Quantitative Analysis of the Functionalization of Gallium Phosphide with Organic Azides David Richards Purdue University

Purdue University Purdue e-Pubs

Birck and NCN Publications Birck Nanotechnology Center

12-20-2011 Quantitative Analysis of the Functionalization of Gallium Phosphide With Organic Azides David Richards Purdue University

Philip Luce Purdue University

Dmitry Zemlyanov Birck Nanotechnology Center, Purdue University, [email protected]

Albena Ivanisevic North Carolina State University

Follow this and additional works at: http://docs.lib.purdue.edu/nanopub Part of the Nanoscience and Nanotechnology Commons

Richards, David; Luce, Philip; Zemlyanov, Dmitry; and Ivanisevic, Albena, "Quantitative Analysis of the Functionalization of Gallium Phosphide With Organic Azides" (2011). Birck and NCN Publications. Paper 1131. http://dx.doi.org/10.1002/sca.21012

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for additional information. SCANNING VOL. 34, 332–340 (2012) C Wiley Periodicals, Inc.

Quantitative Analysis of the Functionalization of Gallium Phosphide With Organic Azides

DAV I D RICHARDS1,PHILIP LUCE1,DMITRY ZEMLYANOV2, AND ALBENA IVANISEVIC3 1 Weldon School of Biomedical Engineering, Purdue University, West Lafayette, Indiana 2 Birck Nanotechnology Center, Purdue University, West Lafayette, Indiana 3 Department of Materials Science and Engineering, North Carolina State University, Joint Department of Biomedical Engineering NCSU/UNC-CH, Raleigh, North Carolina

Summary: Gallium phosphide (GaP) surfaces were coupled plasma mass spectrometry was used to eval- functionalized with two different molecules that uate the gallium concentration in the stability solu- contain an azide moiety at their terminus. Com- tions. While the functionalization with the organic pound 4-azidophenacyl bromide (4AB) is an aryl azides did not provide complete suppression of gal- azide with a bromine group at its opposite termi- lium leaching, both of the azides decreased the leach- nus that provides easy identification of the molecule’s ing by 10–50%. SCANNING 34: 332–340, 2012. C presence on the surface with x-ray photoelec- 2012 Wiley Periodicals, Inc. tron spectroscopy (XPS). O-(2-aminoethyl)-O-(2- azidoethyl)pentaethylene glycol (AAP) is a small poly(ethylene glycol) molecule with an amine group Key words: gallium phosphide, azide, at its opposite terminus. Atomic force microscopy functionalization, XPS, toxicity was used to identify the uniformity of the clean and functionalized GaP surfaces. Water contact angle revealed a more hydrophobic surface with AAP functionalization (33◦) and even more hydrophobic (53◦) Introduction with the 4AB functionalized surface compared to a clean surface (16◦). XPS confirmed the presence A considerable amount of effort in the past decades of each of the organic azides on the surface. XPS has been devoted to developing implantable devices was further used to calculate the adlayer thickness for specific purposes, especially biosensing. There of each functionalization. This analysis revealed an are many considerations for the advancement of im- adlayer thickness of about 8 A˚ for the 4AB func- plantable devices, including the optimization of sensi- tionalized surfaces compared to 1 A˚ for the AAP tivity and selectivity; however, the most important is adlayer, which led to the conclusion that AAP func- most likely biocompatibility. Much research has been tionalization only provided partial coverage. A stabil- dedicated to tailoring inorganic surfaces with organic ity study using 4AB-functionalized surfaces showed molecules so as to minimize toxicity. good stability in saline solutions with varying con- One of the most common classes of materials used centrations of hydrogen peroxide. Finally, inductively in biosensing is semiconductors. Indeed, silicon has been experimented with extensively for biosensors Contract grant sponsor: NSF; Contract grant number: CHE- due to its broad physical and electrical characteri- 1052809. zation (Cui et al., 2001; Kim et al., 2007). In addition, Address for reprints: Albena Ivanisevic, Department of Materials III-V semiconductors have also been used in sensing Science and Engineering, North Carolina State University, Joint De- applications. Gallium arsenide is the most commonly partment of Biomedical Engineering NCSU/UNC-CH, 911 Partner’s used III-V semiconductor, again owing to its mature Way, Raleigh, NC 27695 development (Lee et al., 2008). Gallium phosphide E-mail: [email protected] (GaP) is a III-V semiconductor that is commonly used Received 31 October 2011; Accepted with revision 20 December in optoelectronic and high-temperature devices. It is 2011 an excellent candidate for future biosensor devices because its unmodified surface has shown favorable DOI 10.1002/sca.21012 et al. Published online 9 April 2012 in Wiley Online Library (wiley biocompatibility in previous studies (Hallstrom , onlinelibrary.com) 2007). D. Ricahards et al.: Azide functionalization on GaP 333 Like many semiconductors, GaP is susceptible to molecule. AAP was chosen because of the antibio- corrosion and degradation. This process can lead to fouling capabilities poly(ethylene glycol) (PEG) like the leaching of toxic surface ions into the surround- molecules can confer on a surface. ing solution. Obviously, this poses a problem when Photochemistry on a GaP surface is not uncom- it comes to materials such as GaP being used in im- mon. In our past study, GaP was reacted with a termi- plantable devices. Therefore, our group has used or- nal alkene under UV light resulting in the formation ganic adsorbates to covalently bind to the surface of of a dense adlayer (Ivanisevic et al., 2008). To our GaP to reduce the amount of the toxic and leach- knowledge, functionalization with azides on a GaP able oxide layer (Flores-Perez et al., 2008; Richards surface has not been studied before. et al., 2010). In this article, we propose a new method Besides XPS, other surface sensitive techniques in- to covalently functionalize a GaP surface using azide cluding atomic force microscopy (AFM) and water chemistry. Azides, which are commonly used in click contact angle were used to analyze the functionalized chemistry, have the ability to form a highly reac- surfaces. A stability study was performed using water, tive nitrene intermediate when irradiated with ultra- various pH solutions, and saline with varying concen- violet (UV) light (Ziani-Cherif et al., ’99). Azides trations of hydrogen peroxide (H2O2). Finally induc- have been used to link various organic molecules to tively coupled plasma mass spectrometry (ICP-MS) polystyrene (Nahar et al., 2001), poly(ethylene tereph- was used to quantify the amount of gallium leach- thalate) (Ziani-Cherif, Imachi et al.,’99),glasscover ing from the functionalized surfaces in the stability slips (Chen et al., 2009), quartz (Harmer, ’91), and solutions. silicon (Harmer, ’91). Due to its successful functionalization of a broad range of materials, it is proposed that azides can Materials and Methods provide a new stable route to covalently link organic molecules and biomolecules to a GaP surface. Surface Cleaning and Functionalization We used 4-azidophenacyl bromide (4AB) and O- (2-aminoethyl)-O-(2-azidoethyl)pentaethylene gly- GaP (100) wafers were purchased from Univer- col (AAP) as the azide reactants (see Fig. 1). Specif- sity Wafer (South Boston, MA). The (100) plane ically, 4AB was used to allow for easy identification of GaP was chosen because of its ability to ac- of its presence on a GaP surface. Its bromine group commodate a wide range of adlayers including UV- generates a distinct peak in an x-ray photoelectron reactive adlayers (Ivanisevic et al., 2008; Richards spectroscopy (XPS) spectrum that provides evidence et al., 2010). Compounds 4AB and AAP were pur- that the molecule is indeed attached to the GaP sur- chased from Sigma-Aldrich (St. Louis, MO). The face. In addition, the bromine group is not capable GaP wafers were cut into 5 mm × 5 mm pieces. Be- of bonding to both the GaP surface and the azide fore functionalization, all wafers were degreased. The

Fig 1. The functionalization schemes of the two different azide adsorbates. 334 SCANNING VOL. 34, 5 (2012) degreasing process consisted of ultrasonication in wa- chased from Asylum Research (Santa Barbara, CA). ter for 20 min and ethanol for 20 min followed by dry- Data analysis was performed using the Nanoscope ing with N2. The wafers were then exposed to NH4OH III 5.12r3 software (Veeco, Plainview, NY). for 30 s in order to remove the oxide layer (Morota XPS data were obtained with a Kratos Ultra DLD and Adachi, 2006) The wafers were then rinsed with spectrometer (Shimadzu, Nakagyo-ku, Kyoto, Japan) water and ethanol and dried with N2. Finally, the using monochromatic Al Kα radiation (hν = 1486.6 wafers were exposed to either a 1-mM solution of eV). Survey and high-resolution spectra were col- 4AB in toluene or a 1% solution of AAP in ethanol. lected at a fixed analyzer pass energy of 160 and 20 eV, The reaction vials were exposed to UV light (302 nm) respectively, and acquisition was performed at pho- for 3 h. After the functionalization was complete, the toemission angles of 0◦,45◦, and 60◦ measured with wafers that were exposed to 4AB were rinsed with respect to the surface normal. Binding energy values copious amounts of toluene, water, and ethanol and were referenced to the Fermi edge, and charge cor- dried with N2 after each rinse. The wafers exposed to rection was done using the Carbon 1s peak set at AAP were rinsed with copious water and ethanol with 284.80 eV. Curve fitting was performed after linear subsequent N2 drying after each rinse. All wafers were or Shirley-type background subtraction assuming a stored in a vacuum desiccator for an average of four Gaussian/Lorentzian peak shape. days until subsequent analysis unless otherwise noted. ICP-MS data were obtained with an Element2 ICP mass spectrometer (ThermoFisher, Bremen, Ger- many) equipped with an Aridus desolvating introduc- Stability Experiments tion system (with a T1Hnebulizer) to enhance sen- sitivity and reduce oxide and hydride interferences One milliliter solutions of water; saline; saline with (Cetac Technologies, Omaha, NE). The argon sweep 0.1%, 1%, and 10% H O ; and phosphate-buffered 2 2 gas and nitrogen of the Aridus were adjusted for saline (PBS) solutions of pH 5, 7, and 9 were created maximum peak height and stability using 7Li, 115In, and placed in 1-mL plastic vials. One GaP sample and 238U peaks obtained from a Merck multielement functionalized with 4AB was placed in each vial. Af- standard (1 ng/mL, Merck & Co., Whitehouse sta- ter one and five days of incubation in their respective tion, NJ). After tuning and calibration of the ICP solutions, the wafers were removed and rinsed with instrument, medium resolution analyses were carried ethanol and dried with N . Water contact angle was 2 out on the Ga69 and Ga71 peaks. performed on each sample. The reported contact angles are the average of at least three measurements taken in multiple locations on the sample. Each GaP Results and Discussion wafer was rinsed with ethanol, dried with N2, and then placed back in the respective solution. After Before the GaP wafers were exposed to the azide seven days, all of the GaP wafers were removed from reactants, the cleaned GaP wafers were subject to wa- their solution. Water contact angle was performed for ter contact angle in order to identify that the wafers the last time and the wafers were not placed back in were indeed clean and ready for functionalization. A solution. The solutions of water, saline, saline with clean GaP surface revealed a water contact angle of ◦ 1% H2O2, and PBS buffer solutions of pH 5, 7, and about 16 , which is consistent with previous reports 9 were analyzed to determine how much gallium ion (Flores-Perez et al., 2008; Mukherjee et al., 2010). Af- was present using ICP-MS. AAP-functionalized sam- ter functionalization with AAP and 4AB, the contact ples were exposed to water, saline, saline with 1% angle increased to about 33 ± 5◦ and 53 ± 5◦, respec- H2O2, and PBS buffer solutions of pH 5, 7, and 9. tively (Table I). The contact angle for AAP correlates These solutions were also analyzed with ICP-MS to with the literature related to PEG on various surfaces determine the amount of gallium ion present. (Schlapak et al., 2006; Alibeik et al., 2010). This is a fair assessment since AAP is essentially a low molecular weight PEG molecule. The contact angle for GaP Surface Characterization functionalized with 4AB is higher due to the mostly hydrophobic structure of the 4AB molecule. Water contact angle measurements were performed on a Tantac, Inc. (Acquired by Chemsultants International, Inc., Mentor, OH) contact angle meter, TABLE I Contact angle of a clean GaP sample and a GaP model CAM-PLUS, using the half-angle method. sample after each azide functionalization AFM images were collected using the tapping Functionalization Contact angle (◦) mode setting on a Multi-Mode Nanoscope IIIa atomic force microscope (Veeco, Plainview, NY). A Clean GaP 16 ± 6 ± scan size of 1 μm × 1 μm at a rate of 1 Hz was used GaP with AAP 33 5 GaP with 4AB 53 ± 5 on each sample. Single-beam cantilevers were pur- D. Ricahards et al.: Azide functionalization on GaP 335 tionalizations result in slight nonuniformity. Indeed, the roughness most likely plays a role in the relatively large error in the contact angle measurements. The increased RMS roughness values compared to the clean GaP RMS roughness value are consistent with the chemisorption of up to a monolayer of the respective reactant.

XPS Characterization

Further analysis of the clean and functionalized GaP surfaces was carried out with XPS. XPS is a powerful surface sensitive tool that allows qualitative as well as quantitative characterization of the surface. The 4AB molecule was chosen speciﬁcally because of the presence of its bromine group. Halogens, such as bromine, are easily detectable using XPS, therefore the presence of the bromine signature in the XPS spectra would provide a quick and easy identiﬁca- tion of the successful functionalization of 4AB (Basu et al., 2006). Figure 3 depicts the Br 3d spectra of the clean GaP surface along with a 4AB functionalized GaP surface. An obvious peak at around 69 eV indeed indicates the presence of bromine and the 4AB adsorbate. The clean GaP Br 3d spectrum is lacking the signature peak. More important information can be garnered from the XPS carbon 1s spectra (Fig. 4). First, the freshly cleaned GaP spectrum shows the typical signatures of a clean, etched surface. The peak at 284.8 eV corresponds to C-H bonding. We can conclude indirectly that the relatively large size of this peak compared to the functionalized samples is most likely the result of atmospheric contamination from the lab air. The clean GaP sample was exposed to lab air for several minutes while the sample was being loaded into the XPS chamber. This is likely a result of the

Fig 2. AFM surface topography images after AAP- functionalization (A), clean GaP (B), and after 4AB- functionalization (C). (Scale bar corresponds to 250 nm).

In order to analyze the surface topography of the clean GaP sample and the functionalized samples, we used AFM. A surface topography image allows us to identify a uniformly functionalized surface that lacks any large pits or mounds of debris or excess reactant and solvent. A clean, etched GaP wafer, as seen in Figure 2(B), was shown to be very smooth, with a root mean square (RMS) roughness of 0.21 nm. After functionalization with AAP and 4AB, the roughness increases slightly to 0.48 nm and 0.37 nm, respectively. The corresponding topography images can be seen in the Figure 2(A) and (C). We can conclude Fig 3. XPS Bromine 3d spectra of a clean GaP surface and from these images and measurements that both func- two 4AB-functionalized GaP surfaces. 336 SCANNING VOL. 34, 5 (2012)

bound to the gallium atoms on the surface. This is not easily distinguishable using the XPS spectra, but other studies have reported that organic thiolated molecules preferentially bond to the gallium atoms on a GaP surface compared to the phosphorus atoms forming Ga-S bonds (Zerulla and Chasse, 2009). There- fore, based on this previous study, it is hypothesized that the azide molecules predominantly develop Ga- N bonds at the surface (see Fig. 1). XPS was further used to determine the thickness of the adlayer on the GaP surface using the following equation adapted from Fadley, which assumes complete and homogeneous coverage (Fadley, ’78): −t N = N0 − , Ga2p(P2p) Ga p(P p) 1 (θ) (1) 2 2 e Ga2p(P2p) cos Fig 4. XPS carbon 1s spectra of a clean GaP surface and two 0 of each azide functionalizations. where NGa2p(P2p) and N Ga2p(P2p) are peak intensities from the Ga 2p (P 2p) spectra before and after functionalization, respectively. Ga2p(P2p) is the electron electrostatically charged oxide-free GaP surface that attenuation length of Ga 2p or P 2p photoelectrons attracts oppositely charged contamination (Rein- through the AAP and 4AB adlayers, θ is the photoe- hardt et al., 2008). Moreover, the greater exposure mission angle, and t is the thickness of the adlayer of the functionalized samples to the solvents (ethanol in A.˚ The electron attenuation length was estimated and toluene) and UV light likely removed some of using an equation developed by Cumpson (Cump- the atmospheric contaminants. Loading the sam- son, 2001). The average thickness, t, was calculated ples into the XPS chamber also resulted in lab air by plotting ln(N/N0) versus 1/cos(θ). The slope of exposure. The other carbon components at higher the resulting line corresponded to the value for –t. binding energies are negligible since their area repre- The thickness of the AAP adlayer on the GaP sents an order of magnitude less than the C-H peak surface was determined to be 2 A˚ using the Ga 2p area. high-resolution spectrum and 1 A˚ using the P 2p Figure 4 also depicts two separate samples of each high-resolution spectrum (Table II). Since the thick- type of functionalization. The GaP functionalized ness of the adlayer must be at least the thickness of with 4AB revealed components that are to be ex- an AAP molecule lying on its side (∼2 A),˚ we can pected. The peak at 284.8 eV again corresponds to say that the complete coverage assumption does not the C-H bonding from the 4AB molecule. Indeed, this hold. Therefore, the AAP functionalization results in peak may also represent carbon contamination from partial coverage of the GaP surface. In contrast, the the atmosphere. Shifted to a higher binding energy of thickness of the 4AB adlayer was calculated to be 8 about 286.5 eV lies a shoulder peak that is a result A˚ with the Ga 2p high-resolution spectrum and 11 of the C-N bonding of the 4AB molecule covalently A˚ with the P 2p high-resolution spectrum. The re- bonded to the surface. At an even higher binding en- sults of these calculations demonstrate that the 4AB ergy of 289 eV, we observe the peak corresponding adsorbates are standing in a more perpendicular fash- to C = O, which is relatively similar in area as the ion with respect to the surface than the AAP adsor- C-N peak. This finding correlates well with the 4AB bates, especially considering the fact that the length of structure since there is exactly one of each type of the 4AB molecule is approximately 9 A,˚ whereas the bond. AAP molecule is approximately 27 A.˚ There are a va- The spectra corresponding to the AAP- riety of explanations for these findings, each of which functionalized GaP again showed peaks that may contribute partly. First, it is possible that the were expected. The most noticeable aspect of the spectra is the relatively higher shoulder peaks cor- TABLE II The adlayer thickness of each azide functionaliza- responding to C-O. The AAP is essentially a short tion using the gallium 2p and phosphorus 2p high-resolution PEG molecule and therefore has a large amount of spectra C-O bonds compared to the other types of carbon bonds. Calculated thickness (A)˚ In the case of both functionalization schemes, the Functionalization Ga 2p P 2p nitrene intermediate formed by UV irradiation is re- ± ± active enough to covalently bond to the atoms on the GaP with AAP 2 112 GaP with 4AB 8 ± 311± 2 GaP surface. Specifically, the azides are most likely D. Ricahards et al.: Azide functionalization on GaP 337 positively polarized Ga atoms at the surface react with approaching the contact angle of clean GaP (16◦). the oxygen atoms or the amine terminus of the AAP This trend was observed in our previous study and molecule, which would prevent the AAP from order- analysis by XPS after exposure to pH solutions indi- ing into a well-packed monolayer. Also, aryl azides cated that there was indeed a depletion of the organic are inherently more stable compared to alkyl azides adlayer (data not shown). In the pH 7 and pH 9 solu- and acyl azides (Abramovitch and Davis, ’64). tions, the adlayers are not only completely degraded, Another explanation for the higher thickness of but the GaP surface becomes more hydrophilic. This the 4AB adlayer is related to the UV absorption. In can most likely be attributed to the chemicals in the general, one can assume that a halogenated ring struc- pH 7 and pH 9 PBS buffers, specifically hydrogen 2− ture will have a higher extinction coefficient compared phosphate (HPO4 ) that react with the GaP surface to a linear molecule (Kumar et al., 2006). This phe- atoms and form a hydrophilic oxide layer. Total wet- nomenon allows more efficient activation of the azide ting is not observed in the pH 5 PBS buffer because 2− on the 4AB molecule and thus greater chemisorption there is a lower concentration of HPO4 and a higher − to the GaP surface. concentration of less reactive H2PO4 . Finally, at low densities, PEG molecules organize Exposing the functionalized samples to physiolog- into a “mushroom-like” configuration rather than a ical conditions is important in testing for future de- more perpendicular “brush-like” configuration that velopment of implantable devices. Saline with vary- would result in the low adlayer thickness that is ob- ing concentrations of H2O2 were used to simulate served (Degennes, ’80). While the “mushroom” radius macrophage recruitment at implant sites (Linsmeier of AAP is greater than 1 A,˚ this adlayer thickness et al., 2008). The samples exposed to saline with vary- value is an average of the entire spot size during XPS ing concentrations of H2O2 showed a different trend analysis, which likely contained portions of bare GaP. compared to the water and pH solutions. There is a leveling off of the contact angle after just one day and continues at that level throughout the study. The contact angle values for the samples in plain saline seem Water Contact Angle Stability Study to be a little higher than the values garnered from the water and pH 5 solutions suggesting that the ad- Since the 4AB adlayer has the potential for fu- layer is more resistant to degradation. Remarkably, ture functionalization by replacing the bromine with the addition of H O has little effect on the contact a reactive carboxylic acid, one is concerned with its 2 2 angles, which means that the H O is not reacting with stability in various solutions. In order to demonstrate 2 2 the adlayer. Instead, the H O is most likely reacting the stability of the adlayer on the GaP surface, one 2 2 with the bare GaP on the surface, which is confirmed 4AB-functionalized surface was exposed to each of by the increased gallium leaching from these samples the following solutions in 1 mL vials: water; PBS at as discussed in the following section. Similar trends pH 5, 7, and 9; saline; and saline with 0.1, 1, and were observed in our previous report, in which GaP 10% H O . After, one, five, and seven days, the 4AB- 2 2 surfaces were functionalized with thiols and terminal functionalized samples were removed and subject to alkene molecules (Richards, Zemlyanov et al., 2010). contact angle. A summary of the results can be found in Table III. With respect to the water solution, there is a decrease after one day to a contact angle of around 25◦ that remains stable until the seventh day. A sim- Toxicity Analysis ilar trend is seen when the functionalized surface is exposed to a pH 5 solution. Samples exposed to the Toxicity of the 4AB- and AAP-functionalized sur- pH 7 and pH 9 solutions showed total wetting after faces was investigated by determining the amount of just one day in solution. These trends indicate that gallium leaching into various solutions. One sample the adlayer is degraded to a considerable extent in the of each functionalization was placed in each of the water and pH 5 solutions since the contact angles are following solutions in 1 mL vials: water; saline; saline

TABLE III The contact angle of 4AB-functionalized GaP surfaces after one, ﬁve, and seven days incubation in water and various ◦ pH solutions and saline with varying concentrations of H2O2. (All measurements have an error of ±5 )

Water contact angle (◦)

Water pH 5 pH 7 pH 9 Saline Saline/0.1% H2O2 Saline/1% H2O2 Saline/10% H2O2 Begin535353535353 53 53 One day 29 28 0 0 40 40 52 43 Five days 21 26 0 0 29 30 50 35 Seven days 23 24 0 0 31 36 53 37 338 SCANNING VOL. 34, 5 (2012)

TABLE IV Gallium concentration in water solutions exposed The highest concentration (330 ppb) again occurred to clean and 4AB- and AAP-functionalized GaP surfaces when the solution was exposed to a nonfunctional- Sample Ga concentration (ppb) ized surface. Functionalized surfaces in saline showed significantly lower concentrations of 80 ppb and 140 Clean GaP in water 80 ± 5 ppb for 4AB and AAP, respectively. This is a large de- ± AAP on GaP in water 60 5 crease in the amount of gallium leaching suggesting 4AB on GaP in water 70 ± 5 that the adlayers are protecting the underlying surface from the harsh conditions of the buffer ions without being heavily degraded by the acidity of the in-house with 1% H2O2; PBS buffers of pH 5, 7, and 9. After water. In this case, the AAP-functionalized surface seven days in solution, the samples were removed and leached more, indicating that the AAP molecule is the remaining solution was analyzed for the amount more susceptible to damage from the buffer condi- of gallium within it. Table IV summarizes the gallium tions. Samples that were exposed to a 1% solution of concentration within the water samples for a 4AB- H2O2 in saline had higher concentrations of gallium, functionalized surface, an AAP-functionalized sur- however, they were still lower than the pure saline face, and a clean GaP wafer as a control. There was solution. As previously discussed, it is obvious that approximately 80 ppb of gallium in the water that was H2O2 has a significant effect on the amount of gal- exposed to a clean GaP wafer compared to 70 ppb and lium leaching from the surface and it is irrespective of 60 ppb for the 4AB- and AAP-functionalized surface, the functionalization. respectively. The slight reduction can be explained by Notably, the amount of gallium leaching from the the azide molecules covalently bonding to the sur- 4AB surfaces in the saline/H2O2 solutions is consid- face atoms that prevent the gallium from forming an erably higher than that of the water solution. This unstable oxide that can dissolve into the surround- finding seems somewhat contrary to the results ob- ing solution. The lesser amount of leaching demon- tained in the stability study in Table III. We would strated by the AAP-functionalized surface compared expect to see about the same or lower amount of gal- to the 4AB-functionalized surface seems contradic- lium leaching from the samples exposed to saline so- tory since the AAP demonstrated significantly less lutions since the contact angles remained stable and coverage. However, this finding can be attributed to relatively high until the seventh day. As mentioned the inherent nature of the AAP molecule compared previously, this would mean that the adlayer remained to the 4AB molecule. As suggested earlier, the AAP is mostly intact throughout the seven days, thus pre- likely lying more parallel (or in a more “mushroom- venting gallium leaching. Therefore, we can conclude like” configuration) to the surface, which allows for that the 4AB adsorbates must not be forming a com- interaction of the oxygen atoms within each AAP plete monolayer on the GaP surface exposing large molecule and the underlying gallium atoms. This sig- areas of nonfunctionalized GaP. This allows the ions nificant amount of coordination of the oxygen with and H2O2 in the saline to attack at the open areas the gallium most likely prevents the formation of gal- and release gallium into solution. The lack of com- lium oxide. plete coverage by 4AB is most likely the result of its With respect to the low reduction in leaching (only large bulky structure that prohibits it from forming a about 10 or 20 ppb) compared to the clean GaP sam- densely packed adlayer. ple, we can conclude that the water must be having The gallium leaching in the pH solutions (Table VI) a large effect on the adlayers. The water from the in- shows a trend that coincides with the water contact house purification system is slightly acidic, therefore, angle stability study (Table III). The lowest amount the water is most likely etching the adlayer to an ex- of leaching is seen with the pH 5 and 7 solutions indi- tent and leading to surface degradation. cating that the adlayer is stable and remaining intact. The gallium concentration in the saline solutions This prevents gallium leaching from the surface. The shows a similar trend to the water solutions (Table V). greatest leaching occurs when the samples are exposed

TABLE V Gallium concentration in various saline solutions TABLE VI Gallium concentration in various pH solutions ex- exposed to clean and 4AB- and AAP-functionalized GaP sur- posed to 4AB- and AAP-functionalized GaP surfaces faces Sample Ga Concentration (ppm) Sample Ga concentration (ppb) AAP on GaP in pH 5 5 ± 2 Clean Gap in saline 330 ± 30 AAP on GaP in pH 7 100 ± 10 AAP on GaP in saline 140 ± 20 AAP on GaP in pH 9 150 ± 15 AAP on GaP in 1% H2O2 270 ± 30 4AB on GaP in pH 5 0.5 ± 0.2 4AB on GaP in saline 80 ± 10 4AB on GaP in pH 7 0.1 ± 0.1 4AB on GaP in 1% H2O2 270 ± 30 4AB on GaP in pH 9 40 ± 5 D. Ricahards et al.: Azide functionalization on GaP 339 to pH 9, which is consistent with the total wetting ob- considering the XPS calculations, however both func- served after one day of exposure. The adlayers are tionalizations still had considerable gallium leaching. degrading quickly allowing the gallium to form oxide The lack of complete uniform coverage of the surfaces and dissolve into solution. The 4AB-functionalized could be the result of a number of factors including samples leached less gallium compared to the AAP- the molecular structures of the organic azides. Fur- functionalized samples, which attests to the more in- thermore, the strength and wavelength of the UV irra- complete adlayer formed by AAP and its susceptibil- diation, the reaction time, and the solvents could play ity to the buffer conditions. a role in the amount of coverage. While the results ob- The toxicity of gallium has been studied in pre- tained here are beneficial, there must be considerable vious reports due to gallium’s usefulness in treating work done on increasing the adlayer coverage and de- conditions such as hypercalcemia (Warrell et al., ’88). creasing leaching before the implementation of azide This usefulness even prompted the FDA to approve functionalization on GaP for implantable devices. a drug known as GaniteTM (gallium nitrate) for the treatment of this condition. Gallium has also demonstrated anticancer properties especially against lym- Acknowledgment phomas (Warrell et al., ’83; Weick et al., ’83). In a study by Hart et al., gallium, along with aluminum, We thank Arlene Rothwell and Dr. Karl Wood for indium, and thallium were evaluated for their antitu- help with ICP-MS measurements. mor properties as well as their in vivo toxicity. Gal- lium demonstrated a high effectiveness in its tumor suppression along with being the least toxic (Hart References and Adamson, ’71). The normal gallium levels in human blood and human urine is roughly 0.5 ppb and Abramovitch RA, Davis BA. 1964. Preparation + properties of 0.15 ppb, respectively (Liao et al., 2004). The gallium imido intermediates ( imidogens ). Chem Rev 64:149–185. Alibeik S, Zhu SP, Brash JL. 2010. Surface modification with concentrations reported in Tables IV, V, and VI were PEG and hirudin for protein resistance and thrombin neu- the result of 5 mm × 5 mm GaP wafers in 1 mL of tralization in blood contact. Colloid surf. B-Biointerfaces solution. Future implantable devices will require con- 81:389–396. siderably less GaP and will obviously be exposed to Basu R, Kinser CR, Tovar JD, Hersam MC. 2006. Bromine functionalized molecular adlayers on hydrogen passivated more than 1 mL of solution, further alluding to the silicon surfaces. Chem Phys 326:144–150. potential of this material. Chen RX, Cole N, Willcox MDP, Park J, Rasul R, et al. 2009. Synthesis, characterization and in vitro activity of a surface-attached antimicrobial cationic peptide. Biofoul- ing 25:517–524. Conclusions Cui Y, Wei QQ, Park HK, Lieber CM. 2001. Nanowire nanosensors for highly sensitive and selective detection of In this report, GaP was functionalized with two or- biological and chemical species. Science 293:1289–1292. ganic molecules bearing an azide moiety. The first ad- Cumpson PJ. 2001. Estimation of inelastic mean free paths for polymers and other organic materials: use of quantita- sorbate was an aryl azide containing a bromine group tive structure-property relationships. Surf Interface Anal at its opposite end (4AB) for easy XPS identification. 31:23–34. The other was a small PEG azide with an amine group Degennes PG. 1980. Conformations of polymers attached to an interface. Macromolecules 13:1069–1075. at its opposite end (AAP). AFM, XPS, and water con- Fadley CS. 1978. Basic concepts of X-ray photoelectron spec- tact angle were used to characterize the surface. AFM troscopy. New York: Pergamon Press. topography scans indicated a mostly smooth surface Flores-Perez R, Zernlyanov DY, Ivanisevic A. 2008. Quanti- after functionalization with 4AB and a slightly more tative evaluation of covalently bound molecules on GaP (100) surfaces. J Phys Chem C 112:2147–2155. nonuniform surface after functionalization with AAP. Hallstrom W, Martensson T, Prinz C, Gustavsson P,Montelius XPS confirmed the presence of both azides on the sur- L, et al. 2007. Gallium phosphide nanowires as a substrate face. XPS was also used to determine the thickness of for cultured neurons. Nano Lett 7:2960–2965. each adlayer. From this data, it was suggested that Harmer MA. 1991. Phtomodification of surfaces using hete- rocyclic azides. Langmuir 7:2010–2012. the 4AB formed a more ordered adlayer compared to Hart MM, Adamson RH. 1971. Antitumor activity and tox- the AAP adsorbates whose coverage was significantly icity of salts of inorganic group IIIA metals—aluminum, less. A stability study performed on 4AB functional- gallium, indium, and thallium. Proc Natl Acad Sci USA ized surfaces revealed that the adlayer had an excel- 68:1623–1626. Ivanisevic A, Flores-Perez R, Zemlyanov DY. 2008. DNA lent resistance to degradation in saline solutions with molecules on GaP (100) surfaces: spectroscopic charac- varying concentrations of H2O2. ICP-MS was used terization and biospecificity assessment. Chemphyschem to evaluate the amount of gallium leaching into the 9:1528–1530. et al surrounding stability solutions. In general, the 4AB Kim A, Ah CS, Yu HY, Yang JH, Back IB, . 2007. Ultra- sensitive, label-free, and real-time immunodetection using functionalized surfaces decreased the amount of gal- silicon field-effect transistors. Appl Phys Lett 91:103901- lium in solution to a larger extent, which was expected 1–103901-3. 340 SCANNING VOL. 34, 5 (2012)

Kumar A, Singh S, Mudahar GS, Thind KS. 2006. Molar Richards D, Zemlyanov D, Ivanisevic A. 2010. Assessment extinction coefficients of some commonly used solvents. of the passivation capabilities of two different covalent Radiat Phys Chem 75:737–740. chemical modifications on GaP(100). Langmuir 26:8141– Lee KH, Nair PR, Alam MA, Janes DB, Wampler HP, et al. 8146. 2008. Electrical detection of the biological interaction of a Schlapak R, Pammer P, Armitage D, Zhu R, Hinterdorfer charged peptide via gallium arsenide junction-field-effect P, et al. 2006. Glass surfaces grafted with high-density transistors. J Appl Phys 103:114510-1–114510-7. poly(ethylene glycol) as substrates for DNA oligonu- LiaoYH,YuHS,HoCK,WuMT,YangCY,et al. 2004. Bi- cleotide microarrays. Langmuir 22:277–285. ological monitoring of exposures to aluminium, gallium, Warrell RP, Coonley CJ, Straus DJ, Young CW. 1983. Treat- indium, arsenic, and antimony in optoelectronic industry ment of patients with advanced malignant-lymphoma us- workers. J Occup Environ Med 46:931–936. ing gallium nitrate administered as a 7-day continuous Linsmeier CE, Wallman L, Faxius L, Schouenborg J, Bjursten infusion. Cancer 51:1982–1987. LM, et al. 2008. Soft tissue reactions evoked by implanted Warrell RP, Israel R, Frisone M, Snyder T, Gaynor JJ, et al. gallium phosphide. Biomaterials 29:4598–4604. 1988. Gallium nitrate for acute treatment of cancer-related Morota H, Adachi S. 2006. Properties of GaP(001) surfaces hypercalcemia - a randomized, double-blind comparison chemically treated in NH4OH solution. J Appl Phys to calcitonin. Ann Intern Med 108:669–674. 100:054904-1–054904-6. Weick JK, Stephens RL, Baker LH, Jones SE. 1983. Mukherjee J, Peczonczyk S, Maldonado S. 2010. Wet chemical Gallium nitrate in malignant-lymphoma —A Soutwest functionalization of III-V semiconductor surfaces: alkyla- Oncology Group study. Cancer Treat Rep 67:823– tion of gallium phosphide using a grignard reaction se- 825. quence. Langmuir 26:10890–10896. Zerulla D, Chasse T. 2009. Structure and self-assembly of alka- Nahar P, Wali NM, Gandhi RP. 2001. Light-induced activa- nethiols on III-V semiconductor (110) surfaces. J Electron tion of an inert surface for covalent immobilization of a Spectrosc 172:78–87. protein ligand. Anal Biochem 294:148–153. Ziani-Cherif H, Imachi K, Matsuda T. 1999. Preparation Reinhardt KA, Kern W. 2008. Handbook of silicon wafer of aryldiazonium-, aryldiazo-, and arylazido-derivatized cleaning technology. Materials science & process technol- copolymers and their surface photografting. Macro- ogy series. Norwich, NY: William Andrew. molecules 32:3438–3447.