoelectron spectra is provided elsewhere.34

’ RESULTS AND DISCUSSION Mn Content of MnOx(s)-Coated Anthracite Media. The total Mn content of the MnOx(s)-coated anthracite media collected for this study was 30.7 mg extractable Mn (g dry weight

Figure 1. (a) Mn(II) uptake of columns used to conduct the experiments performed with an influent Mn(II) concentration of 0.3 mg 3 L1 at pH 6.3 in the absence and presence of chlorine (target = 3 mg 3 L1 HOCl). (b) Release of Mn(II) from columns originally operated at pH 6.3 with an influent Mn(II) concentration of 0.3 mg 3 L1 in the absence and presence of chlorine (target = 3 mg 3 L1 HOCl) after elution with a mildly acid solution (pH 5.0).

of sample)1, showing that there is considerable total Mn content in the surface of this sample. This value is comparable to those reported in other studies.15,20 Control Column Experiments. Observed levels of effluent Mn varied from very mild (