Iron Oxide Surface Chemistry: Effect of Chemical Structure on Binding

N. Cross, and Amy Szuchmacher Blum. Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, Quebec H3A 0B8, Canada. Lang...
6 downloads 14 Views 1MB Size
Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Iron Oxide Surface Chemistry: The Effect of Chemical Structure on Binding in Benzoic Acid and Catechol Derivatives Katalin V Korpany, Dorothy D. Majewski, Cindy T Chiu, Shoronia N Cross, and Amy Szuchmacher Blum Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.6b03491 • Publication Date (Web): 18 Feb 2017 Downloaded from http://pubs.acs.org on February 20, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Iron Oxide Surface Chemistry: The Effect of Chemical Structure on Binding in Benzoic Acid and Catechol Derivatives Katalin V. Korpany, Dorothy D. Majewski, Cindy T. Chiu, Shoronia N. Cross, and Amy Szuchmacher Blum* Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, QC H3A 0B8, Canada

ABSTRACT

Excellent performance of functionalized iron oxide nanoparticles in nanomaterial and biomedical applications often relies on achieving attachment of ligands to the iron oxide surface both in sufficient number and with proper orientation. Towards this end, we determine relationships between ligand chemical structure and surface binding on magnetic iron oxide nanoparticles for a series of related benzoic acid and catechol derivatives. Ligand exchange was used to introduce the model ligands, and the resulting nanoparticles were characterized by FTIR-ATR, transmission electron microscopy (TEM), and nanoparticle solubility behavior. An in-depth analysis of ligand electronic effects and reaction conditions reveals that the nature of ligand binding does not solely depend on the presence of functional groups known to bind to iron oxide

1 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 53

nanoparticles. The structure of the resulting ligand-surface complex was primarily influenced by the relative positioning of hydroxyl and carboxylic acid groups within the ligand as well as whether or not HCl(aq) was added to the ligand exchange reaction. Overall, this study will help guide future ligand design and ligand exchange strategies towards realizing truly custom-built iron oxide nanoparticles.

INTRODUCTION Iron oxide nanoparticles (IONPs) have demonstrated immense promise for medical applications such as magnetic resonance imaging (MRI), drug targeting, labeling, and tumor hyperthermia, and also for magnetic separation, water treatment, as components of

magnetic responsive

hydrogels and load-bearing functional materials, recoverable catalysts, and other materials applications.1‒7 Interest in using magnetic IONPs in inorganic or biological nanomaterials has fueled research into the stabilization of nanoparticles under different solution conditions and modification of the ligand sphere to introduce meaningful functional groups.8, 9 To this end, there is great interest in synthesizing tailored ligands for IONPs to control their solubility, reactivity, functionality, and magnetic properties. Monodisperse iron oxide nanoparticles are easily synthesized in organic solvents using oleic acid as a ligand for steric stabilization.10 Subsequent ligand exchange of oleic acid for target ligands to confer aqueous stability of the resultant nanoparticles,11‒14 or to introduce functional groups amenable to further chemical modification,11, 13, 15‒17 has been very successful. Although there have been many investigations into stabilizing iron oxide nanoparticles with designer ligands, there are very few examples of systematic research regarding the relationship between

2 ACS Paragon Plus Environment

Page 3 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

ligand structure and binding. As a result, there is no clear understanding as to the specific ligand structure required to achieve specific and oriented attachment of a target ligand to the iron oxide surface, especially in cases where ligands possess multiple functional groups. This lack of knowledge has severely hindered the design of new ligands for the successful stabilization and functionalization of IONPs. Our main focus is to better understand the binding of functional groups commonly used to anchor ligands to IONPs. Functional groups typically used as anchors for ligand attachment include carboxylic acid and catechol groups, and to a lesser extent amine, silane, and phosphonate.

4‒6, 13, 18‒22

As such, our study focuses on a family of carboxylic acids, which

include some ligands containing catechol groups, as well as dopamine (which contains both a catechol and amine) (Scheme 1). To date, there has been no thorough study that has examined in detail the binding of a family of similar ligands on iron oxide nanoparticles. By studying a set of chemically related ligands, we are able to investigate the effect of the identity, relative position of substituents, and substituent chain length on binding to IONPs. The binding of dihydroxybenzoic acids has been previously studied for other metal oxides,23‒30 however, prior work on salicylic acid (a similar benzoic acid derivative) has suggested that the chemical composition of the metal oxide determines which ligand surface complex forms.31 Therefore an evaluation of ligand binding to magnetite is warranted, due to its popular use in inorganic and biological nanomaterials, even if previous studies of the target ligands with other metal oxides have been performed.

3 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 53

Scheme 1. Investigated ligands. A family of dihydroxybenzoic acid (DHBA) positional isomers are outlined by a dotted grey line. Ligands that could potentially bind in a salicylate binding mode, using both the carboxylic acid and an adjacent hydroxyl group, are highlighted by diagonal stripes.

Herein we explore the surface modification of magnetite nanoparticles with dopamine, 3,4dihydoxyphenylacetic acid (3,4-DHPA), salicylic acid (SA), and the entire dihydroxybenzoic acid (DHBA) family of positional isomers. The ligands presented serve as good models for binding to iron oxide, as they possess a variety of functional groups (carboxylic acid, catechol, and amine) that are known to bind to iron oxide. We developed a general ligand exchange method to introduce the ligands under investigation, and the resulting ligand binding modes were determined by solubility studies and Fourier-transform infrared attenuated internal reflectance spectroscopy (FTIR-ATR). The target ligands are all small, so interpretation of FTIR spectra 4 ACS Paragon Plus Environment

Page 5 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

should be simplified in comparison to extended molecules. Inspired by previous results concerning the effect of the addition of HCl upon ligand exchange,11 we also evaluated the influence of HCl addition on ligand binding for each of the ligands studied. Results described will aid in the synthesis and evaluation of future catechol, carboxylic acid, and salicylic acidbased ligands for the functionalization of magnetic iron oxide nanoparticles.

EXPERIMENTAL SECTION Materials. For the synthesis of IONP-OA, toluene (anhydrous, ≥99.8%), hexane (CHROMASOLV®, for HPLC, ≥95% ), iron(III) chloride hexahydrate (FeCl3·6H2O) (reagent grade, ≥98%, purified lumps), iron(II) chloride tetrahydrate (FeCl2·4H2O) (puriss. p.a., ≥99.0%) were obtained from Sigma-Aldrich (St. Louis, MO, USA), ethanol (anhydrous) from Commercial Alcohols Inc. (Brampton, ON, CAN), and sodium oleate (C18H33ONa) (>97.0%) from TCI America (Portland, OR, USA). For the preparation and characterization of ligand exchanged nanoparticles, dichloromethane (DCM) (anhydrous, ≥99.8%, contains 50‒150 ppm amylene as stabilizer), oleic acid (90%, technical grade), dopamine hydrochloride (DA·HCl) (98%), salicylic acid (SA) (puriss. p.a., ≥99%), 3,4-dihydroxyphenylacetic acid (3,4-DHPA) (98%), 3,5-dihydroxybenzoic acid (3,5-DHBA) (97%), 2,3-dihydroxybenzoic acid (2,3-DHBA) (99%), 2,4-dihydroxybenzoic acid (2,4-DHBA) (97%), 2,5-dihydroxybenzoic acid (2,5-DHBA) (98%), 2,6-dihydroxybenzoic acid (2,6-DHBA) (98%), sodium phosphate monobasic (BioPerformance Certified, ≥99%), Trizma® base (Primary Standard and Buffer, ≥99%), and 4,5-dihydroxy-1,3-benzenedisulfonic acid disodium salt monohydrate (Tiron) (97%) were purchased from Sigma-Aldrich (St. Louis, MO, USA). 3,4-dihydroxybenzoic acid (3,4-DHBA) (97%) was obtained from Alfa Aesar (Ward Hill, MA, USA). Isopropanol (ACS Plus), sodium

5 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 53

phosphate dibasic heptahydrate (ACS), Tris hydrochloride (Molecular Biology), and sodium acetate trihydrate (99‒101%), were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Hydrochloric acid (HCl), methanol (MeOH), acetone, dimethylformamide (DMF), dimethyl sulfoxide (DMSO), acetic acid, sodium hydroxide (NaOH), hydrogen peroxide (H2O2) (30%), and potassium hydroxide (KOH) were obtained at ACS reagent grade from ACP Chemicals Inc. (Montreal, QC, CAN). All chemicals were used without further purification. When required, DCM was dried with molecular sieves 3Å (Sigma-Aldrich, pellets, 1.6 mm) for 20 minutes. Deionized (DI) water used in this study, to prepare buffers and as a reagent, was obtained from a BarnsteadTM Diamond TII water purification system (>15 MΩ·cm) (Thermo Fisher Scientific, Waltham, MA, USA), Milli-Q® Reference (18.2 MΩ·cm), or Milli-Q® Academic (>18 MΩ·cm) water purification system (EMD Millipore Corporation, Billerica, MA, USA). Sonication steps were performed using a Bransonic® model 2510 ultrasonicator (Branson Ultrasonic Corp., Danbury, CT). Preparation of Oleic Acid (OA) Functionalized Iron Oxide Nanoparticles (IONP-OA). Iron oxide nanoparticles stabilized with oleic acid (IONP-OA) were prepared as outlined in Korpany et al.12 Purified IONP-OA was stored at -20 °C under N2(g) until use. IONP-OA concentration was determined spectrophotometrically by nanoparticle dissolution and iron complexation with Tiron via absorption of iron-Tiron complex at 480 nm (ε(Fe3+, 480nm) = 109.5 (mg/mL)-1 cm-1) as outlined in Ref. 12 and Supporting Information. Synthesis of Ligand Exchanged Nanoparticles. Detailed synthesis and purification of ligand exchanged iron oxide nanoparticles is provided in the Supporting Information. Exchange reactions were based on a method reported in Nagesha et al.32 Ligand exchange reactions were conducted, in general, as follows. The ligand of interest (0.132‒0.272 mmol) (Scheme 1) was

6 ACS Paragon Plus Environment

Page 7 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

mixed with 5 mg of IONP-OA in 15 mL of dry DCM in a Teflon capped glass vial. Either 10 µL of 1 M HCl(aq) or DI water was added to the exchange mixture, then shaken by vortex overnight (~ 18.5 h) at room temperature. After shaking, the nanoparticles were magnetically isolated by a handheld NdFeB magnet. In some cases, the resulting nanoparticles could not be magnetically isolated and were instead recovered by evaporating DCM under N2(g) flow. To purify, recovered nanoparticles

were

repeatedly

washed

with

hexane

and

methanol,

acetone,

or

methanol/isopropanol (1:1) as described in the Supporting Information. Synthesis of Ligand Sodium Salts. Sodium salts of 3,4-DHBA, 2,4-DHBA, 2,3-DHBA, and 3,4-DHPA were prepared by first measuring 0.250 mmol of each ligand into 1.5 mL microcentrifuge tubes, and dissolving in 0.366 ml of MeOH. A solution of 1.865 M NaOH in 80% MeOH was prepared, and 0.134 mL (0.250 mmol) added to each ligand solution to neutralize. After vigorous mixing, all salts began to precipitate out of solution over time. Transmission Electron Microscopy (TEM) Imaging. TEM images were acquired using an FEI TechnaiTM T12 TEM at 120 kV. All TEM samples were prepared using 400-mesh carboncoated copper TEM grids (Canemco Inc. Lakefield, QC, CAN). Samples of IONP-OA were prepared by incubating a TEM grid for 30 seconds at room temperature in a solution of purified IONP-OA in hexane. The grid was subsequently removed from the solution and dried completely in air at room temperature before imaging. Samples of purified ligand exchanged nanoparticles were prepared by depositing 20 µL of nanoparticle solution (as outlined in SI, Table S1) on the carbon-side of a TEM grid and incubating the droplet for 5 minutes at room temperature. In some cases, nanoparticle purification for TEM samples differed slightly from the general purification method detailed in the Supporting Information (SI) as outlined in Table S2 (SI). After 5 minutes the nanoparticle solution was removed from the grid, any excess droplet

7 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 53

removed by wicking with filter paper, and dried completely in air at room temperature before imaging. TEM images were used to obtain nanoparticle (equivalent) diameters using the software Pebbles v.2.0.1,33 and statistics calculated using PebbleJuggler v1.033 or using OriginPro v.8.5.0 and v.8.5.1 (OriginLab Corporation, Northampton, MA, USA). Nanoparticle diameter (d; in nm) was calculated from the nanoparticle equivalent radius (r; in pixels) acquired using the equation d = r·2·SF, where SF is the image scale factor (nm/pixel).12 Nanoparticle Solubility Determination. Solubility of ligand exchanged nanoparticles was determined by the visual examination of redispersed purified ligand exchange products. In some cases, nanoparticle purification for solubility determination differed slightly from the general purification method detailed in the Supporting Information (SI) as outlined in Table S3 (SI). Purified products obtained from each ligand exchange were dried completely under N2(g) flow and were subsequently redispersed in various solvents as outlined in Table 1. Products were briefly vortexed and sonicated (10, 1 second pulses each) before solubility determinations were made. Fourier-Transform Infrared Attenuated Internal Reflectance Spectroscopy (FTIR-ATR). FTIR-ATR spectra were recorded (4000‒400 cm-1, 4 cm-1 resolution, 64 scans per spectrum) using a Spectrum TwoTM FTIR spectrometer equipped with a diamond ATR accessory and processed using SpectrumTM FTIR software (PerkinElmer Inc., Waltham, MA, USA). Ligand reference spectra, and the spectrum of sodium oleate, were obtained from unmodified ligand solids or liquids as applicable. As-prepared ligand sodium salt mixtures were drop-cast onto the ATR crystal, and the solvent allowed to evaporate completely in air at room temperature, leaving a thin film of salt for spectrum acquisition. Nanoparticle spectra were

8 ACS Paragon Plus Environment

Page 9 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

obtained by drop-casting films of redispersed nanoparticles onto the ATR surface, drying the film in air at room temperature, followed by spectra acquisition. IONP-OA was dropcast from CHCl3 or DCM, and ligand exchanged nanoparticles from methanol or ethanol resuspensions. Some nanoparticle preparations were subject to additional washing prior to spectra acquisition, as outlined in the Supporting Information.

RESULTS AND DISCUSSION Ligand Exchange on Iron Oxide Nanoparticles. IONPs were modified with the ligands under study (Scheme 1) using the method outlined in Scheme 2. It was previously shown that ligand exchange procedures involving the catechols dopamine and Tiron11 required the addition of a small amount of HCl to be added to the reaction mixture when amylene stabilized solvents were used, in order for the ligand exchange to proceed to the extent that phase transfer of the nanoparticles could be achieved. Therefore, 10 µL of 1 M of HCl(aq) was added to all ligand exchange reaction mixtures and, to test the effect of HCl addition on the ligand exchange, another set of ligand exchange reactions were examined where 10 µL of DI water was added instead. Twice the molar amount of ligand was required to obtain aqueous stable nanoparticles in the case of 3,4-DHPA versus DA. Due to the successful ligand exchange of 3,4-DHPA with 2 DA molar equivalents all of the following DHBA and SA ligand exchanges were performed with ~0.264 mmol of target ligand.

9 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 53

Scheme 2. General method for the ligand exchange of oleic acid (OA) for target ligand. Overnight shaking at room temperature was used to facilitate the ligand exchange. Ligand exchange reaction products that visually appeared to form stable solutions were examined by TEM to check for visible signs of aggregation and the effect of ligand exchange on nanoparticle core size. TEM images and size analyses are presented in the Supporting Information (Figures S1‒S3, Table S1). All ligand exchanged nanoparticles were generally the same size, within experimental error, and shape as the parent IONP-OA and well dispersed. Some areas of higher contrast were observed on the grid for IONP-DA (HCl), IONP-SA (DI water added), IONP-3,5-DHBA (HCl), IONP-2,6-DHBA (DI water) prepared samples. The fine higher contrast material is probably deposited free iron, likely a side product that was formed during the ligand exchange reaction. Previous researchers have observed the degradation of iron oxide nanoparticles by DA34 and for other ligands known to have a high affinity for iron.11, 22 Identification of binding mode by solubility studies and FTIR-ATR. After the ligand exchange mixtures were purified, the solubility of the resulting products was determined by visual inspection. IONP-OA is soluble in organic solvents such as hexane, DCM, chloroform, and toluene. In general, changes in nanoparticle solubility from DCM (the initial reaction solvent) to more polar solvents demonstrate that ligand exchange of oleic acid for the target ligand was successful. Solubility properties for all nanoparticles studied are provided in Table 1.

Table 1. Solubility of Nanoparticles Under Study in a Variety of Polar Solvents and Buffers Ligand (additive)

DMF

DMSO

DI water

0.2 M Tris 20 mM Sodium pH 9.26 phosphate

20 mM Sodium acetate pH 5

10 ACS Paragon Plus Environment

Page 11 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

pH 7.01 3,5-DHBA (HCl)

MS

S

S

SS‒PS

PS‒MS

MS

3,5-DHBA (DI water)

MS‒S

S

PS

PS





DA (HCl)

SS

MS‒S

S

INS

INS

S

3,4-DHBA (HCl)

INS



INS

INS

INS



3,4-DHBA (DI water)

SS‒PS

PS‒MS

INS‒SS MS‒S

INS‒PS

INS‒SS

3,4-DHPA (HCl)

MS (UNS)

MS‒S

INS‒SS S

PS‒MS

INS‒PS

3,4-DHPA (DI water)

SS

PS

INS‒SS S

S

INS‒SS

2,3-DHBA (HCl)

S

S

INS

SS

SS



2,3-DHBA (DI water)

S (UNS)

S

SS

PS

S

SS

2,5-DHBA (HCl)

S

S

INS

PS‒SS

PS



2,5-DHBA (DI water)

S





PS

PS



SA (HCl)

S

S



SS





SA (DI water)

S

S



INS/SS





2,4-DHBA (HCl)

MS

S

INS

INS





2,4-DHBA (DI water)

S

S

INS

INS‒SS

INS‒SS

INS

2,6-DHBA (HCl)

INS

INS‒SS INS

SS

PS

INS

2,6-DHBA (DI water)

PS

MS‒S

INS‒SS

SS

INS

SS

Key: In order of decreasing solubility, S (soluble); MS (mostly soluble); PS (partially soluble); SS (sparingly soluble); INS (insoluble). UNS (unstable), nanoparticle suspension flocculated over the course of a day. “—”, no data available.

To better understand and possibly predict how each ligand affects the surface properties of the IONPs, it is important to know how the ligand interacts with the iron oxide surface. This information can potentially be extracted from FTIR spectra. Carboxylate stretch position and separation (∆v) can assist in the assignment of the binding mode for carboxylic acid based 11 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 53

ligands to metals.35‒41 Here, the separation (∆v) observed for each nanoparticle or ligand salt is calculated (∆v = v(COO-)as - v(COO-)s) and compared alongside literature examples of ironligand complexes and ionic values for the ligands under study. A summary of the correlations of separation to binding mode is given in Table 2, and a pictorial description of carboxylate, along with salicylate and catechol, binding modes is presented in Figure 1.

Figure 1. Carboxylate (A) bidentate chelate, (B) bidentate bridging, and (C) monodentate coordination. Both catechol and salicylate binding modes can be mononuclear or binuclear in nature.26 Bidentate mononuclear “salicylate” binding mode is shown in (D), bidentate binuclear salicylate binding in (E), bidentate mononuclear catechol chelate in (F), and bidentate binuclear catechol binding in (G). Comparing the position and separations of a complex’s carboxylate stretches to the ligand’s corresponding sodium salt is valuable towards a more accurate assignment of binding mode.35 Vibrations associated with the sodium or potassium salt of the ligand are generally similar to the free carboxylate ion,37 as the structure (C‒O bond lengths and OCO bond angle) of the salt does

12 ACS Paragon Plus Environment

Page 13 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

not significantly deviate from the free ion.36 Salicylate binding mode (Figure 1D and 1E) is supported when data shows monodentate binding of the ligand carboxylate with the metal as well as deprotonation and involvement of the 2-OH upon ligand association. Catechol binding may occur if ligands possess two phenol groups ortho to one another (Figure 1F and 1G). As the generalizations in Table 2 are primarily based on the analysis of theoretical36 and empirical37 data of metal-acetate complexes, care should be taken to use other characterization modalities to achieve a more confident picture of ligand binding.39 Analysis of carboxylate stretching modes is not straightforward as the vibrations can be influenced by several factors including: hydrogen bonding, electronic effects, symmetry of the coordination, coupling with nearby COO- groups or other vibrations, and the metal involved in binding.35‒37, 39, 42 Furthermore, spectra interpretation is itself difficult because of many overlapping bands in the region.42

Table 2. Correlation of Separation (∆v) and Carboxylate Binding Mode35‒37, 40

Binding mode

Relationship between ∆v versus ∆v (free carboxylate ion)

Bidentate chelate ∆v ∆v > 150 cm-1

Monodentate

∆v > ∆v (free)

∆v > 200 cm-1

Ionic

∆v ~ ∆v (free)

∆v ~ 150‒210 cm-1

Bidentate bridging

A summary of binding modes identified for each ligand studied, under a given reaction condition, is provided in Table 3. An analysis of the binding patterns for these ligands is the subject of the following sections.

13 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 53

Table 3. Summary of Ligand Binding Modes for a Given Ligand Exchange Preparation Ligand

Binding mode

Oleic acid (OA)

CA (chelate) HCl(aq) addeda

DI water added

3,5-DHBA

CA

CA (chelate or bridging)

Dopamine (DA)

CAT

did not produce water stable IONP

3,4-DHBA

did not produce stable IONP

CAT, CA (monodentate, minorb)

3,4-DHPA

CAT, CA

CAT, CA

2,3-DHBA

CA (bridging), CAT (very minor)

CA (bridging), CAT (minor)

2,5-DHBA

SA, CA (chelate)

SA, CA (chelate)

Salicylic acid (SA)

CA (bridging and chelate)

CA (bridging and chelate)

2,4-DHBA

CA (bridging)

CA (bridging)

2,6-DHBA

did not produce stable IONP

Inconclusive

a

Either HCl(aq) or DI water is added to the exchange reaction. Key: CA, carboxylate; CAT, catecholate; SA, salicylate. bIndicates a particular binding mode is likely present in minor quantities.

The position of hydroxyl and carboxylic acid groups influence ligand binding. 1. The main determinant of ligand binding structure for the studied ligands is the presence and relative positioning of functional groups (hydroxyl, carboxylic acid, and amine) in the molecule. The simplest examples of this observation is the carboxylate binding of OA and 3,5-DHBA, where the presence of an isolated carboxylic acid is responsible for binding of the ligand to iron oxide. Solubility of IONP-OA in non-polar solvents is consistent with oleic acid being bound to the surface through the carboxylic acid group with the extended hydrophobic chain in solution. Evidence of oleate binding through the carboxylic acid in both batches of IONP-OA is clearly given in analysis of the FTIR spectra acquired. The C=O stretch found at 1708 cm-1 in oleic acid 14 ACS Paragon Plus Environment

Page 15 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

is not found in the IONP-OA, and instead replaced by two COO- stretches at 1520/1519 cm-1 and 1428/1425 cm-1, corresponding to the asymmetric and symmetric COO- stretch (respectively) of the attached oleate ligand. Bands associated with the aliphatic tail of oleic acid, v(=C‒H), vas(CH3), vas,s(CH2), and δs(CH3) are observed in oleic acid, sodium oleate, and remain in both IONP-OA spectra. These findings are consistent with oleic acid attaching to the nanoparticle through the carboxylic acid. The Fe‒O stretch is observed at 591 and 580 cm-1 for IONP-OA, corresponding well with literature of other Fe3O4 nanoparticles.43‒46 Further FTIR analysis of the structure of the iron oxide nanoparticle core is given in the Supporting Information (SI). A summary of bands and assignments for IONP-OA (Table S4, SI), as well for all ligands and associated nanoparticles (Tables S4‒S12, SI), is provided in the Supporting Information. In addition to the works cited, insights regarding the assignments of all ligands and carboxylate stretches were also provided by Larkin47 and Lambert et al.48 All IONP-OA and ligand exchanged nanoparticle spectra presented can be found in the Supporting Information as single spectra (Figures S4‒S14, SI). The calculated ∆v (92, 94 cm-1) for IONP-OA is less than the ionic value (sodium oleate, ∆v = 115 or 135 cm-1) and below 105 cm-1, a region typically associated with chelating bidentate association (Table 2). The ∆v observed for both preparations of IONP-OA is very similar, and both IONP-OA preparations align with the separation observed for isolated iron-oleate complex.38 Other preparations of oleic acid capped iron oxide nanoparticles have also assigned the association of oleic acid to chelating bidentate, based on the separation values observed (∆v = 80 to 110 cm-1).13, 14, 49, 50 Based on the observed solubility of the nanoparticles, bands observed in the FTIR spectra, and ∆v for IONP-OA (92 cm-1 and 94 cm-1, Table 4), the binding mode of oleic acid to IONP is likely chelating bidentate (Figure 1A).

15 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 53

Table 4. Asymmetric (vas(COO-)) and Symmetric (vs(COO-)) Carboxylate Stretches, ∆v, and Proposed Binding Mode for IONP-OA and IONP-3,5-DHBA Determined by FTIR

Nanoparticle or complex

v(COO-)as (cm-1)

v(COO-)s ∆v (cm-1) (cm-1)

Proposed binding mode

Referencea

IONP-OA (Batch 1) IONP-OA (Batch 2) Fe–oleate

1520

1428

92

Carboxylate (chelate)

This work

1519

1425

94

Carboxylate (chelate)

This work

1527

1436

91

Carboxylate (chelate)

Na–oleate

1559 1523

115 or 135 127

Ionic

IONP-3,5DHBA (DI water, water washed) Na-3,5-DHB

1444 or 1424 1396

Bronstein et al.38 This work

1565

1413

152

Ionic

Carboxylate (chelateb or bridging)

This work

Kalinowska et al.51 a Comparisons to metal-ligand complexes in the literature are provided as reference. bMost likely coordination mode is underlined. Key: DHB, dihydroxybenzoate.

Similar to OA, binding of 3,5-DHBA to the iron oxide nanoparticle appears to be occurring solely through the carboxylic acid, although the analysis is complicated by the likely presence of non-specifically adsorbed ligand in addition to the surface bound material. Given that a bound carboxylate can no longer participate in charge stabilization of the nanoparticle in water or buffered solutions, IONP-3,5-DHBA (water washed DI water preparation) particles are only soluble in DMSO and DMF, as expected. In comparison, IONP-3,5-DHBA prepared with HCl is soluble in DI water, and to some extent in buffer, suggesting that there is non-specific binding of the ligand to the nanoparticle.

16 ACS Paragon Plus Environment

Page 17 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Infrared spectra of IONP-3,5-DHBA (water washed DI water preparation) (Figure 2) show the removal of oleic acid from the ligand sphere, as bands associated with CH2 groups of oleic acid (vas, s(CH2), 2920 cm-1 and 2851 cm-1) have greatly decreased or are not present, and are instead replaced with vibrations associated with the aromatic ring of 3,5-DHBA (Table S5). The C=O stretch observed in 3,5-DHBA (1682 cm-1) is still present in all but the DI water washed nanoparticles, and along with the presence of the v(C‒O) (O=C‒OH) mode in the HCl added nanoparticles (at 1693 cm-1 or 1694 cm-1) suggests incomplete removal of the non-specifically adsorbed ligand in most cases. Washing with water is able to remove non-specifically adsorbed ligand from IONP-3,5-DHBA (DI added) nanoparticles, which is unexpected, as 3,5-DHBA itself appears much more soluble in alcohol compared with water. The nature of the association of the excess ligand with the nanoparticle is unclear, however, our data is consistent with a hydrogen bonding type interaction, since water disrupts the association while alcohol does not. Unfortunately, washing of nanoparticles prepared with HCl is impossible, as the resulting nanoparticles are very soluble in DI water. In this case, even extensive washing with methanol or a mixture of 1:1 methanol/isopropanol is not destabilizing enough to the adsorbed ligand, as washing with alcohol does not significantly change the FTIR spectra obtained (data not shown). In all 3,5-DHBA nanoparticle spectra, the C‒O (Ar‒OH) stretch remains (1204‒1209 cm-1), strengthening the case that complexation to the nanoparticle is not through the 3- or 5-OH groups, and aligning with prior literature results that demonstrate limited involvement of the isolated hydroxyl groups in ligand binding to other iron oxides.24, 25, 52 Bands assigned to vas(COO-) are observed to some degree in all spectra of 3,5-DHBA functionalized nanoparticles, however vs(COO-) is only clear in the spectrum of DI water washed nanoparticles. Identifying vs(COO-) in the other spectra is difficult, due to very broad absorption

17 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 53

in the region and close proximity to the β(C‒OH) vibrations. In contrast to the other IONP-3,5DHBA spectra, in the spectra of DI water washed nanoparticles the C=O vibration is replaced with modes for v(COO-) at 1523 and 1396 cm-1 (∆v = 128 cm-1). The asymmetric and symmetric COO- vibrations are both red-shifted with respect to the sodium salt, and the ∆v calculated for IONP-3,5-DHBA is below that of the sodium salt51 (∆v Na-3,5-DHB = 152 cm-1, Table 4) and in the range for values typically associated with bidentate bridging coordination. Typically when the vas(COO-) observed is lower than that of the sodium salt chelate binding is implicated.35, 39 Thus, chelate coordination cannot be ruled out completely, but the vs(COO-) band is rather broad so this result is not diagnostic. Based on solubility observations and FTIR data, 3,5-DHBA appears to bind to the IONP surface in a bidentate chelate or bridging fashion. Modification of TiO2 with SA, 2,5-DHBA, 2,3-DHBA, 3,4-DHBA, and catechol were previously found to result in the formation of metal-ligand charge transfer complexes accompanied by a change in optical properties.26 In contrast, significant changes in optical absorption were not observed for upon the surface modification of IONP with 3,5-DHBA, 3,4DHBA, or 3,4-DHPA (SI, Figure S15 and Table S13). There is no evidence of additional bands in the UV‒visible region that would suggest the formation of a charge transfer complex upon ligand exchange of OA for the listed ligands.

18 ACS Paragon Plus Environment

Page 19 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 2. FTIR spectra showing the replacement of oleic acid (OA) for 3,5-dihydroxybenzoic acid (3,5-DHBA) on the surface of OA functionalized iron oxide nanoparticles (IONP-OA). The resulting 3,5-DHBA functionalized nanoparticles (IONP-3,5-DHBA, water washed DI water preparation) present bands consistent with attachment of 3,5-DHBA by the carboxylic acid and the removal of most of the OA originally present. 2. Ligands possessing two adjacent hydroxyl groups (DA, 3,4-DHPA and 3,4-DHBA, 2,3DHBA) demonstrate catechol binding to varying degrees.

19 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 53

While dopamine (DA) and 3,4-DHPA clearly show catechol binding to the nanoparticle surface, 3,4-DHBA binding occurs mostly through the catechol with a minor contribution from carboxylate (monodentate) binding, and mainly carboxylate (bridging) binding is observed for 2,3-DHBA with some evidence of catechol binding. DA is thought to coordinate to the iron oxide surface via the phenol groups versus the amine,11, 32 evidenced here through analysis of the solubility behavior and FTIR spectra of DA functionalized nanoparticles (IONP-DA). IONP-DA is completely soluble in DI water and in sodium acetate pH 5. These results are consistent with the protonation of the dopamine amine at pH < pKa NH2 (pKa NH2 = 9.05),53 which would supply a net positive charge to the nanoparticle, rendering it soluble under neutral and acidic conditions. It is notable that for the ligand exchange to occur to an extent that produces water soluble nanoparticles, the addition of HCl to the reaction mixture appears to be required. This behavior is consistent with previous results of Fe3O4 functionalization with DA and Tiron when amylene stabilized solvents are used.11 Bands assigned to v(CH2)as and v(CH2)s are much less intense in the IONP-DA spectrum, compared with IONP-OA, and carboxylate stretches are absent (at 1519 cm-1) or weak (1427 cm1

) showing the removal of oleate from the ligand sphere of IONP-DA. In-plane C‒H bending

vibrations found in both DA (1146, 1114 cm-1) and IONP-DA (1150, 1121 cm-1) are similar to those found in IONP-3,4-DHPA (1147 or 1148 cm-1, 1114 cm-1) an analogous ligand, and provide more evidence for the attachment of DA to the iron oxide nanoparticle. The bands for δs(NH2) (scissor) (1605 cm-1, in IONP-DA), β(C‒C‒N) (1267 or 1222 cm-1), and v(C‒N) (1088 or 1012 cm-1) for DA remain in the spectrum for IONP-DA, consistent with the observed solubility properties of IONP-DA. If the dopamine ligand were binding to the iron oxide through the amine, these vibrations would be affected, and likely the resulting nanoparticle would not be

20 ACS Paragon Plus Environment

Page 21 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

soluble in buffer. In contrast, bands for β(C‒O‒H) (1393, 1321 cm-1) and v(C‒O) (Ar‒OH) (1285 cm-1) observed in the DA spectrum are not found in the spectrum of IONP-DA, consistent with catechol binding of the ligand. It is possible that the vibration at 1267 cm-1 is v(C‒O) when DA is adsorbed, since this value is similar to that previously observed by Amstad et al. for v(C‒ O) (1265 cm-1) upon DA binding to Fe3O4 nanoparticles.22 Comparison of relevant bands for binding, those relating to the carboxylic acid and phenol groups, for 3,4-DHPA, 3,4-, and 2,3-DHBA (Table 5 and Figures S16 and S17, SI) help illustrate the differences in binding among these similar ligands. Bands for v(C‒O) (O=C‒OH) were not included in the comparison, since these bands were generally difficult to assign due to many overlapping or coupling absorbances in the area making any comparisons likewise difficult.

Table 5. Comparison of Carboxylic Acid and Phenol Related Bands (in cm-1) and ∆v Observed in the FTIR Spectra of the Ligands 3,4-DHPA, 3,4-DHBA, and 2,3-DHBA and their Corresponding Functionalized Nanoparticles.

Ligand or functionalized nanoparticle 3,4-DHBA

v(C=O)

v(COO-)as

v(COO-)s

∆v

v(C‒O) (Ar‒OH)

β(C‒OH)

1654







1392, 1364

Na-3,4-DHB

1679w

1541

1380

161

1244sh, 1221 1242, 1208

1350

194

not visible

1388a

IONP-3,4-DHBA (DI water) 3,4-DHPA Na-3,4-DHP

IONP-3,4-DHPA (HCl)b (DI water)

a

1654

1582

1388

1682 1692

— 1522 (~1560‒ 1522)

1416, 1373 1434, 1370

1524

— 123 (123‒ 190) 118

1244 1254

1700

— 1399 (~1399‒ 1370) 1406

1216

1702

1524

1400

124

1217

not resolvable 1423sh

21 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2,3-DHBA

1678, 1656





Page 22 of 53



1255, 1232

1433, 1381, 1349 1383, 1355 not visible

Na-2,3-DHB 1644w 1568 1421 147 1230 IONP-2,3-DHBA not visible 1543 1391 152 1258, 1233 (HCl) (DI water) not visible 1537 1397 140 1257, 1231 not visible a b Band may be part of a larger absorption. Data shown for 2x washed nanoparticles, see SI for details. Key: v, stretch; β, in-plane bending; sh, shoulder; w, weak; DHB, dihydroxybenzoate; DHP, dihydroxyphenylacetate; “—”, not applicable.

Since the structures of 3,4-DHPA and 3,4-DHBA are very similar and possess the same functional groups, they were expected to bind to the IONP in a similar fashion, although the increased carboxylic acid substituent length may affect ligand complexation. Indeed, the resulting solubility and features in their FTIR spectra for IONP-3,4-DHBA (DI water preparation) and IONP-3,4-DHPA are similar. Solubility of these nanoparticles is in agreement with the carboxylic acid group being free and ionized in solution, since 3,4-DHPA ligand exchange products are aqueous stable and IONP-3,4-DHBA (DI water preparation) is mostly soluble to soluble in 0.2 M Tris pH 9.26 implying the nanoparticles are electrostatically stabilized by the negative carboxylate ion. This solubility behavior agrees with electrokinetic measurements of IONP-3,4-DHBA (DI water preparation), which exhibits a highly negative zeta potential (-55.09 ± 2.77 mV) and electrophoretic mobility (-4.30 ± 0.22 10-8 m2s-1V-1) in 0.2 M Tris pH 9.21 (measurement details in SI, with nanoparticle preparation modifications Table S13). IONP-3,4-DHPA (DI water preparation) also showed a negative zeta potential (-33.94 ± 1.90 mV) and mobility (-2.65 ± 0.15 10-8 m2s-1V-1) with the same buffer and nanoparticle concentration, however its magnitude was significantly smaller. The observed difference may be a result of increased carboxylic acid binding versus catechol binding for 3,4-DHPA in comparison to 3,4-DHBA, and would agree with the relative representation of binding modes

22 ACS Paragon Plus Environment

Page 23 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

determined for each ligand (Table 3). Alternatively, it is possible that lower ligand loading occurs in the case of 3,4-DHPA. Both possibilities would lead to a smaller amount of solution exposed carboxylic acid for IONP-3,4-DHPA, and therefore decreased negative surface charge, in agreement with the electrokinetic data obtained. Similar solubility properties were observed in previous preparations of Fe3O4 nanoparticles stabilized with 3,4-DHBA54 and 3,4-DHPA,16 where the resulting 3,4-DHPA coated nanoparticles could be redispersed in 10 mM phosphate buffered saline (PBS), pH 6.3 and 3,4-DHBA nanoparticles in water and PBS. For the 3,4-DHPA particles prepared by Yun et al.,16 the successful functionalization of the resulting 3,4-DHPA stabilized nanoparticles with carboxyl-(polyethylene glycol)8-amine, via carbodiimide (sulfoNHS/EDC) coupling chemistry showed that the carboxylic acid of 3,4-DHPA remained solution accessible after ligand binding to the iron oxide nanoparticle. In contrast, IONP-3,4-DHBA prepared with the addition of HCl was insoluble in all of the solvents tested. Likely some ligand exchange of 3,4-DHBA occurs under HCl added conditions, since the nanoparticles are no longer soluble in DCM, implying some of the stabilizing OA is removed from the nanoparticle, but that an insufficient amount of 3,4-DHBA is exchanged to yield aqueous stable nanoparticles. By FTIR, it was observed that the initial purification of IONP-3,4-DHPA (HCl preparation) is not sufficient to remove excess ligand or by-products from the reaction mixture. Extra washing with acetone (SI experimental and Figure S18, SI) results in HCl and DI water preparations of nanoparticles with similar spectra, demonstrating the utility of FTIR to monitor purification. Bands for v(C=O) are still visible in the spectra of IONP-3,4-DHBA (1654 cm-1) and IONP-3,4DHPA (1700, 1702 cm-1) nanoparticles, indicating the presence of free carboxylic acid (Figure 3). Carboxylate stretches for IONP-3,4-DHBA, Na-3,4-DHP, and IONP-3,4-DHPA are difficult to assign because of their close positioning to other ring vibrations or C‒OH deformation modes

23 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 53

and broad absorbance in the area. For IONP-3,4-DHPA, the bands at 1524 cm-1 and 1406/1400 cm-1 mirror closely those found for v(COO-) in the ligand sodium salt (1522 cm-1 and 1399 cm-1), which is unusual as such vibrations are usually very sensitive to the metal involved.39 If they are in fact carboxylate stretches, the presence of such bands indicates some contribution from carboxylate binding, or alternatively, these bands may be attributed to other unidentified modes for the ligand. In that case, the DHPA ligand is binding solely through the catechol, and the difference between the HCl and DI water preparations is a difference in ligand loading and not binding mode. The bands for v(C‒O) (Ar‒OH) are either disturbed (3,4-DHPA) or not visible (3,4-DHBA) in the nanoparticle spectra, suggesting these groups are involved in ligand binding.

24 ACS Paragon Plus Environment

Page 25 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 3. FTIR spectra of 3,4-DHBA and 3,4-DHPA outlining key bands relating to the carboxylic acid and phenol groups with evidence of semiquinone (SQ) v(C‒O) bands ~1490 cm1

. In addition, we see evidence of semiquinone v(C‒O) bands in the spectra of IONP-DA (1485

cm-1), IONP-3,4-DHPA (1487, 1486 cm-1), IONP-3,4-DHBA (1491 cm-1), and Na-3,4-DHB (~1490 cm-1). These values match well with previous observations of semiquinone v(C‒O) vibrations between 1500 cm-1 and 1480 cm-1,55, 56 and with other preparations of DA modified Fe3O4 (1484 cm-1).22 A relatively more intense semiquinone band exists in the spectra of IONP3,4-DHPA prepared without HCl, versus HCl added, which implies greater semiquinone production under these conditions. All ligand reactions and ligand salt syntheses here were performed under aerobic conditions, which may have encouraged semiquinone generation. Behavior of bands for the in-plane bending mode for C‒OH parallels the proposed binding behavior of 3,4-DHPA and 3,4-DHBA. Previous reports for substituted benzoic acids have assigned the higher β(C‒OH) vibration to that originating from COOH, and the lower vibration from Ar‒OH.57 For IONP-3,4-DHPA, only the higher band β(C‒OH) is still observed (1423 cm1

), originating from the free carboxylic acid, and in IONP-3,4-DHBA (DI water) the lower band

at 1350 cm-1 in the sodium salt has disappeared suggesting phenol binding for both ligands. In contrast, most of the 2,3-DHBA ligand appears to be binding via a carboxylate mode, since the v(C‒O) (Ar‒OH) vibrations continue to have high intensity. Some catechol binding may be occurring, as a broad shoulder from approximately 1700‒1610 cm-1 is observed with a greater relative intensity in the DI water preparation, especially around 1650 cm-1, which is associated with v(C=O) of unionized carboxylic acid. This observation follows the solubility behavior for 2,3-DHBA functionalized nanoparticles, where particles are more soluble in aqueous buffers in

25 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 53

the DI water preparation, suggesting more of the carboxylic acid is solvent accessible and therefore not participating in binding, electrostatically stabilizing the nanoparticles in solution. Since the absorbance around 1650 cm-1 is much less prominent than what is seen in 3,4-DHBA or 3,4-DHPA, likely more carboxylic acid binding occurs than for either of the two above ligands. Instead, prominent bands for v(COO-) are located at 1543/1537 cm-1 and 1391/1397 cm-1 which are red-shifted with respect to the sodium salt and with ∆v values close to ionic, indicative of carboxylate bridging coordination. For 2,3-DHBA both sets of β(C‒OH) bands are not strong, possibly part of a larger absorption, or not present, suggesting both the carboxylic acid and phenol groups are involved in binding.

3. The presence of an ortho hydroxyl group allows for the ligand interaction to proceed through both the carboxylic acid and the adjacent 2-OH. This observation is most clearly demonstrated by comparing the proposed binding modes of 3,5-DHBA and 2,5-DHBA, where 2,5-DHBA shows evidence of ligand binding through a combination of the ortho-hydroxyl and the carboxylic acid (“salicylate”-type binding) in addition to bidentate coordination through the carboxylic acid. This agrees with earlier work where 2,5DHBA was also found to bind to aluminum hydroxide27 and Fe(III)58 in a salicylate configuration. Here, small shoulders at 1628 cm-1 and 1630 cm-1 for the IONP-2,5-DHBA HCl and DI water preparations, respectively, can be assigned to the asymmetric COO- stretch, and bands associated with the symmetric COO- stretch at 1387 cm-1 and 1386 cm-1, give ∆v values (241 cm-1 and 244 cm-1) far above that of Na-2,5-DHB (∆v = 202 cm-1)59 (Table 6) suggesting monodentate coordination with respect to the carboxylate (Figure 4). Weak bands around 1350 cm-1 seen in both preparations of IONP-2,5-DHBA may be vs(COO-), and would still result in

26 ACS Paragon Plus Environment

Page 27 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

values of separation consistent with monodentate binding of the carboxylate. In general, the position of the carboxylate stretching vibrations are difficult to assign for benzoates given that the attached aromatic ring possesses atoms of similar mass, comparable bond strengths, and vibrational symmetry as the carboxylate group which can often lead to coupling between vibrations60 and making assignments in this area (~1650‒1300 cm-1) difficult or at the very least complex. A rather broad v(C‒O) (Ar‒OH) band remains in both IONP-2,5-DHBA preparations, however a 14 cm-1 red-shift in band position for the DI water case indicates a change in the strength of the C‒O bond and implies more of the coordination proceeds through the o-OH under different reaction conditions. This is also consistent with prior work, where a shift of this band to lower frequencies was observed for the binding of salicylic acid to goethite in a salicylate binding mode.61 A red-shift in this vibration is also seen in the binding of 2,5-DHBA to Fe(III),28 compared to free 2,5-DHBA, suggesting the ortho phenol is involved in binding and further strengthening our claim here. Our conclusions are also consistent with the more pronounced shoulders at around 1630 cm-1 (vas(COO-) in the DI water preparation (Figure 4) that have been associated to salicylate-type binding of 2,5-DHBA.

Table 6. Asymmetric (vas(COO-)) and Symmetric (vs(COO-)) Carboxylate Stretches, ∆v, and Proposed Binding Mode for IONPs Functionalized with SA, 2,5- and 2,4-DHBA as Determined by FTIR

Nanoparticle or complex IONP-2,5-DHBA (HCl)

v(COO-)as (cm-1) 1628sh, 1548

v(COO-)s (cm-1) 1459, 1387/1351

∆v (cm-1) 241/277, 89

Proposed binding mode

Referencea

Salicylate, carboxylate (chelate)

This work

27 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

IONP-2,5-DHBA (DI water)

1630sh, 1541

1454, 1386/1350

244/280, 87

Na-2,5-DHB

1587

1385

202

IONP-SA (HCl)

1602, 1527

~1409sh, w, 1390

212, 118

IONP-SA (DI water)

1601, 1527

~1409sh, w, 1390

211, 118

Na-salicylate

1596

1403

193

~1585 1602

1391 1389

~194 213

IONP-2,4-DHBA (HCl, 6 washes acetone) IONP-2,4-DHBA 1600 1386 214 (DI water) Na-2,4-DHB 1557 1345 212 a Ligand sodium salts from this work or the literature are shoulder; w, weak; DHB, dihydroxybenzoate.

Page 28 of 53

Salicylate, carboxylate (chelate) Ionic Carboxylate (bridging and chelate) Carboxylate (bridging and chelate) Ionic Ionic Carboxylate (bridging)

This work

Regulska et al.59 This work

This work

Lewandowski et al.62 Lindberg63 This work

Carboxylate This work (bridging) Ionic This work provided for comparison. Key: sh,

28 ACS Paragon Plus Environment

Page 29 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 4. Comparison of nanoparticle (IONP-2,5-DHBA) FTIR spectra (2000‒1000 cm-1) obtained from both ligand exchange reactions of IONP-OA with 2,5-DHBA. Bands at 1628/1630 cm-1 are slightly more pronounced in the spectra for the DI water preparation. This observation coupled with a 14 cm-1 redshift, and change in band shape, of the v(C‒O) (Ar‒OH) absorption at 1223 cm-1 suggests that more salicylic binding takes place under no HCl added conditions. It is not surprising, based on their structure, that the ligand exchange products for SA, 2,4-, and 2,5-DHBA are similar in terms of their FTIR spectra and solubility. No C=O stretch is visible in any of the purified SA, 2,5, or 2,4-DHBA nanoparticle spectra. Instead this stretch is replaced with bands associated with v(COO-), therefore supporting the involvement of the carboxylic acid 29 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 53

in ligand binding for all the above ligands (Table 6). Weak methylene vas,s(CH2) bands, from oleic acid, are present in the spectra of SA, 2,4-, and 2,5-DHBA functionalized nanoparticles (Figure S19, SI), indicating that the ligand exchange does not go fully to completion. This is not an unexpected finding, especially given the results from Davis et al.64 that demonstrate incomplete exchange for a variety of ligands on oleic acid functionalized iron oxide nanoparticles. Solubility of SA, 2,5-, and 2,4-DHBA functionalized nanoparticles are similar, pointing to a similar ligand binding structure, and consistent with the carboxylic acid being sequestered in binding. The binding of 2,6-DHBA is unlike the other o-OH ligands. This is likely a consequence of having two ortho hydroxyl groups versus one, and results in ligand exchange products that are insoluble in any solvents tested (HCl preparation) or with moderate solubility in polar aprotic solvents (DI water preparation). Unfortunately FTIR data of 2,6-DHBA adsorbed to iron oxide nanoparticles could not be obtained due to nanoparticle insolubility and therefore incompatibility with our experimental setup. However, the evolution of deep blue coloration during the reaction of 2,6-DHBA with the iron oxide nanoparticles suggests that either too strong of a bond forms between the ligand and nanoparticle, or that acidic dissolution of the nanoparticle occurs resulting in removal of iron from the nanoparticle surface in either case. The UV‒visible spectra of supernatants (λmax,vis = 572 nm) recovered from the ligand exchange reaction are indeed very similar to the solution spectra of prepared Fe(III)-2,6-DHBA complexes (λmax,vis = 565 nm) (Figure S20, SI). 4. For a given ligand, mixed binding modes are possible when multiple binding modes are available.

30 ACS Paragon Plus Environment

Page 31 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

There is evidence that mixed binding of 2,5-DHBA to the iron oxide nanoparticle surface occurs, through the participation of the ortho-hydroxyl and monodentate carboxylate as well as through a bidentate chelate coordination. The separations between asymmetric and symmetric stretch observed in the bidentate chelate coordination of 2,5-DHBA with FeCl328 at pH 5.2 (1549 cm-1 - 1454 cm-1 = 95 cm-1) correspond well with the calculated ∆v for both preparations of IONP-2,5-DHBA (89/87 cm-1 (HCl/DI water)) and provide further support for the chelate mode of coordination for 2,5-DHBA and Fe3O4 nanoparticles. Likewise, mixed binding is also noted for 2,3-DHBA where both carboxylate (bridging) and a minor contribution from catechol binding are indicated, with more catechol binding occurring under HCl free reaction conditions. 5. Even though some ligand structures would allow multiple possible binding modes, this does not guarantee that the ligands will assume each or any given possible interaction. The most surprising binding case is for salicylic acid, as “salicylate”-type binding has been observed in the literature for other iron oxides such as goethite (α-FeOOH),31,

61

hematite (α-

Fe2O3),31 and lepidocrocite (γ-FeOOH).31 Instead, for both preparations of IONP-SA, the C‒O (Ar‒OH) stretch is not perturbed upon ligand adsorption, and values for ∆v are consistent only with bidentate coordination of the carboxylic acid. Similarly, 2,4-DHBA only shows evidence of carboxylate binding (bridging) on the nanoparticle surface, as ∆v values are close to that of the sodium salt of the ligand, and values of vas,s(COO-) both symmetrically shift to higher wavenumbers with respect to the sodium salt.39 Both sets of v(C‒O) (Ar‒OH) bands, at ~1250 cm-1 (2-OH) and ~1225 cm-1 (4-OH), remain visible in both IONP-2,4-DHBA nanoparticle spectra and at similar positions, indicating they are not significantly involved in binding. The value of v(C‒O) (Ar‒OH) for 2-OH in 2,4-DHBA is probably ~1250 cm-1, as this band blueshifts upon preparation of the sodium salt, likely a consequence of a stronger intramolecular

31 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 53

hydrogen bond in the ionic carboxylate resulting in increased electron density at the phenol oxygen and a stronger C‒O bond.31 Similarly, it was not expected that 2,3-DHBA would primarily bind to the iron oxide surface through the carboxylate, given the generally high interaction strength of catechols for iron ([Fe(cat)3]3-, Kstability = 1044.9 versus [Fe(oxalate)3]3-, Kstability = 1018.49).21 Like for 2,4-DHBA, only a carboxylate bridging interaction agrees with the values of ∆v observed for IONP-2,3DHBA, even though the ligand possesses an ortho hydroxyl that could participate with the carboxylic acid to form a salicylate-type bond to the nanoparticle surface. Absence of a strong C=O stretch in 2,3-DHBA spectra suggests the carboxylic acid is involved in binding regardless of reaction conditions. Ligand binding is affected by the addition of HCl(aq). Adding HCl(aq) to the reaction mixture affects the solubility, FTIR spectra of the resulting nanoparticle products, and apparent ligand binding for all the ligands studied except for SA and 2,4-DHBA. The most striking examples of this behavior are that of DA, where water stable nanoparticles are not produced without the addition of HCl, and that of 2,6-DHBA and 3,4-DHBA, whose ligand exchange products are not appreciably soluble in any of the solvents tested when HCl is added. In general, greater shifts in solubility and FTIR spectra are observed for the catechol containing ligands (DA, 2,3-DHBA, 3,4-DHBA, and 3,4-DHPA) versus those containing only a 2-OH (2,4-DHBA, 2,5-DHBA, SA). The ligand exchange reaction on the iron oxide nanoparticle can be broken down into three interacting parts: (1) deprotonation of the oncoming ligands, (2) the competition and removal of the original ligand, OH-, or H2O coordinating the surface iron sites, and (3) the interaction between iron and the new ligand. The effect of adding HCl(aq) to the reaction mixture likely

32 ACS Paragon Plus Environment

Page 33 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

affects these stages through a few connected mechanisms, namely via electronic effects, by altering the protonation state of the ligands, and through adding Cl- to the reaction. Electronic effects are likely modulating metal-ligand bond strength and proton dissociation, and pH effects primarily affecting ligand protonation state. Electronic effects, due to the number and position of substituents, are reflected by the pKa values of ionizable groups of the ligands (Table 7) and influence both the deprotonation of the substituents and the Lewis basicity of the ligand. This property is demonstrated in both the lower pKa values for the 2-OH ligands, with the extreme being 2,6-DHBA, which has two ortho OH groups and the lowest pKa value, and the observed dominance of carboxylate binding for these ligands on iron oxide nanoparticles. Increased electron withdrawal at a particular ionizable group results in a decreased pKa value65 and, as in the case of salicylic acid, the presence of a 2-OH results in stabilization of the carboxylate anion due to intramolecular hydrogen bonding further decreasing the carboxylic acid pKa.66 This intramolecular hydrogen bonding is also responsible for the unusually high pKa of the 2-OH.31 Ortho and para phenol groups in 2-, 4-, and 6-OH DHBAs also have a strong electron-donating resonance effect upon the carboxylic acid group, increasing electron density at the carboxyl, making it more attractive for the carboxylic acid to bind to the metal oxide,27 and ortho phenolic groups possess a weak inductive electron withdrawing effect that can decrease the carboxylic acid pKa. These effects may explain in part the successful binding and stabilization of nanoparticles with 2,3-, 2,4-, and 2,5-DHBA, in contrast to 3,4-DHBA, under HCl added conditions. Even though 3,4-DHPA is similar in structure to 3,4-DHBA, successful binding of 3,4-DHPA through the catechol even under acid added conditions may be a result of the moderate electron donation (by resonance) of the alkyl group para to the 4-OH position.

33 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 53

Table 7. pKa Values for Ligands Under Studya

Ligand

pKa1

pKa2

pKa3

Dopamine (DA)53 9.05 10.58 12.07 67 3,4-DHPA 4.17 9.49 13.7 3,4-DHBA 4.46 9.08 12.58 3,5-DHBA 4.08 9.90 11.09 2.8 13.4 N/A Salicylic acid (SA)61 2,4-DHBA 3.23 8.84 14.04 2,5-DHBA 2.96 10.46 13.41 2,3-DHBA 2.41 10.30 13.48 13.95 2,6-DHBA 1.43 13.03 a 25 Ligand pKa values from Evanko and Dzombak unless otherwise indicated. N/A: not applicable.

Solution pH has been observed to influence the coordination of aluminum oxide,27 iron oxides,25, 29, 68, 69 other metal oxides,69 and Fe(III)67 with benzoic acid and catechol derivatives possessing phenol groups adjacent to one another or to a carboxylic acid. This effect is, in part, attributed to easier deprotonation of the phenolic substituents with increasing pH,27, 29 and the competition between protons and iron for binding to basic catechols67 and phenols at low pH. Indeed, the binding of 3,4-DHBA, 3,4-DHPA, 2,3-DHBA, 2,5-DHBA, and 2,6-DHBA is altered by adding HCl to the reaction medium, which may be a result of less deprotonation of basic phenol groups, which would disfavor catechol (2,3-DHBA, 3,4-DHBA, and 3,4-DHPA) or salicylate (2,5-DHBA, 2,6-DHBA) type ligand binding to the nanoparticle under these conditions. Differences in solubility, between HCl and HCl free preparations, are generally more apparent with the catechol containing ligands under study. In previous work for situations where both catechol and carboxylate groups are present, carboxylic acid groups of dihydroxybenzoic acid dominated binding at low pH for aluminum hydroxide27 and hematite29 surfaces. Solubility 34 ACS Paragon Plus Environment

Page 35 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

and FTIR data for IONP-2,3-DHBA (no HCl) and IONP-3,4-DHPA also suggest that more catechol-type coordination occurs when reacting at more alkaline conditions, as the resulting nanoparticles are more stable in aqueous buffer and show FTIR data consistent with this observation. More engagement of the phenol in binding is also observed for 2,5-DHBA under HCl free conditions (Figure 4). These results are similar to previous observations of the complexation of 3,4-DHBA to hematite surfaces,29 where at pH ≥ 9, the surface complexation appeared to have occurred through the phenolic groups, and at pH values up to pH 8, through the carboxylic acid. In our work, the successful stabilization of the iron oxide nanoparticle by 3,4DHBA could only be accomplished when HCl is not added to the reaction medium, whereby the complexation is observed to primarily proceed through the catechol. Poor coordination of 3,4DHBA to the nanoparticle under HCl added conditions may be a consequence of its relatively high carboxylic acid pKa (4.46) with respect to the other DHBA ligands. Previously, higher phenol pKa values67, 70 have been correlated with stronger bonding to Fe3+ and higher carboxylic acid pKa values result in stronger metal-ligand σ bonds.71 However, these effects do not necessarily guarantee more ligand will adsorb,70 as we have seen for 3,4-DHBA which has the highest carboxylic acid pKa but does not bind under HCl added conditions. Future IR studies on purified dried solid IONP-3,4-DHBA (HCl added) may reveal the nature of ligand adsorption at lower pH conditions. Correlations between pKa and bond strength should be expressed with caution, however, as different orbital overlaps are involved in σ bonding between a ligand and a proton versus with Fe.21 It is instead wiser to reflect differences in ligand pKa values as indicators of changes in electron density distribution due to the addition or position of substituents, or as measures to evaluate the relative tendency for ligands to deprotonate.

35 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 53

Decreased deprotonation may also occur via electron donation (through resonance) by carboxylate (COO-) and hydroxyl groups, by lowering the acidity of the ligand at ortho and para positions. This effect can control the adsorption profile, limiting ligand binding, even if the coordination step is more favored due to the increased Lewis basicity by electron donation.70 This situation is reflected in the high pKa values for the phenol groups of 2,6-DHBA (pKa2,3), SA (pKa2), 2,4-DHBA (pKa3), and may explain in part why salicylic acid binding is not observed for SA or 2,4-DHBA, since the phenol group has more difficulty deprotonating. It is possible that we see salicylate binding for 2,5-DHBA and not SA or 2,4-DHBA due to a lower pKa3 (at the 2-OH), compared with 2,4-DHBA, and the presence of a hydroxyl group in the para position for 2,5-DHBA increasing electron density at the 2-OH. For the four most acidic ligands, 2,6-, 2,5-, 2,3-DHBA and SA, supernatants recovered from the exchange reaction were highly colored. The colors of the supernatants are consistent with the formation of Fe-ligand complexes in solution, purple-violet for SA, indigo for 2,3-DHBA and 2,6-DHBA, and green-blue for 2,5-DHBA (Table S14, SI). The more acidic ligands likely contribute to nanoparticle dissolution through a combination of proton and ligand mediated processes. Dissolution was found to be a significant feature at low pH (pH < 5) for a host of metal oxides, including Fe2O3.69 Here, colors are more intense for the HCl added reactions, which is consistent with our understanding, as adding more H+ would only increase the amount of dissolution. Iron released into solution during dissolution can be complexed with free ligand, resulting in the observed Fe-ligand complexes. In addition, strong binding of ligand to surface iron can result in the removal of the complex from the iron oxide surface as was previously observed for ligand exchange reactions on Fe3O4 with mimosine,22 a small catechol and carboxylic acid containing aromatic ligand, citric, and galactaric acids.17 These effects likely

36 ACS Paragon Plus Environment

Page 37 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

explain, in part, the anomalous binding behavior for 2,6-DHBA, which is both highly acidic and possesses two electron donating substituents ortho to the carboxylic acid. Both of these characteristics would increase the strength of the metal-ligand bond, and result in enhanced nanoparticle dissolution especially when HCl is added to the exchange reaction. Another mechanism whereby HCl can influence the ligand exchange reaction is through the addition of more Cl- to the reaction mixture. Chloride, if in large enough amounts, could compete with oncoming ligands for binding sites and decrease their adsorption, as was suggested for the electrolyte ions Na+ and Cl- by Vasudevan and Stone.65 This behavior would be similar to that previously observed for adsorbed sulfate, which inhibited the adsorption of p-hydroxybenzoic acid on noncrystalline iron oxide.52 Chloride ions are also thought to be incorporated during initial nanoparticle synthesis via side reactions with reaction byproducts, easily occupying free spots on nanoparticle surfaces due, in part, to their small footprint.72 Here, IONP-OA was synthesized using iron chloride precursors making the adsorption of Cl- at the surface a very likely scenario. It was observed in the binding of carboxylic acids to HfO2 nanoparticles that a more basic environment was required to encourage more ligand deprotonation in order for the ligand to outcompete and replace adsorbed Cl- on the nanoparticle surface.73 Therefore, the effect of adding HCl to the ligand exchange reaction mixture would be twofold, both inhibiting ligand deprotonation as well as supplying additional Cl- ions that could compete with ligand adsorption. Ligands that possess catechol groups would be largely affected as these groups are generally more difficult to deprotonate than carboxylic acids. This explanation suits the ligand binding behavior observed for 3,4-DHBA, which showed insufficient ligand exchange when HCl was added to the ligand exchange mixture, and the decreased aqueous solubility of 3,4-DHPA and 2,3-DHBA under HCl added conditions. The addition of Cl- has also been reported to speed up

37 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 53

nanoparticle dissolution in the presence of protons, due to the formation of Fe‒Cl surface complexes,74 and could contribute to the increased production of free iron-ligand complexes for the ligand exchange reactions of 2,6-, 2,5-, 2,3-DHBA and SA when HCl is added. Changes observed in solubility and in FTIR spectra of ligand exchanged nanoparticles are clearly connected to the addition of HCl to the reaction mixture. These changes may manifest either due to alterations in the ratio of ligands bound by a particular mode or to the total amount of ligands bound. For example, all ligands with di o-OH groups showed changes when HCl was added to the ligand exchange reaction mixture. This effect appears to be universal within this ligand group but to varying degrees with 3,4-DHPA and 2,3-DHBA less perturbed upon change in reaction conditions, in contrast to DA and 3,4-DHBA where the ligand exchange does not result in solution stable nanoparticles under HCl free and HCl added conditions respectively. These results suggest that another binding mode is being utilized by 3,4-DHPA and 2,3-DHBA under HCl added reaction conditions. Solution studies of Fe(III) with 3,4-DHPA do show evidence of carboxylate complexation in acidic solution, even though the binding of catechol to iron is typically a stronger interaction than for carboxylates. The possibility that carboxylate binding could be occurring in more acidic conditions may explain our solubility observations for IONP-3,4-DHPA. However, even if there is no difference in binding mode between the HCl added versus DI water added IONP ligand exchange products, it is possible that the increase in pH leads to more ligand exchange of OA for 3,4-DHPA. This would explain the increase in solubility of the DI water added IONP-3,4-DHPA as there is simply more 3,4-DHPA ligand on the surface of the nanoparticle, and correspondingly, more negative charge. Likewise, the FTIR spectra of 2,3-DHBA look qualitatively similar, therefore it is possible that simply greater ligand

38 ACS Paragon Plus Environment

Page 39 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

loading occurs when DI water is added resulting in greater stability of the particles in aqueous buffer.

SUMMARY AND CONCLUSIONS For a series of related benzoic acid and catechol derivatives we have determined relationships between chemical structure and ligand binding for iron oxide nanoparticles. The ligand exchange method was applied, relatively unchanged, to all of the ligands under study, and would serve as a good starting point for other catechol, salicylic acid, and benzoic acid derivatives. The method of ligand exchange and purification is easy, does not require sophisticated equipment or synthetic knowledge to perform, all of the solvents, reagents, and ligands are commercially available, and all reactions and manipulations are performed safely at room temperature. To our knowledge, solution stable magnetite iron oxide nanoparticles stabilized with the ligands 2,3-, 2,4-, 2,5-, and 3,5-DHBA have not been achieved until now. Many of the ligand exchanged nanoparticles were soluble under a range of solution conditions, and some were very soluble in aqueous media and could be directly used in materials applications that demand ultrasmall, monodisperse, aqueous stable magnetic iron oxide nanoparticles. Ligand exchanged nanoparticles that appear to present solution exposed carboxylic acid groups (IONP-3,4-DHPA, IONP-3,4-DHBA, and IONP-2,3DHBA) or amine groups (IONP-DA) may be good candidates for further chemical modification, such as EDC coupling, to attach molecules of interest. FTIR spectroscopy demonstrated great utility in verifying ligand exchange success, assisting in binding mode determination, and for gauging product purity. In situ FTIR spectroscopy could expand on this method to obtain realtime ligand exchange reaction data and determine exchange kinetics. In addition, Raman spectroscopy could also be utilized to confirm the presence and determine the nature of ligand

39 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 53

catechol binding to the iron oxide nanoparticle, as has been demonstrated in studies of catechol binding to iron in marine mussels75 and to TiO2 surfaces.76 Binding mode was mainly influenced by the relative position and nature of ring substituents and was surprisingly variable depending on whether or not HCl was added to the reaction medium, demonstrating that reaction conditions can promote one form of binding over another when multiple binding modes are available. Ligands with adjacent phenol groups and phenols ortho to the carboxylic acid have the potential to bind as a catecholate or salicylate, respectively, however the presence of these substituent geometries does not guarantee binding by the particular mode. In these cases, electronic effects and the influence of HCl addition to the reaction served to explain the apparent deviations in binding mode. The determined correlations between ligand structure and binding will ultimately aid in the rational design of ligands to achieve custom iron oxide nanoparticles.

ASSOCIATED CONTENT Supporting Information. The Supporting Information (SI) is available free of charge on the ACS Publications website at DOI: . Detailed synthesis and purification of ligand exchanged iron oxide nanoparticles. Details regarding electrokinetic measurements, iron oxide nanoparticle concentration determination, and nanoparticle UV‒visible (UV‒vis) spectroscopy. Preparation of Fe(II) and Fe(III) 2,6-DHBA complexes and UV‒vis spectroscopy of complexes and ligand exchange reaction supernatants. Nanoparticle TEM images, calculated diameters, TEM sample plating conditions, and modified nanoparticle purification parameters for some TEM samples and solubility determinations. FTIR

40 ACS Paragon Plus Environment

Page 41 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

band assignments for IONP-OA, ligand exchanged nanoparticles, investigated ligands and ligand sodium salts. Isolated FTIR spectra of all ligands, ligand functionalized nanoparticles, and prepared ligand sodium salts. UV‒vis solution spectra comparing adsorption profiles of selected nanoparticles. Modified purification parameters for some samples prepared for UV‒vis and electrokinetic measurements. FTIR spectra comparing bands relevant to ligand binding for 3,4DHPA, 3,4-DHBA, and 2,3-DHBA. FTIR spectra showing the purification process for IONP3,4-DHPA. FTIR spectra comparing IONP-SA, IONP-2,4-DHBA, and IONP-2,5-DHBA prepared with HCl or DI water. Solution spectra of 2,6-DHBA ligand exchange supernatant, Fe2,6-DHBA complexes, and spectral analysis of ligand exchange supernatants. Additional FTIR analysis of iron oxide nanoparticle core structure. AUTHOR INFORMATION Corresponding Author *Email: [email protected]. ACKNOWLEDGMENTS Financial support for this research was provided by the Center for Self-Assembled Chemical Structures (CSACS), the Canada Foundation for Innovation (CFI), and the National Sciences and Engineering Research Council of Canada (NSERC). Funding was provided to K. V. Korpany by the Richard T. Mohan Scholarship. We thank Omar Zahr, Joshua Lucate, and Julia Del Re for their assistance in TEM imaging. REFERENCES

41 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(1)

Page 42 of 53

Hudson, R.; Feng, Y.; Varma, R. S.; Moores, A. Bare magnetic nanoparticles: sustainable

synthesis and applications in catalytic organic transformations. Green Chem. 2014, 16 (10), 4493‒4505. (2)

Xu, P. A.; Zeng, G. M.; Huang, D. L.; Feng, C. L.; Hu, S.; Zhao, M. H.; Lai, C.; Wei, Z.;

Huang, C.; Xie, G. X.; Liu, Z. F. Use of iron oxide nanomaterials in wastewater treatment: A review. Sci. Total Environ. 2012, 424, 1‒10. (3)

Gupta, A. K.; Gupta, M. Synthesis and surface engineering of iron oxide nanoparticles

for biomedical applications. Biomaterials 2005, 26 (18), 3995‒4021. (4)

Ling, D.; Hyeon, T. Chemical design of biocompatible iron oxide nanoparticles for

medical applications. Small 2013, 9 (9‒10), 1450‒1466. (5)

Ghadban, A.; Ahmed, A. S.; Ping, Y.; Ramos, R.; Arfin, N.; Cantaert, B.; Ramanujan, R.

V.; Mieserez, A. Bioinspired pH and magnetic responsive catechol-functionalized chitosan hydrogels with tunable elastic properties. Chem. Commun. 2016, 52 (4), 697‒700. (6)

Li, Q.; Barrett, D. G.; Messersmith, P. B.; Holten-Andersen, N. Controlling Hydrogel

Mechanics via Bio-Inspired Polymer-Nanoparticle Bond Dynamics. ACS Nano 2016, 10 (1), 1317‒1324. (7)

Zvarec, O.; Purushotham, S.; Masic, A.; Ramanujan, R. V.; Miserez, A. Catechol-

Functionalized Chitosan/Iron Oxide Nanoparticle Composite Inspired by Mussel Thread Coating and Squid Beak Interfacial Chemistry. Langmuir 2013, 29 (34), 10899‒10906.

42 ACS Paragon Plus Environment

Page 43 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(8) Iron

Laurent, S.; Forge, D.; Port, M.; Roch, A.; Robic, C.; Elst, L. V.; Muller, R. N. Magnetic Oxide

Nanoparticles:

Synthesis,

Stabilization,

Vectorization,

Physicochemical

Characterizations, and Biological Applications. Chem. Rev. 2008, 108 (6), 2064‒2110. (9)

Wu, W.; He, Q.; Jiang, C. Magnetic Iron Oxide Nanoparticles: Synthesis and Surface

Functionalization Strategies. Nanoscale Res. Lett. 2008, 3 (11), 397‒415. (10) Wen, X.; Yang, J.; He, B.; Gu, Z. Preparation of monodisperse magnetite nanoparticles under mild conditions. Curr. Appl. Phys. 2008, 8 (5), 535‒541. (11) Korpany, K. V.; Habib, F.; Murugesu, M.; Blum, A. S. Stable water-soluble iron oxide nanoparticles using Tiron. Mater. Chem. Phys. 2013, 138 (1), 29‒37. (12) Korpany, K. V.; Mottillo, C.; Bachelder, J.; Cross, S. N.; Dong, P.; Trudel, S.; Friščić, T.; Blum, A. S. One-step ligand exchange and switching from hydrophobic to water-stable hydrophilic superparamagnetic iron oxide nanoparticles by mechanochemical milling. Chem. Commun. 2016, 52 (14), 3054‒3057. (13) Bloemen, M.; Brullot, W.; Luong, T. T.; Geukens, N.; Gils, A.; Verbiest, T. Improved functionalization of oleic acid-coated iron oxide nanoparticles for biomedical applications. J. Nanopart. Res. 2012, 14 (9), 1100. (14) Hofmann, A.; Thierbach, S.; Semisch, A.; Hartwig, A.; Taupitz, M.; Rühl, E.; Graf, C. Highly monodisperse water-dispersable iron oxide nanoparticles for biomedical applications. J. Mater. Chem. 2010, 20 (36), 7842‒7853. (15) Peacock, A. K.; Cauët, S. I.; Taylor, A.; Murray, P.; Williams, S. R.; Weaver, J. V. M.; Adams, D. J.; Rosseinsky, M. J. Poly[2-(methacryloyloxy)ethylphosphorylcholine]-coated iron 43 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 44 of 53

oxide nanoparticles: synthesis, colloidal stability and evaluation for stem cell labelling. Chem. Commun. 2012, 48 (75), 9373‒9375. (16) Yun, J.; Sonabend, A. M.; Ulasov, I. V.; Kim, D. H.; Rozhkova, E. A.; Novosad, V.; Dashnaw, S.; Brown, T.; Canoll, P.; Bruce, J. N.; Lesniak, M. S. A novel adenoviral vector labeled with superparamagnetic iron oxide nanoparticles for real-time tracking of viral delivery. J. Clin. Neurosci. 2012, 19 (6), 875‒880. (17) Lattuada, M.; Hatton, T. A. Functionalization of Monodisperse Magnetic Nanoparticles. Langmuir 2007, 23 (4), 2158‒2168. (18) Klokkenburg, M.; Hilhorst, J.; Erné, B. H. Surface analysis of magnetite nanoparticles in cyclohexane solutions of oleic acid and oleylamine. Vib. Spectrosc. 2007, 43 (1), 243‒248. (19) Pujari, S. P.; Scheres, L.; Marcelis, A. T. M.; Zuilhof, H. Covalent Surface Modification of Oxide Surfaces. Angew. Chem., Int. Ed. 2014, 53 (25), 6322‒6356. (20) Daou, T. J.; Grenèche, J. M.; Pourroy, G.; Buathong, S.; Derory, A.; Ulhaq-Bouillet, C.; Donnio, B.; Guillon, D.; Begin-Colin, S. Coupling Agent Effect on Magnetic Properties of Functionalized Magnetite-Based Nanoparticles. Chem. Mater. 2008, 20 (18), 5869‒5875. (21) Yuen, A. K. L.; Hutton, G. A.; Masters, A. F.; Maschmeyer, T. The interplay of catechol ligands with nanoparticulate iron oxides. Dalton Trans. 2012, 41 (9), 2545‒2559. (22) Amstad, E.; Gehring, A. U.; Fischer, H.; Nagaiyanallur, V. V.; Hähner, G.; Textor, M.; Reimhult, E. Influence of Electronegative Substituents on the Binding Affinity of CatecholDerived Anchors to Fe3O4 Nanoparticles. J. Phys. Chem. C 2011, 115 (3), 683‒691.

44 ACS Paragon Plus Environment

Page 45 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(23) Sarma, J.; Mahiuddin, S. Specific ion effect on the point of zero charge of α-alumina and on the adsorption of 3,4-dihydroxybenzoic acid onto α-alumina surface. Colloids Surf., A 2014, 457, 419‒424. (24) Tejedor-Tejedor, M. I.; Yost, E. C.; Anderson, M. A. Characterization of Benzoic and Phenolic Complexes at the Goethite Aqueous-Solution Interface Using Cylindrical InternalReflection Fourier-Transform Infrared-Spectroscopy. 2. Bonding Structures. Langmuir 1992, 8 (2), 525‒533. (25) Evanko, C. R.; Dzombak, D. A. Influence of Structural Features on Sorption of NOMAnalogue Organic Acids to Goethite. Environ. Sci. Technol. 1998, 32 (19), 2846‒2855. (26) Janković, I. A.; Šaponjić, Z. V.; Čomor, M. I.; Nedeljlković, J. M. Surface Modification of Colloidal TiO2 Nanoparticles with Bidentate Benzene Derivatives. J. Phys. Chem. C 2009, 113 (29), 12645‒12652. (27) Guan, X. H.; Shang, C.; Chen, G. H. ATR-FTIR investigation of the role of phenolic groups in the interaction of some NOM model compounds with aluminum hydroxide. Chemosphere 2006, 65 (11), 2074‒2081. (28) Hanna, K.; Quilès, F. Surface Complexation of 2,5-Dihydroxybenzoic Acid (Gentisic Acid) at the Nanosized Hematite-Water Interface: An ATR-FTIR Study and Modeling Approach. Langmuir 2011, 27 (6), 2492‒2500. (29) Saikia, N.; Sarma, J.; Borah, J. M.; Mahiuddin, S. Adsorption of 3,4-dihydroxybenzoic acid onto hematite surface in aqueous medium: Importance of position of phenolic -OH groups

45 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 46 of 53

and understanding of the same using catechol as an auxiliary model. J. Colloid Interface Sci. 2013, 398, 227‒233. (30) Borah, J. M.; Sarma, J.; Mahiuddin, S. Adsorption comparison at the α-alumina/water interface: 3,4-Dihydroxybenzoic acid vs. catechol. Colloids Surf., A 2011, 387 (1‒3), 50‒56. (31) Biber, M. V.; Stumm, W. An In-Situ ATR-FTIR Study: The Surface Coordination of Salicylic Acid on Aluminum and Iron(III) Oxides. Environ. Sci. Technol. 1994, 28 (5), 763‒768. (32) Nagesha, D. K.; Plouffe, B. D.; Phan, M.; Lewis, L. H.; Sridhar, S.; Murthy, S. K. Functionalization-induced improvement in magnetic properties of Fe3O4 nanoparticles for biomedical applications. J. Appl. Phys. 2009, 105 (7), 07B317. (33) Mondini, S.; Ferretti, A. M.; Puglisi, A.; Ponti, A. PEBBLES and PEBBLEJUGGLER: software for accurate, unbiased, and fast measurement and analysis of nanoparticle morphology from transmission electron microscopy (TEM) micrographs. Nanoscale 2012, 4 (17), 5356‒ 5372. (34) Shultz, M. D.; Reveles, J. U.; Khanna, S. N.; Carpenter, E. E. Reactive Nature of Dopamine as a Surface Functionalization Agent in Iron Oxide Nanoparticles. J. Am. Chem. Soc. 2007, 129 (9), 2482‒2487. (35) Zeleňák, V.; Vargová, Z.; Györyová, K. Correlation of infrared spectra of zinc(II) carboxylates with their structures. Spectrochim. Acta, Part A 2007, 66 (2), 262‒272. (36) Nara, M.; Torii, H.; Tasumi, M. Correlation between the Vibrational Frequencies of the Carboxylate Group and the Types of Its Coordination to a Metal Ion:  An ab Initio Molecular Orbital Study. J. Phys. Chem. 1996, 100 (51), 19812‒19817. 46 ACS Paragon Plus Environment

Page 47 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(37) Deacon, G. B.; Phillips, R. J. Relationships between the carbon-oxygen stretching frequencies of carboxylato complexes and the type of carboxylate coordination. Coord. Chem. Rev. 1980, 33 (3), 227‒250. (38) Bronstein, L. M.; Huang, X.; Retrum, J.; Schmucker, A.; Pink, M.; Stein, B. D.; Dragnea, B. Influence of Iron Oleate Complex Structure on Iron Oxide Nanoparticle Formation. Chem. Mater. 2007, 19 (15), 3624‒3632. (39) Lewandowski, W.; Kalinowska, M.; Lewandowska, H. The influence of metals on the electronic system of biologically important ligands. Spectroscopic study of benzoates, salicylates, nicotinates and isoorotates. Review. J. Inorg. Biochem. 2005, 99 (7), 1407‒1423. (40) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds: Part B: Applications in Coordination, Organometallic, and Bioinorganic Chemistry, 5th ed.; John Wiley & Sons: New York, NY, 1997; pp 59‒62. (41) Nakamoto, K.; McCarthy, P. J. Spectroscopy and structure of metal chelate compounds; John Wiley & Sons: New York, NY, 1968. (42) Martínez, D.; Motevalli, M.; Watkinson, M. Is there really a diagnostically useful relationship between the carbon-oxygen stretching frequencies in metal carboxylate complexes and their coordination mode? Dalton Trans. 2010, 39 (2), 446‒455. (43) Hu, L.; Percheron, A.; Chaumont, D.; Brachais, C. H. Microwave-assisted one-step hydrothermal synthesis of pure iron oxide nanoparticles: magnetite, maghemite and hematite. J. Sol-Gel Sci. Technol. 2011, 60 (2), 198‒205.

47 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 48 of 53

(44) Iwasaki, T.; Sato, N.; Nakamura, H.; Watano, S. Mechanochemical formation of superparamagnetic magnetite nanoparticles from ferrihydrite over a wide range of pH environments. Mater. Chem. Phys. 2013, 140 (2‒3), 596‒601. (45) Cai, W.; Wan, J. Facile synthesis of superparamagnetic magnetite nanoparticles in liquid polyols. J. Colloid Interface Sci. 2007, 305 (2), 366‒370. (46) Sundrarajan, M.; Ramalakshmi, M. Novel Cubic Magnetite Nanoparticle Synthesis Using Room Temperature Ionic Liquid. E-Journal of Chemistry 2012, 9 (3), 1070‒1076. (47) Larkin, P. Infrared and Raman Spectroscopy: Principles and Spectral Interpretation; Elsevier: Waltham, MA, 2011. (48) Lambert, J. B.; Shurvell, H. F.; Lightner, D. A.; Cooks, R. G. Introduction to Organic Spectroscopy; Macmillan: New York, 1987; pp 133‒245. (49) Mahdavi, M.; Ahmad, M. B.; Haron, M. J.; Namvar, F.; Nadi, B.; Rahman, M. Z.; Amin, J. Synthesis, Surface Modification and Characterisation of Biocompatible Magnetic Iron Oxide Nanoparticles for Biomedical Applications. Molecules 2013, 18 (7), 7533‒7548. (50) Roonasi, P.; Yang, X.; Holmgren, A. Competition between sodium oleate and sodium silicate for a silicate/oleate modified magnetite surface studied by in situ ATR-FTIR spectroscopy. J. Colloid Interface Sci. 2010, 343 (2), 546‒552. (51) Kalinowska, M.; Świsłocka, R.; Borawska, M.; Piekut, J.; Lewandowski, W. Spectroscopic (FT-IR, FT-Raman, UV) and microbiological studies of di-substituted benzoates of alkali metals. Spectrochim. Acta, Part A 2008, 70 (1), 126‒135.

48 ACS Paragon Plus Environment

Page 49 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(52) Kung, K. H.; McBride, M. B. Coordination-Complexes of Para-Hydroxybenzoate on Fe Oxides. Clays Clay Miner. 1989, 37 (4), 333‒340. (53) Corona-Avendaño, S.; Alarcón-Angeles, G.; Rosquete-Pina, G. A.; Rojas-Hernández, A.; Gutierrez, A.; Ramírez-Silva, M. T.; Romero-Romo, M.; Palomar-Pardavé, M. New Insights on the Nature of the Chemical Species Involved During the Process of Dopamine Deprotonation in Aqueous Solution: Theoretical and experimental study. J. Phys. Chem. B 2007, 111 (7), 1640‒ 1647. (54) Hao, R.; Yu, J.; Ge, Z.; Zhao, L.; Sheng, F.; Xu, L.; Li, G.; Hou, Y. Developing Fe3O4 nanoparticles into an efficient multimodality imaging and therapeutic probe. Nanoscale 2013, 5 (23), 11954‒11963. (55) Suzuki, H.; Nagasaka, M.; Sugiura, M.; Noguchi, T. Fourier Transform Infrared Spectrum of the Secondary Quinone Electron Acceptor QB in Photosystem II. Biochemistry 2005, 44 (34), 11323‒11328. (56) Bauscher, M.; Nabedryk, E.; Bagley, K.; Breton, J.; Mäntele, W. Investigation of models for photosynthetic electron acceptors: Infrared spectroelectrochemistry of ubiquinone and its anions. FEBS Lett. 1990, 261 (1), 191‒195. (57) Volovšek, V.; Colombo, L.; Furić, K. Vibrational Spectrum and Normal Coordinate Calculations of the Salicylic Acid Molecule. J. Raman Spectrosc. 1983, 14 (5), 347‒352. (58) Porwal, S. K.; Furia, E.; Harris, M. E.; Viswanathan, R.; Devireddy, L. Synthetic, potentiometric and spectroscopic studies of chelation between Fe(III) and 2,5-DHBA supports salicylate-mode of siderophore binding interactions. J. Inorg. Biochem. 2015, 145, 1‒10.

49 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 50 of 53

(59) Regulska, E.; Kalinowska, M.; Wojtulewski, S.; Korczak, A.; Sienkiewicz-Gromiuk, J.; Rzączyńska, Z.; Świsłocka, R.; Lewandowski, W. Theoretical (in B3LYP/6-3111++G** level), spectroscopic (FT-IR, FT-Raman) and thermogravimetric studies of gentisic acid and sodium, copper(II) and cadmium(II) gentisates. Spectrochim. Acta, Part A 2014, 132, 713‒725. (60) Dunn, G. E.; McDonald, R. S. Infrared spectra of aqueous sodium benzoates and salicylates in the carboxyl-stretching region: chelation in aqueous sodium salicylates. Can. J. Chem. 1969, 47 (24), 4577‒4588. (61) Yost, E. C.; Tejedor-Tejedor, M. I.; Anderson, M. A. In Situ CIR-FTIR Characterization of Salicylate Complexes at the Goethite/Aqueous Solution Interface. Enviro. Sci. Technol. 1990, 24 (6), 822‒828. (62) Lewandowski, W.; Dasiewicz, B.; Koczoń, P.; Skierski, J.; Dobrosz-Teperek, K.; Świsłocka, R.; Fuks, L.; Priebe, W.; Mazurek, A. P. Vibrational study of alkaline metal nicotinates, benzoates and salicylates. J. Mol. Struct. 2002, 604 (2–3), 189‒193. (63) Lindberg, B. J. IR-Spectra of Substituted Sodium Benzoate in Aqueous and Deuterium Oxide Solution. Acta Chem. Scand. 1968, 22, 571‒580. (64) Davis, K.; Qi, B.; Witmer, M.; Kitchens, C. L.; Powell, B. A.; Mefford, O. T. Quantitative Measurement of Ligand Exchange on Iron Oxides via Radiolabeled Oleic Acid. Langmuir 2014, 30 (36), 10918‒10925. (65) Vasudevan, D.; Stone, A. T. Adsorption of Catechols, 2-Aminophenols, and 1,2Phenylenediamines at the Metal (Hydr)Oxide/Water Interface: Effect of Ring Substituents on the Adsorption onto TiO2. Enviro. Sci. Technol. 1996, 30 (5), 1604‒1613.

50 ACS Paragon Plus Environment

Page 51 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(66) Chanley, J. D.; Feageson, E. Fe(III) Complexes of o-Hydroxy Aromatic Acids. J. Am. Chem. Soc. 1956, 78 (10), 2237‒2241. (67) Avdeef, A.; Sofen, S. R.; Bregante, T. L.; Raymond, K. N. Coordination Chemistry of Microbial Iron Transport Compounds. 9. Stability Constants for Catechol Models of Enterobactin. J. Am. Chem. Soc. 1978, 100 (17), 5362‒5370. (68) Gu, B.; Schmitt, J.; Chen, Z.; Liang, L.; McCarthy, J. F. Adsorption and desorption of different organic matter fractions on iron oxide. Geochim. Cosmochim. Acta 1995, 59 (2), 219‒ 229. (69) Gulley-Stahl, H.; HoganII, P. A.; Schmidt, W. L.; Wall, S. J.; Buhrlage, A.; Bullen, H. A. Surface Complexation of Catechol to Metal Oxides: An ATR-FTIR, Adsorption, and Dissolution Study. Enviro. Sci. Technol. 2010, 44 (11), 4116‒4121. (70) McBride, M. B.; Kung, K. H. Adsorption of phenol and substituted phenols by iron oxides. Environ. Toxicol. Chem. 1991, 10 (4), 441‒448. (71) Vestal, C. R.; Zhang, Z. J. Effects of Surface Coordination Chemistry on the Magnetic Properties of MnFe2O4 Spinel Ferrite Nanoparticles. J. Am. Chem. Soc. 2003, 125 (32), 9828‒ 9833. (72) Boles, M. A.; Ling, D.; Hyeon, T.; Talapin, D. V. The surface science of nanocrystals. Nat. Mater. 2016, 15 (2), 141‒153. (73) De Roo, J.; Van den Broeck, F.; De Keukeleere, K.; Martins, J. C.; Van Driessche, I.; Hens, Z. Unravelling the Surface Chemistry of Metal Oxide Nanocrystals, the Role of Acids and Bases. J. Am. Chem. Soc. 2014, 136 (27), 9650‒9657. 51 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 52 of 53

(74) Sidhu, P. S.; Gilkes, R. J.; Cornell, R. M.; Posner, A. M. Dissolution of iron oxides and oxyhydroxides in hydrochloric and perchloric acids. Clays Clay Miner. 1981, 29 (4), 269‒276. (75) Harrington, M. J.; Masic, A.; Holten-Andersen, N.; Waite, J. H.; Fratzl, P. Iron-Clad Fibers: A Metal-Based Biological Strategy for Hard Flexible Coatings. Science 2010, 328 (5975), 216‒220. (76) Hwang, D. S.; Harrington, M. J.; Lu, Q.; Masic, A.; Zeng, H.; Waite, J. H. Mussel foot protein-1 (mcfp-1) interaction with titania surfaces. J. Mater. Chem. 2012, 22 (31), 15530‒ 15533.

52 ACS Paragon Plus Environment

Page 53 of 53

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Insert Table of Contents Graphic and Synopsis Here

53 ACS Paragon Plus Environment