Microstructured Hydroxyapatite Membranes for Ion … · 2016-03-25 · Microstructured...

Microstructured Hydroxyapatite Membranes

for Ion Conducting and Orthopedic Applications

by

Keith Savino

Submitted in Partial Fulfillment of the

Requirements for the Degree

Doctor of Philosophy

Supervised by Professor Matthew Z. Yates

Department of Chemical Engineering

Arts, Science and Engineering Edmund A. Hajim School of Engineering and Applied Sciences

University of Rochester Rochester, New York

2013

II

Biographical Sketch

The author was born in Rochester, New York on September 21, 1985. In 2004, he

graduated from Pittsford Sutherland High School with honors. In 2008, he received his Bachelor

of Science degree in Chemical Engineering and minor in mathematics from Clarkson University

in Potsdam, NY with honors. He came to the University of Rochester in the fall of 2008 as a PhD

candidate in the Department of Chemical Engineering under the supervision of Professor

Matthew Z. Yates. He received a Master of Science degree in Chemical Engineering in 2011.

Funding for the author’s stipend was provided by a Horton Fellowship from the University of

Rochester Laser Lab for Energetics. He was also awarded a one year Breadth fellowship in 2009-

2010 to teach a WRT 105 freshman composition course. In the summer of 2013, he participated

in an NSF I-Corps program to investigate the commercial potential for his research. His area of

research is focused on synthesizing and characterizing novel hydroxyapatite membranes for

potential applications such as electrets, sensors, fuel cells, piezoelectrics, and orthopedic

implants with antimicrobial properties.

III

Publications and Presentations

1) Savino, K., Gabrys, P., Gillespie, M., and M.Z. Yates. Electrochemical reduction of silver

nanoparticles onto hydroxyapatite films. In draft stages. 2) Savino, K. and M.Z. Yates. Improved thermal stability of fluoride doped hydroxyapatite

membranes. In draft stages. 3) Fu, C., Savino, K., Zeng, A., M.Z. Yates, Olvera, D., and Awad, H. Large stored charge in

electrochemically synthesized hydroxyapatite coatings. In draft stages. 4) Guan, B., Savino, K., and M.Z. Yates. Dense and highly textured ytterbium-doped

hydroxyapatite coatings. In draft stages. 5) NSF I-Corps final presentation of lessons learned from customer interviews. August 22, 2013. 6) Microanalytical Reference Materials topical conference,The Microanalysis Society. A New

Synthesis Method of Doped Hydroxyapatite Reference Materials. May 15, 2012. 7) Kong, B., Yu, J., Savino, K., Zhu, Y., and Guan, B. Synthesis of α-calcium sulfate hemihydrate

submicron-rods in water/n-hexanol/CTAB reverse microemulsion. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2012. 409: p. 88-93.

8) Wei, X., et al., Carbonated hydroxyapatite coatings with aligned crystal domains. Crystal

Growth & Design, 2012. 12: p. 3474-3480. 9) Wei, X., et al., Fully dense yttrium-substituted hydroxyapatite coatings with aligned crystal

domains. Crystal Growth & Design, 2012. 12: p. 217-223. 10) Liu, D., K. Savino, and M.Z. Yates, Coating of hydroxyapatite films on metal substrates by

seeded hydrothermal deposition. Surface and Coatings Technology, 2011. 205: p. 3975-3986.

11) Liu, D., K. Savino, and M.Z. Yates, Microstructural engineering of hydroxyapatite membranes

to enhance proton conductivity. Advanced Functional Materials, 2009. 19: p. 3941-3947.

IV

Acknowledgements

The work of this thesis was accomplished only because of the amazing support network

of friends, family, and colleagues I am blessed to have. My support network helped make

graduate school a time in my life for growth and enlightenment that I truly enjoyed. Foremost, I

would like to thank my advisor, Professor Matthew Z. Yates. Preparing this thesis has made me

realize how much I have learned in terms of specific technical knowledge, as well as the ability

to study broad and complex issues. The fact that I have grown so much in my years under his

supervision also highlights how little I knew when I started. I am forever grateful for him

entrusting a naïve first year student to take over this challenging project from his gifted

graduating PhD student, Dongxia Liu. His advising style gave me the freedom to grow as a

researcher and explore side projects that became the cornerstone of this work, yet he also

critiqued me enough so I would not stray too far from my goals. I would also like to thank my

committee members, Professor Jacob Jorne, Professor Paul Funkenbusch, and the committee

chair Professor Amy Lerner for their time, guidance, and feedback with my dissertation.

My family, as they are with everything I do in life, were instrumental in helping me

overcome the struggles that inherently come with being a PhD student. I am forever grateful for

my mom’s terrific meals and motherly insights to help me keep persevering. I appreciate my

sister Dani being a great role model as a hard worker who knows how to have fun, and a

constant friend who is always there for me no matter what, when, and where. And even though

my dad did not have the chance to see me attend the University of Rochester, his impact on my

life was great enough where he already gave me all the tools I need to be successful, in

particular bestowing on me a patient mindset and the ability to keep life in perspective.

V

Perhaps the biggest perk of being a graduate student is interacting daily with extremely

refined and well-rounded colleagues. I would like to thank the other graduate students in the

Yates lab, including Alex Lee, Dr. Sherry Tsai, Dr. Weisi Yin, Dr. Dongxia Liu, Dr. Xue Wei, Pavithra

Ramanarayanan, Professor Baohong Guan, Professor Aibin Zeng, Xuefei Zhang, Cong Fu, Boao

Song, Chaoyi Wan, and Wei Peng.

I would like to thank all of the undergraduates that I was fortunate to work with. While

they may have been in our lab for a learning experience, I feel that my interactions with them

helped me so much more, from getting extra experiments done, to questioning something I’ve

never thought of before, to being a social outlet to help the day go by a little better. I’d like to

thank Sean Rodrigues, Alex Pratt, Anna Couglin, Christian Basil, Rebecca Kelly, Mike Peritz, Ben

Seigal, Victoria Zapata, Zhekai Deng, Mikhail Gillespie, Samara Vilar Da Costa, and Paul Gabrys.

Various staff, students, and associates helped with specific projects or to just get

through the daily grind. I’d like to thank Jason Inzana, Professor Hani Awad, Diana Olvera, Jack

Daiss, and John Varrone for help in the medical center, John Hunt for composition analysis help,

Christine Pratt for XRD measurements, Brian McIntyre for SEM training, Nathan Cooley for fuel

cell testing help, John Miller at the machine shop, Dr. Joseph Bringley of Transparent Materials

for supplies, and all the ChemE secretaries for keeping the department afloat.

Finally, I’d like to thank Jim Bonafini for being a great mentor and offering me my first

job after graduating, David Bowen for all the emails and life insights, The John Ditullio Show for

making 3-6 PM more entertaining, my loyal and loving cat Luke for always being there no matter

what, the four V’s, most of my roommates for making life more entertaining, and finally any

family or friend who I am not specifically mentioning, you all know who you are, so thank you

for being you.

VI

Abstract

Hydroxyapatite (Ca5(PO4)3OH) is a highly studied material because of its similar

composition to bone and its ionic conducting capabilities. This thesis investigates a novel

electrochemical-hydrothermal synthesis method for depositing a hydroxyapatite membrane on

metal substrates. The deposited membrane is highly crystalline, oriented along the

crystallographic c-axis, which is perpendicular to the substrate, and is a uniform array of single

crystals. The novel synthesis method was modified by adding the dopants yttrium, ytterbium,

and fluoride during the hydrothermal synthesis. The thermal stability of the films was

characterized using scanning electron microscopy and X-ray diffraction. It was shown that

fluoride helped reduce dehydroxylation of the membrane at elevated temperatures. The stored

charge of yttrium-fluoride co-doped membranes was measured using the thermally stimulated

depolarization current test. The samples were shown to have extremely high stored charge

values even without electrical polarization. Ytterbium doped membranes were characterized

using scanning electron microscopy, energy-dispersive X-ray spectroscopy, and X-ray diffraction.

Conductivity tests were also performed using electrochemical impedance spectroscopy, as well

as fuel cell performance tests. Finally, a new method of electrochemically depositing silver

nanoparticles onto hydroxyapatite films was developed and characterized in order to make an

antimicrobial coating. Preliminary experiments with Staphylococcus aureus demonstrated the

silver coated sample’s ability to retard bacteria growth.

VII

Contributors and Funding Sources

This work was supervised by a dissertation committee consisting of the author’s advisor

Professor Matthew Z. Yates from the department of chemical engineering, Professor Jacob

Jorne from the department of chemical engineering, Professor Paul Funkenbusch from the

department of mechanical engineering, and the committee chair Professor Amy Lerner from the

department of biomedical engineering. Samples that were used for SEM and XRD analysis in

chapter 4 were prepared by visiting scholars Professor Baohong Guan and Aibin Zeng and led to

a paper in the draft stages as listed in Publications and Presentations. In chapter 5, some of the

silver coated hydroxyapatite samples were prepared by undergraduates Paul Gabrys and Mikhail

Gillespie and led to a paper in the draft stages as listed in Publications and Presentations. In

chapter 6, Samara Vilar Da Costa and Paul Gabrys helped prepare the coated screws and Mikhail

Gillespie helped prepare a hydroxyapatite coating with copper nanoparticles. Graduate study

was supported by the Horton Fellowship from the University of Rochester Laser Lab for

Energetics. A one year Breadth fellowship in 2009-2010 to teach a WRT 105 freshman

composition course was also awarded. Research grants were also provided by the NSF grant

“Surface Crystallization to Optimize Nanostructure of Proton Conductors in Hydrogen

Membrane Fuel Cells” NSF-CMMI award number 0856128, DOE grant DE-FG02-05ER15722, DOE

grant DE-FC03-92SF19460 through the Laboratory for Laser Energetics, a University of Rochester

Provost’s Multidisciplinary Research Award in collaboration with Professor Hani Awad’s group

in the orthopedics department, and a six month grant by the NSF I-Corps program, award

number 1343083.

VIII

Table of Contents

Biographical Sketch…………………………………………………………………………………………………………………….II

Publications and Presentations………………………………………………………………………………………………….III

Acknowledgements……………………………………………………………………………………………………………………IV

Abstract……………………………………………………………………………………………………………………………………..VI

Contributors and Funding Sources……………………………………………………………………………………………VII

Table of Contents…………………………………………………………………………………………………………………….VIII

List of Tables………………………………………………………………………………………………………………………………XI

List of Figures……………………………………………………………………………………………………………………………XII

List of Abbreviations and Symbols………………………………………………………………………………………….XVI

Chapter 1. Introduction………………………………………………………………………………………………………1

1.1. Hydroxyapatite Background…………………………………………………………………………………….1

1.2. Hydroxyapatite Synthesis and Coating Methods………………………………………………………3

1.2.1. Electrochemical Deposition……………………………………………………………………..4

1.2.2. Hydrothermal Synthesis…………………………………………………………………………..5

1.2.3. Electrochemical-hydrothermal Synthesis…………………………………………………6

1.3. Fuel Cell Background……………………………………………………………………………………………….7

1.3.1. Fuel Cell Components and Types……………………………………………………………..8

1.3.2. Fuel Cell Operation…………………………………………………………………………………..9

1.3.3. Thermodynamics and Performance……………………………………………………….10

1.3.4. Hydroxyapatite as an Intermediate Temperature Fuel Cell Electrolyte…..11

1.4. Polarizing Hydroxyapatite …………………………………………………………………………………….13

1.5. Reducing Orthopedic Implant Infections……………………………………………………………….15

IX

1.6. Research Objectives………………………………………………………………………………………………16

1.7. Thesis Outline………………………………………………………………………………………………………..16

References……………………………………………………………………………………………………………………18

Chapter 2. Improved Thermal Stability of Novel Hydroxyapatite Membranes via Fluoride

Doping................................................................................................................22

2.1. Introduction…………………………………………………………………………………………………….22

2.2. Experimental……………………………………………………………………………………………………23

2.3. Results and discussion……………………………………………………………………………………..26

2.4. Conclusions……………………………………………………………………………………………………..39

References…….……………………………………………………………………………………………………………..40

Chapter 3. Stored Charge Analysis of Yttrium-fluoride Co-doped Hydroxyapatite

Membranes……….……………………………………………………………………………………………42

3.1. Introduction…………………………………………………………………………………………………….42

3.2. Experimental……………………………………………………………………………………………………44


3.4. Conclusions……………………………………………………………………………………………………..59

References…………….……………………………………………………………………………………………………..60

Chapter 4. Characterization of Ytterbium and Ytterbium-fluorine Doped Hydroxyapatite

Membranes for Ion Conducting Applications…………………………………………………63

4.1. Introduction…………………………………………………………………………………………………….63

X

4.2. Experimental……………………………………………………………………………………………………67


4.4. Conclusions……………………………………………………………………………………………………..89

References…………….……………………………………………………………………………………………………..91

Chapter 5. Electrochemical Reduction of Silver Nanoparticles onto Hydroxyapatite

Films………………………………………………………………………………………………………………..94

5.1. Introduction…………………………………………………………………………………………………….94

5.2. Experimental……………………………………………………………………………………………………96


5.4. Conclusions……………………………………………………………………………………………………105

References…………….……………………………………………………………………………………………………106

Chapter 6. Conclusions and Future Work……………………………………………………………………….109

6.1. Conclusions………………………………………………………………………………………………………….109

6.2. Future Work………………………………………………………………………………………………………..111

References…………….……………………………………………………………………………………………………113

XI

List of Tables

Table Title Page

Table 1.1. Summary of different fuel cell types and their characteristics. 9

Table 2.1. EDX composition analysis of as-synthesized HAP, FHAP, YHAP, 27 and YFHAP samples.

Table 2.2. Relative intensities of β-TCP formation. Relative intensities were 38 calculated using the ratio of the integral intensity of the (0 2 10), (2 2 0), or (2 1 4) peak of β-TCP with respect to the integral intensity of the (0 0 2) HAP peak.

Table 2.3. Lattice parameters of as-synthesized HAP, FHAP, YHAP, and YFHAP 39 samples.

Table 4.1. EDX results of HAP, YbFHAP, and YbHAP samples with various Yb3+ 76 concentrations. CaHydrothermal/YbHydrothermal is the Ca2+/Yb3+ atomic ratio in the starting hydrothermal solution. CaEDX/YbEDX is the atomic ratio of the synthesized apatite membrane. Table 4.2. Lattice parameter calculations. 77 Table 4.3. Area specific resistance of YbHAP(0.01M)-PSS-Pd-LSCF and 86

YbFHAP-PSS-Pd-LSCF. Table 5.1. Recipe for simulated body fluid. 98 Table 5.2. EDX of HAP and AgHAP. 101 Table 5.3. EDX of new apatite layer after immersion in SBF for HAP, AgHAP, 104

and Ti.

XII

List of Figures

Figure Title Page

Figure 1.1. Illustration of Ca2+ and OH- orientation in HAP and proton migration to 2 a vacancy site. (Adapted from Nakamura et al.)

Figure 1.2. Electrochemical deposition setup. 5

Figure 1.3. HAP seeded substrate undergoing hydrothermal synthesis. 7 Figure 1.4. Electrochemical-hydrothermal synthesis diagram. 7

Figure 1.5. Fuel cell diagram of a proton conducting electrolyte. 10

Figure 1.6. Typical I-V curve for a fuel cell and the region of polarization losses. 11

Figure 1.7. Process for polarizing a HAP crystal. 14 Figure 1.8. Measuring depolarization current of HAP. 14 Figure 2.1. HAP seed layer: a) side view and b) top view. 27 Figure 2.2. HAP samples as-synthesized and heated in air and steam: a) 28

as-synthesized side view, b) as-synthesized top view, c) heated to 600oC in air, d) heated to 700oC in air, e) heated to 800oC in air, f) heated to 900oC in air, g) heated to 900oC in steam, and h) heated to 1,000oC in steam.

Figure 2.3. XRD of unheated HAP. Peaks were labeled using JCPDS 09-0432 for 30

HAP and JCPDS 01-1197 for Ti.

Figure 2.4. XRD of HAP samples as-synthesized, heated to 900oC in steam, heated 31 to 1,000oC in steam, heated to 600oC in air, heated to 700oC in air, heated to 800oC in air, and heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for hydroxyapatite, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.5. FHAP samples as-synthesized and heated in air: a) as-synthesized side 32

view, b) as-synthesized top view, c) heated to 700oC in air, d) heated to 800oC in air, and e) heated to 900oC in air

Figure 2.6. XRD of FHAP samples as-synthesized, heated to 700oC in air, heated 33

to 800oC in air, and heated to 900oC in air. Peaks were labeled using

XIII

JCPDS 09-0432 for hydroxyapatite, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.7. YHAP samples as-synthesized and heated in air: a) as-synthesized side 34

view, b) as-synthesized top view, c) heated to 600oC in air, d) heated to 700oC in air, e) heated to 800oC in air, and f) heated to 900oC in air.

Figure 2.8. XRD of YHAP samples as-synthesized, heated to 600oC in air, heated 35 to 700oC in air, heated to 800oC in air, and heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for hydroxyapatite, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.9. YFHAP samples as-synthesized and heated in air: a) as-synthesized 36 side view, b) as-synthesized top view, c) heated to 800oC in air, and d) heated to 900oC in air.

Figure 2.10. XRD of YFHAP samples as-synthesized, heated to 800oC in air, and 37

heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for hydroxyapatite, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 3.1. a) Illustration of OH- aligned along the crystallographic c-axis 43 surrounded by Ca2+ triangles, b) H+ randomly positioned on the top or bottom of O2-, c) H+ all occupying the bottom position on O2- due to an applied electric field at elevated temperatures, and d) H+ migration to a neighboring O2- with an H+ vacancy. (Adapted from Nakamura et al.)

Figure 3.2. SEM of seed layers: a) EdepSeed top view, b) EdepSeed side view, 48

c) NanoSeed top view, and d) NanoSeed side view. Figure 3.3. SEM of YFHAP samples after hydrothermal growth and PlasmaHAP: a) 50

EdepYFHAP top view, b) EdepYFHAP side view, c) NanoYFHAP top view, d) NanoYFHAP side view, e) PlasmaHAP top view, and f) PlasmaHAP side view.

Figure 3.4. XRD patterns of all samples prepared, with peaks labeled according to 51

JCPDS 09-0432 for HAP and JCPDS 01-1197 for Ti. Figure 3.5. Preferentially oriented crystal growth via an “evolutionary selection” 52

mechanism, as proposed by van der Drift. (Adapted from Gang et al.) Figure 3.6. TSDC curve of EdepYFHAP, NanoYFHAP, and PlasmaHAP. 53 Figure 3.7. Schematic of Ca2+ attraction to the negatively charged surface of a 55

seed layer; initiating the calcium-rich base of a hydrothermally grown

XIV

crystal. Figure 3.8. Potential composition concentration profile of a YFHAP crystal grown 55

during hydrothermal synthesis. Figure 3.9. TSDC curve of unpolarized EdepYFHAP sample heated to 400oC. 57 Figure 3.10. TSDC curve depolarizing the EdepYFHAP sample by keeping it at 400oC 57

for several hours. Inset shows a zoomed in shot of the TSDC curve at the end of the TSDC curve.

Figure 3.11. TSDC curve after polarizing the EdepYFHAP sample at 400oC and 30V 58

for 1 hour. Figure 4.1. Hydrogen membrane fuel cell, with palladium anode, YbHAP or 67

YbFHAP electrolyte, and LSCF cathode.

Figure 4.2. SEM of a) La3+, b) Ce3+, c) Nd3+, d) Eu3+, and e) Yb3+ doped HAP. 73

Figure 4.3. SEM top view images of crystal layers after one hydrothermal growth 75 as a function of Yb concentration: a) [Yb3+] = 0.0M, b) [Yb3+] = 0.005M, c) [Yb3+] = 0.01M, d) [Yb3+] = 0.015M, e) [Yb3+] = 0.02M, and f) [Yb3+] = 0.03M.

Figure 4.4. XRD patterns of HAP powder, HAP seed layer, and HAP, YbHAP, and 77

YbFHAP after one hydrothermal growth on titanium: (a) Analytical grade HAP powder, (b) HAP seed layer, (c) YbFHAP with [Yb] = 0.01M and [F] = 0.02M, (d) HAP, (e) YbHAP(0.005M), (f) YbHAP(0.01M), (g) YbHAP (0.02M), and (h) YbHAP(0.03M). XRD peaks are identified from the hydroxyapatite standard pattern (JCPDS 09-0432), titanium (Ti) is identified with (JCPDS 01-1197), while secondary phases Na3Yb2(PO4)3 and Ca3Yb2O6 are determined from (JCPDS 23-0534) and (JCPDS 19-0263), respectively.

Figure 4.5. Top and side view SEM images of YbHAP(0.01M) after 1, 2, 3, and 4 79

hydrothermal growths of 15 hours each: a) 1 growth top view, b) 1 growth side view, c) 2 growths top view, d) 2 growths side view, e) 3 growths top view, f) 3 growths side view, g) 4 growths top view, h) 4 growths side view.

Figure 4.6. Top and side view SEM images of HAP after four hydrothermal 80

growths: a) HAP top view, b) HAP side view. Figure 4.7. Top and side view SEM images of ytterbium-fluorine co-doped 81

sample (YbFHAP) with [Yb] = 0.01M and [F] = 0.02M, after 1

XV

hydrothermal growth and a fluoride doped sample (FHAP) with [F] = 0.02M, after 4 hydrothermal growths: a) YbFHAP top view, b) YbFHAP side view, c) FHAP top view, d) FHAP side view.

Figure 4.8. SEM images of porous stainless steel and palladium (Pd): a) Porous 82

stainless steel, zoomed in, b) Porous stainless steel, zoomed out, c) Pd on porous stainless steel, zoomed in, d) Pd on porous stainless steel, zoomed out.

Figure 4.9. Nyquist plots of YbHAP-PSS-Pd-LSCF from 400-700oC. 84

Figure 4.10. Nyquist plots of YbFHAP-PSS-Pd-LSCF from 400-700oC. 84

Figure 4.11. Arrhenius plot of YbHAP(0.01M)-PSS-Pd-LSCF and 85 YbFHAP-PSS-Pd-LSCF.

Figure 4.12. I-V and power curves of YbHAP(0.01M)-PSS-Pd-LSCF from 400-700oC. 88

Figure 4.13. I-V and power curves of YbFHAP-PSS-Pd-LSCF from 400-700oC. 89

Figure 5.1. SEM of a) HAP side view, b) HAP top view, c) AgHAP side view, d) 100 AgHAP top view, and e) AgHAP zoomed in.

Figure 5.2. EDX spectrum of AgHAP. 101 Figure 5.3. XRD of HAP and AgHAP. 102 Figure 5.4. SEM of new HAP growth after being immersed in SBF for 24 hours, a) 104

HAP after SBF immersion, b) AgHAP after SBF immersion, c) representative top view of samples in images a and b, and d) Ti after SBF immersion.

Figure 5.5. Growth of S. Aureus bacteria when exposed to AgHAP, HAP, and in 105

suspension. Figure 6.1. Electrochemically deposited hydroxyapatite coating on a titanium 111

dental screw. Figure 6.2. Electrochemical reduction of copper nanoparticles onto the HAP 112

seed layer.

XVI

List of Abbreviations and Symbols

AFM Atomic force microscopy ASR Area specific resistance At% Atomic percent C Coulomb cm Centimeter DI Deionized EDX Energy-dispersive X-ray spectroscopy EIS Electrochemical impedance spectroscopy FHAP Fluoride doped hydroxyapatite h Hour HAP Hydroxyapatite I-V Current-Voltage JCPDS Joint Committee on Powder Diffraction Standards LSCF La0.6Sr0.4Co0.2Fe0.8O3-x M Molar (mol/liter) mC Milicoulomb mg Miligram min Minute mL Mililiter MSC Mesenchymal stem cell nA Nanoamp nm Nanometer OCP Open circuit potential PEM Proton exchange membrane Pd Palladium PBS Phosphate buffered solution PSS Porous Stainless Steel S. aureus Staphylococcus aureus S/cm Sieman/centimeter SEM Scanning electron microscope SOFC Solid oxide fuel cell T Temperature TCP Tricalcium phosphate (Ca3(PO4)2) Ti Titanium TSDC Thermally stimulated depolarization current TTCP Tetracalcium phosphate (Ca4P2O9) V Volt XRD X-ray diffraction YbFHAP Ytterbium-fluoride co-doped hydroxyapatite

XVII

YbHAP Ytterbium doped hydroxyapatite YFHAP Yttrium-fluoride co-doped hydroxyapatite YHAP Yttrium doped hydroxyapatite YSZ Yttria-stabilized zirconia Å Angstrom β Heating rate (oC/minute) σ Conductivity (Siemen/cm) µg Microgram µl Microliter Ω Ohms µm Micrometer

1

Chapter 1

Introduction

1.1. Hydroxyapatite Background

Hydroxyapatite (Ca5(PO4)3OH, HAP) has been studied in a variety of applications such as

sensors, fuel cell electrolytes, filters, and electrets due to its unique electrical properties, as well

as a bone implant material and coating on orthopedic implants due to its similar composition to

bone and osteoconductive capabilities.[1-6] HAP belongs to the hexagonal P63/m space group,

with OH- ions forming one-dimensional chains parallel to the c-axis in the center of Ca2+

triangles.[7] At elevated temperatures (T > 300oC), protons can be conducted through

hydroxyapatite by rotating along hydroxyl end groups, as shown in Figure 1.1.[8] However,

while the ion conducting capabilities of HAP provide a unique platform for a variety of

technologies, the ability to exploit HAP’s ion conducting potential is limited since the

conductivity values are relatively low for HAP produced by traditional methods. Traditional HAP

films have grain boundaries within HAP crystals and high interfacial resistance between crystals,

causing conductivity values to be reduced.[9] Also, many HAP synthesis techniques produce

amorphous coatings, which further reduces proton conductivity since different calcium-

phosphate phases either have poor ionic conduction or have a different ion conduction

mechanism from HAP.

2

Figure 1.1. Illustration of Ca2+ and OH- orientation in HAP and proton migration to a vacancy

site. (Adapted from Nakamura et al.)[8]

Another unique property of HAP is its ability to have ions substituted into its crystal

structure. The chemical formula can be rewritten as Ca5-xAx(PO4)3-yByOH1-zCz, where A represents

a monovalent, bivalent, or trivalent cation that substitutes for Ca2+, B represents a trivalent

anionic oxide that substitutes for (PO4)3-, and C represents a monovalent anion that substitutes

for OH-. One reason for doping HAP is to improve ionic conduction. Low amounts of Y3+

substitution for Ca2+, for example, has been shown to improve proton conductivity.[10, 11]

However, with too much yttrium substitution, the HAP membrane’s ion conducting species can

change from being a proton conductor to an oxygen ion conductor since the OH- ions are

converted to O2- ions to maintain the crystal’s charge neutrality, thus changing the ion

conduction mechanism. A similar improvement in conduction and a change in the conduction

3

mechanism occurs when carbonate substitutes for PO43-.[12] Another reason to substitute HAP

with various ions is to modify its chemical and mechanical properties. Magnesium, strontium,

and carbonate are often substituted in order to make synthetic hydroxyapatite more similar in

composition to natural bone.[13-16] Fluoride is often added to improve dissolution properties,

thermal stability, and hardness.[14, 17-20] Fluoride has also been shown to stimulate

proliferation of bone cells, as well as prevent dental carries.[21]

1.2. Hydroxyapatite Synthesis and Coating Methods.

HAP is typically deposited onto a support substrate since HAP is a brittle material. Ideally, a

coating method would be able to control various film properties such as crystallinity,

composition, film thickness, crystal orientation, and porosity, while also utilizing inexpensive

equipment, be fast, scalable, and able to coat a variety of substrate compositions and

shapes.[22] Numerous coating techniques have been investigated to deposit hydroxyapatite

onto substrates. Thermal spraying is the most commonly used method for coating orthopedic

implants and has the benefit of a high deposition rate with control over the porosity, but the

films often have non-uniform crystallinity due to the high temperature of the spraying

process.[23] Sputter coating produces uniform films, but it is a slow deposition process and can

be costly.[23, 24] Sputter coating and thermal spraying are also line of sight techniques, making

uniform coatings on complex and porous substrates difficult. Other techniques such as sol-gel

coating and dip-coating require a high temperature sintering step to improve crystallinity.[23]

This sintering step can cause the HAP coating to crack, unwanted phase changes of the HAP

coating and substrate, and degradation of mechanical properties of the HAP coating and

substrate.

4

1.2.1. Electrochemical Deposition

Electrochemical deposition of HAP is an inexpensive and fast method to deposit crystals

onto various shaped metallic substrates.[25-29] The substrate and a counter electrode are

placed in an electrolyte solution containing calcium ions, phosphate ions, a buffer to regulate

pH, and salt to improve the solution’s ionic conductivity, as shown in Figure 1.2. When a current

is passed through the electrodes, a local pH increase occurs at the cathode/substrate surface

due to the increase of OH- ions from the hydrolysis of water:[30]

2H2O + 2e- H2 + 2OH- (1)

In addition, a variety of electrochemical reactions can occur at the cathode surface depending

on the local potential and pH:[29, 30]

O2 + 2H2O + 4e- 4OH- (2)

2H+ + 2e- H2 (3)

H2PO4- + H2O + 2e- H2PO3

- + 2OH- (4)

H2PO4- + e- HPO4

2- + 1/2H2 (5)

H2PO4- + 2e- PO4

3- + H2 (6)

HPO42- + e- PO4

3- + 1/2H2 (7)

The local pH increase causes the calcium and phosphate ions to become insoluble, thus initiating

heterogeneous nucleation. As the pH at the surface increases, a variety of calcium-phosphate

apatite phases can form such as dicalcium phosphate dihydrate, octacalcium phosphate, calcium

deficient hydroxyapatite, and stoichiometric hydroxyapatite:[29, 30]

Ca2+ + HPO42- + 2H2O CaHPO4*2H2O (8)

8Ca2+ + 2HPO42- + 4PO4

3- + 5H2O Ca8H2(PO4)6*5H2O (9)

(10-x)Ca2+ + (6-x)PO43- + xHPO4

- + (2-x)OH- Ca(10-x)(HPO4)x(PO4)6-xOH2-x (10)

5

10Ca2+ + 6PO43- + 2OH- Ca10(PO4)6(OH)2 (11)

Adsorbed hydroxyl groups also play a critical role in the nucleation of hydroxyapatite onto the

cathode surface, providing sites for direct nucleation of hydroxyapatite:[29, 30]

TiO2 + 2H2O + e- Ti(OH)3 + (OH)-ads (12)

10Ca2+ + 6PO43- + 2(OH)-

ads Ca10(PO4)6(OH)2 (13)

However, hydrogen gas bubbles that form can obstruct the nucleation and growth of HAP

crystals and can even dislodge previously deposited crystals. Adding H2O2 or ethanol to the

electrolyte solution has been investigated to suppress hydrogen gas production, but it is still

difficult to form a dense and uniform coating of large hydroxyapatite crystals.[31]

Figure 1.2. Electrochemical deposition setup.

1.2.2. Hydrothermal Synthesis

Hydrothermal synthesis is a method for creating large, high purity crystals from aqueous

solution at high temperatures and vapor pressures.[32] The hydrothermal solution is placed

into a pressure vessel, or autoclave, which is a steel cylinder that contains an inner Teflon liner.

In the case of HAP, a calcium chelating agent is often added to help control supersaturation and

6

thus the growth rate.[25] Once heated to a high enough temperature, the calcium-chelating

agent bond breaks, releasing the calcium slowly into the solution. The calcium eventually reacts

with phosphate to grow into larger HAP crystals in solution. A variety of dopants can also be

introduced to the hydrothermal solution to modify the crystal’s composition. While

hydrothermal synthesis provides a means to make high purity crystals with control over their

size, composition, and morphology, it is not a good method for depositing HAP onto metallic

substrates. For most metals, the HAP crystals do not adhere to the surface during the

hydrothermal growth, and if they do adhere, they are sparse and weakly attached.[25]

1.2.3. Electrochemical-hydrothermal Synthesis

To mitigate the problems of electrochemical and hydrothermal synthesis, most of the HAP

samples investigated in this thesis were produced by a novel technique that combines both

methods.[13, 25, 26] First, a HAP seed layer was quickly applied to the substrate using

electrochemical deposition. This produces a uniform, thin HAP coating on the entire substrate

and eliminates the issue of the film cracking since the coating is applied so quickly; typically less

than five minutes. Then, this seeded substrate is placed into an autoclave to undergo a

hydrothermal synthesis, as shown in Figure 1.3. The HAP seed layer on the substrate can then

grow during the hydrothermal synthesis. The hydrothermal synthesis can be repeated with

fresh solution to continue growing the crystals to form an even more dense and thick film if

desired, as shown in Figure 1.4. This method not only produces a uniform film of highly

crystalline HAP crystals with good adhesion to the substrate, the crystals produced are also

highly aligned along the c-axis and contains no grain boundaries along the crystal length. This

unique crystal orientation and morphology is the ideal structure for applications that require

higher proton conductivity.[10, 26]

7

Figure 1.3. HAP seeded substrate undergoing hydrothermal synthesis.

Figure 1.4. Electrochemical-hydrothermal synthesis diagram.

1.3. Fuel Cell Background

Fuel cells offer a promising means to cleanly and efficiently convert a variety of fuels

such as hydrogen, methanol, and natural gas into energy.[33] Since fuel cells produce energy

electrochemically, they are not limited by Carnot Cycle limitations, thus having the potential to

be highly efficient. Unlike batteries, a fuel cell is theoretically capable of operating as long as

reactants are supplied. When hydrogen is the fuel source, the only byproduct is water. If

hydrogen is produced via electrolysis from renewable energy sources such as wind and solar,

then no harmful emissions will enter the environment during the energy production-use cycle.

8

Also, using sources such as nuclear or coal to produce hydrogen via electrolysis or using a fuel

such as natural gas, which is domestically more prevalent, can help reduce dependence on

foreign energy sources.

1.3.1. Fuel Cell Components and Types

While there are a variety of fuel cell types, all fuel cells contain the same three basic

components of an anode, cathode, and an ion conducting membrane, or electrolyte. The anode

and cathode need to be catalytic materials that can convert fuel into ions and electrons, and

then recombine the ions and electrons again on the other side of the electrolyte. The

electrolyte is the heart of the fuel cell and determines the cell’s operating conditions and

determines what kind of anode and cathode will be used. The electrolyte is responsible for ion

conduction, being a barrier that is impermeable to fuel and electrons, while also preventing the

anode and cathode from short circuiting. Electrolytes are typically categorized by the

temperature at which they operate. A fuel cell must operate at a temperature in which the

electrolyte has a high enough ionic conductivity, yet remains chemically and mechanically

stable. Table 1.1 summarizes a few fuel cell types and their characteristics.

9

Table 1.1. Summary of different fuel cell types and their characteristics.[33]

1.3.2. Fuel Cell Operation

Figure 1.5 shows how a proton conducting membrane fuel cell converts hydrogen and

oxygen into electricity. Hydrogen is oxidized into protons and electrons at the anode catalyst.

The electrolyte must then conduct the protons to the cathode, while being impermeable to the

hydrogen gas and electrons. The electrons are then forced back through the anode to a current

collector, which harnesses the electrons to an external circuit to generate electricity. Oxygen

from the air is reduced at the cathode catalyst, where it combines with the electrons and

protons that were formed at the anode to create water. A single-cell can only produce a limited

10

amount of power and produce a theoretical maximum voltage; therefore single-cells can be

combined in series to form a stack for higher voltage and power output.

Figure 1.5. Fuel cell diagram of a proton conducting electrolyte.

1.3.3. Thermodynamics and Performance

For any single-cell of a fuel cell, there is a theoretical maximum reversible voltage

(Vreversible) that the cell can produce, which is given by Equation 1.1:

Vreversible =∆

(1.1)

where ΔG is the Gibbs free energy change of the reaction, z is the number of electrons

transferred in the overall reaction, and F is Faraday’s constant.[33] For a hydrogen fuel cell, this

maximum voltage at room temperature is approximately 1.23 V. However, real fuel cells never

achieve this theoretical maximum due to irreversible loses. The three main types of losses in a

fuel cell are due to activation polarization, ohmic polarization, and concentration polarization.

At low current densities, the activation polarization is the dominant mechanism for a voltage

drop due to the slow reaction kinetics of the electrodes. At intermediate current densities, the

11

voltage will typically have a gradual decrease due to ohmic resistance of the membrane and

interfacial resistance between the electrolyte and electrodes. Finally, at high current densities,

the voltage will drop off to zero because of concentration polarization due to poor transport and

diffusion of reactants to the electrode/electrolyte interface.[33] Figure 1.6 shows a fuel cell’s

performance on a voltage vs. current density (I-V) curve, with a comparison between the

theoretical voltage and an actual fuel cell curve with polarization losses.

Figure 1.6. Typical I-V curve for a fuel cell and the region of polarization losses.

1.3.4. Hydroxyapatite as an Intermediate Temperature Fuel Cell Electrolyte

While there are a wide variety of fuel cells for different applications, as mentioned in

Table 1, developing a commercially viable fuel cell to operate in the intermediate temperature

range (400oC < T < 700oC) has been elusive. The benefit of operating in this temperature range

would be that nonprecious metal catalysts could be used as opposed to platinum in low

temperature fuel cells, the temperatures would be low enough to avoid the complications

induced by high temperature operation, and a variety of fuel sources could be used.[34, 35] The

12

main reason no intermediate temperature fuel cell has been developed is due to the fact that

no electrolyte material has either a high enough conductivity or is chemically and mechanically

stable in this temperature range. HAP is a potential candidate as a proton conducting material

in the intermediate temperature range since proton mobility has been shown to begin as low as

200oC.[8] However, very little work has been done to investigate HAP as a fuel cell electrolyte

since its conductivity values in the intermediate temperature range are too low. For example, a

typical yttria stabilized zirconia (YSZ) membrane of a solid oxide fuel cell (SOFC) has oxygen ion

conductivity values of ~10-2 S/cm at 700oC, whereas a typical hydroxyapatite pellet sintered in

air has proton conductivity values of ~10-6 S/cm at 700oC.[9, 36] Also, when HAP is exposed to

elevated temperatures for an extended period of time, it will often undergo dehydroxylation

and decompose into secondary phases.[9] Since the hydroxyl groups are essential for proton

mobility, this will further decrease HAP’s conductivity.

In this thesis, the issues regarding HAP as an ion conductor were addressed in a variety

of ways. First, the novel electrochemical-hydrothermal synthesis method produces highly

oriented crystals with very few grain boundaries. This provides a direct path for ion conduction,

which reduces grain boundary resistance. Second, the synthesis method produces a very thin,

but dense film. The thin nature of the film reduces the distance that the protons need to travel,

while the density helps maintain a gas tight barrier to prevent fuel crossover. Finally, by doping

the crystal structure, vacancies can be created in the membrane. With the proper amount of

doping, ionic conducitivity can increase compared to an undoped sample. Also, to address the

issues of dehydroxylation, the HAP membranes were doped with fluoride to investigate its

impact on thermal stability.

13

1.4. Polarizing Hydroxyapatite

While HAP coatings have been shown to improve integration of implants in the body, much

research has been done to speed up the integration process. In recent years, one method that

has gained a lot of attention is manipulating HAP’s ion conducting abilities to control the surface

charge.[8, 37] Many studies have shown that a charged surface, particularly negatively charged

surfaces, can help attract positively charged proteins and cations such as Ca2+ to speed up

osteointegration between the implant and the body.[12, 37] While it has been shown that the

a-surface of a HAP crystal typically contains a slight positive charge due to Ca2+ ions and the c-

surface is slightly negative due to PO43- ions exposed at the surface, the electrostatic charge of

both sides is fairly weak.[5] However, HAP’s ion conducting capabilities can be exploited by

applying a high electric field (> 1 kV/cm) to HAP at elevated temperatures (>300oC), forcing

HAP’s protons to migrate toward the negative electrode. By applying the electric field until the

sample cools to room temperature, the electric polarization will remain fixed since the protons

are immobile at room temperature, forming an electret, which is analogous to a magnet having

a permanent magnetic polarization. The process of polarizing HAP is summarized in Figure 1.7.

To measure the amount of stored charge within a polarized material, the polarized sample is

simply connected to a picoammeter and heated to a high enough temperature until the sample

depolarizes, as shown in Figure 1.8. The picoammeter measures the current produced as the

sample is depolarizing, thus giving an estimate of the total stored charge in the material.

However, polarizing traditional HAP membranes with low conductivity values typically produces

stored charges below 100 uC/cm2,[8, 37-40] although recent efforts have increased the stored

charge into the low mC/cm2 range by modifying HAP’s composition with carbonate doping.[41]

14

Figure 1.7. Process for polarizing a HAP crystal.

Figure 1.8. Measuring depolarization current of HAP.

15

1.5. Reducing Orthopedic Implant Infections

While HAP coatings on orthopedic implants can potentially help improve

osteointegration, they do very little to prevent infections. Over 100,000 implants in the U.S.

alone become infected annually, adding $3 billion in costs to the health care system and causing

immeasurable grief to patients who often need revision surgeries.[42] The most common

pathogens include gram-positive staphylococcus aureus and staphylococcus epidermidis, as well

as gram-negative Escherichia coli and pseudomonas aeruginosa.[43] The bacteria can come

from a variety of sources, including the operating room atmosphere, surgical equipment,

clothing from medical staff, and even bacteria on the patient’s skin and already in their body.

Therefore, even in the most sterilized conditions, it can be next to impossible to prevent

bacteria from entering a surgical site. When bacteria enters the body, there is a competition

between the body to integrate with the implant and bacteria to adhere to the implant

surface.[44] For an implant to be implemented successfully, the body needs to integrate with

the implant before any appreciable bacterial adhesion occurs. When bacterial adhesion occurs,

it usually coats the implant in a three step process.[43] 1-2h after implantation, non-specific

and reversible bonding occurs through gravitational, van der Waals, electrostatic, hydrogen

bond, dipole-dipole, ionic bond, and hydrophobic interactions. In the second phase, roughly 2-3

hours later, stronger adhesion occurs via specific chemical interactions between the bacteria

and substrate surface, forming irreversible bonds. Finally, if sufficient nutrients are supplied, a

biofilm can form on the implant surface. If a biofilm forms, it has been found that killing the

bacteria requires roughly 1,000 times the antibiotic dose as would be required for killing cells in

suspension.[45] For this reason, it is critical to provide a means to prevent bacterial adhesion

16

and to stop bacteria growth. In this thesis, HAP samples coated with silver nanoparticles was

investigated as a means to prevent bacterial growth and adhesion. Silver ions have long been

known to contain antimicrobial properties, interfering with the bacteria’s metabolic functions

and replication.[30, 46, 47]

1.6. Research Objectives

This thesis investigates a novel technique for synthesizing highly oriented and grain

boundary-free HAP crystals on metal substrates. The impact of doping fluoride, yttrium, and

ytterbium on various properties such as morphology, composition, crystallinity, thermal

stability, stored charge, and conductivity were investigated using a variety of characterization

techniques such as scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy

(EDX), X-ray diffraction (XRD), electrochemical impedance spectroscopy (EIS), and thermally

stimulated depolarization current (TSDC) measurements. Applications for these membranes

were demonstrated by synthesizing and testing a fuel cell using the HAP membrane as a fuel cell

electrolyte. A new technique of depositing silver nanoparticles on HAP membranes was also

investigated for their potential as an antimicrobial orthopedic coating.

1.7. Thesis Outline

Chapters 2, 3, and 4 all use an electrochemical-hydrothermal synthesis method for

depositing highly c-axis oriented HAP membranes onto metal substrates. Electrochemical

deposition produces a seed layer of HAP onto the metallic substrate, while hydrothermal

synthesis grows the seed layer into a thicker and denser crystal film. The hydrothermal step is

also used to dope HAP with a variety of ions. Chapter 2 looks into the impact that fluoride

17

doping has on the thermal stability of the HAP membranes. Samples were also doped with

yttrium and co-doped with yttrium and fluoride. The differences between doped and undoped

samples were analyzed using XRD and SEM. Chapter 3 explores the stored charge in yttrium-

fluoride co-doped membranes. Samples were prepared using two different seeding steps, one

an electrochemical deposition, the other by letting a suspension of HAP nanoparticles form a

thin film on the substrate, followed by a hydrothermal growth. The stored charge of the films

was characterized using the TSDC method. Since the films were not polarized, a mechanism for

their large stored charge was proposed. Chapter 4 investigates the impact of ytterbium doping

on the crystal film’s structure, morphology, and composition. Fuel cells were fabricated and

tested using a ytterbium and ytterbium-fluoride co-doped membrane coated onto a palladium

substrate. Finally, chapter 5 describes a new technique to reduce silver nanoparticles onto

electrochemically deposited HAP crystals. The samples were characterized using SEM and XRD,

their bioactivity was evaluated by placing the samples into a simulated body fluid, and

preliminary experiments were done to investigate their antimicrobial properties.

18

References

1. Tanka, Y., et al., Ionic conduction mechanism in Ca-deficient hydroxyapatite whiskers.

Materials Science and Engineering, 2009. 161: p. 115-119.

2. Gittings, J., et al., Electrical characterization of hydroxyapatite-based bioceramics. Acta

Biomaterialia, 2009. 5: p. 743-754.

3. Palazzo, B., et al., Biomimetic hydroxyapatite–drug nanocrystals as potential bone

substitutes with antitumor drug delivery properties. Advanced Functional Materials, 2007.

17(2180-2188).

4. Kim, S., et al., The characteristics of a hydroxyapatite-chitosan-PMMA bone cement.

Biomaterials, 2004. 25: p. 5715-5723.

5. Kawasaki, T., Hydroxyapatite as a liquid chromatographic packing. Journal of

Chromatography, 1991. 544: p. 147-184.

6. Owada, H., et al., Humidity-Sensitivity of Yttrium Substituted Apatite Ceramics. Solid

State Ionics, 1989. 35: p. 401-404.

7. Horiuchi, N., et al., Proton conduction related electrical dipole and space charge

polarization in hydroxyapatite. Journal of Applied Physics, 2012. 112: p. 1-6.

8. Nakamura, S., H. Takeda, and K. Yamashita, Proton transport polarization and

depolarization of hydroxyapatite ceramics. Journal of Applied Physics, 2001. 89: p. 5386-5391.

9. Yamashita, K., K. Kitagaki, and T. Umegaki, Thermal Instability and Proton Conductivity of

Ceramic Hydroxyapatite at High Temperatures. J. Am. Ceram. Soc., 1995. 78: p. 1191-97.

10. Wei, X. and M.Z. Yates, Yttrium-doped hydroxyapatite membranes with high proton

conductivity. Chemistry of Materials, 2012. 24: p. 1738-1743.

11. Yamashita, K., et al., Protonic conduction in yttrium-substituted hydroxyapatite ceramics

and their applicability to H2-O2 fuel cell. Solid State Ionics, 1990. 40/41: p. 918-921.

12. Tanaka, Y., et al., Fast oxide ion conduction due to carbonate substitution in

hydroxyapatite. Journal of American Ceramic Society, 2010. 93: p. 3577-3579.

13. Wei, X., et al., Carbonated hydroxyapatite coatings with aligned crystal domains. Crystal

Growth & Design, 2012. 12: p. 3474-3480.

14. Samia, N., et al., Effect of fluorine on the thermal stability of the magnesium-substituted

hydroxyapatite. Annales De Chimie-Science Des Materiaux, 2011. 36: p. 159-176.

19

15. Liu, W., et al., Gentamicin-loaded strontium-containing hydroxyapatite bioactive bone

cement—An efficient bioactive antibiotic drug delivery system. Journal of Biomedical Materials

Research B: Applied Biomaterials, 2010. 95B: p. 397-406.

16. C. Ergun, et al., Hydroxylapatite with substituted magnesium, zinc, cadmium, and

yttrium. I. Structure and microstructure. Journal of Biomedical Materials Research, 2001. 59: p.

305-311.

17. Eslami, H., M. Solati-Hashjin, and M. Tahriri, Effect of fluorine ion addition on structural,

thermal, mechanical, solubility and biocompatibility characteristics of hydroxyapatite

nanopowders. Advances in Applied Ceramics, 2010. 109: p. 200-212.

18. Hammari, L.E., et al., Chrystallinity and Fluorine Substitution Effects on the Proton

Conductivity of Porous Hydroxyapatites. J. Solid State Chemistry, 2004. 177: p. 134-138.

19. Laghzizil, A., et al., Comparison of Electrical Properties between Fluoroapatite and

Hydroxyapatite Materials. Journal of Solid State Chemistry, 2001. 156: p. 57-60.

20. Chen, Y. and X. Miao, Thermal and chemical stability of fluorohydroxyapatite ceramics

with different fluorine contents. Biomaterials, 2005. 26: p. 1205-1210.

21. Bianco, A., et al., F-substituted hydroxyapatite nanopowders: Thermal stability, sintering

behaviour and mechanical properties. Ceramics International, 2010. 36: p. 313-322.

22. Zakaria, S., et al., Nanophase hydroxyapatite as a biomaterial in advanced hard tissue

engineering: A review. Tissue Eng. Part B Rev., 2013.

23. Sun, L., et al., Material fundamentals and clinical performance of plasma-sprayed

hydroxyapatite coatings: A review. Journal of Biomedical Materials Research, 2001. 58: p. 570-

592.

24. Yang, Y., k. Kim, and J. Ong, A review on calcium phosphate coatings produced using a

sputtering process—an alternative to plasma spraying. Biomaterials, 2005. 26: p. 327-337.

25. Liu, D., K. Savino, and M.Z. Yates, Coating of hydroxyapatite films on metal substrates by


26. Liu, D., K. Savino, and M.Z. Yates, Microstructural engineering of hydroxyapatite

membranes to enhance proton conductivity. Advanced Functional Materials, 2009. 19: p. 3941-

3947.

20

27. Kuo, M. and S. Yen, The process of electrochemical deposited hydroxyapatite coatings on

biomedical titanium at room temperature. Materials Science and Engineering C, 2002. 20: p.

153-160.

28. S.Ban and S. Maruno, Hydrothermal-electrochemical deposition of hydroxyapatite.

Journal of Biomedical Materials Research, 1998. 42: p. 387-395.

29. Wang, J., et al., Fluoridated hydroxyapatite coatings on titanium obtained by

electrochemical deposition. Acta Biomaterialia, 2009. 5: p. 1798-1807.

30. Bir, F., et al., Electrochemical depositions of fluorohydroxyapatite doped by Cu2+, Zn2+,

Ag+ on stainless steel substrates. Applied Surface Science, 2012. 258: p. 7021-7030.

31. Chen, J., H. Juang, and M. Hon, Calcium phosphate coating on titanium substrate by a

modified electrocrystallization process. Journal of Materials Science Materials in Medicine, 1998.

9: p. 297-300.

32. Neira, I., et al., An effective morphology control of hydroxyapatite crystals via

hydrothermal synthesis. Crystal Growth and Design, 2009. 9: p. 466-474.

33. Barbir, F., PEM Fuel Cells: Theory and Practice, ed. R. Dorf2005: Elsevier Academic Press.

34. Ito, N., et al., Electrochemical analysis of hydrogen membrane fuel cells. Journal of

Power Sources, 2008. 185: p. 922-926.

35. Ito, N., et al., New intermediate temperature fuel cell with ultra-thin proton conductor

electrolyte. Journal of Power Sources, 2005. 152: p. 200-203.

36. Kwon, O. and G. Choi, Electrical conductivity of thick film YSZ. Solid State Ionics, 2006.

177: p. 3057-3062.

37. Tanaka, Y., et al., Polarization and microstructural effects of ceramic hydroxyapatite

electrets. Journal of Applied Physics, 2010. 107: p. 1-10.

38. Wolf-Brandstetter, C., et al., The impact of heat treatment on interactions of contact-

poled biphasic calcium phosphates with proteins and cells. Acta Biomaterialia, 2012. 8: p. 3468-

3477.

39. Kumar, D., et al., Polarization of hydroxyapatite: Influence on osteoblast cell

proliferation. Acta Biomaterialia, 2010. 6: p. 1549-1554.

40. Bodhak, S., S. Bose, and A. Bandyopadhyay, Bone cell–material interactions on metal-ion

doped polarized hydroxyapatite. Materials Science and Engineering C, 2011. 31: p. 755-761.

21

41. Nagai, A., et al., Electric polarization and mechanism of B-type carbonated apatite

ceramics. Journal of Biomedical Materials Research Part A, 2011. 99A: p. 116-124.

42. Darouiche, R., Treatment of infections associated with surgical implants. New England

Journal of Medicine, 2004. 350: p. 1422-1429.

43. Hetrick, E. and M. Schoenfisch, Reducing implant-related infections: active release

strategies. Chemical Society Reviews, 2006. 35: p. 780-789.

44. Gristina, A., Biomaterial-centered infection: microbial adhesion versus tissue integration.

Science, 1987. 237: p. 1588-1595.

45. Smith, A., Biofilms and antibiotic therapy: Is there a role for combating bacterial

resistance by the use of novel drug delivery systems? Advanced Drug Delivery Reviews, 2005. 57:

p. 1539-1550.

46. Lu, X., et al., Nano-Ag-loaded hydroxyapatite coatings on titanium surfaces by

electrochemical deposition. J. R. Soc. Interface, 2011. 8: p. 529-539.

47. Rai, M., A. Yadav, and A. Gade, Silver nanoparticles as a new generation of

antimicrobials. Biotechnology Advances, 2009. 27: p. 76-83.

22

Chapter 2

Improved Thermal Stability of Novel Hydroxyapatite Membranes via Fluoride Doping

2.1. Introduction

Due to hydroxyapatite’s (Ca10(PO4)6(OH)2, HAP) similar composition to bone and

osteoconductive properties, it has been extensively studied as a synthetic bone substitute and

drug delivery carrier.[1, 2] Recently, HAP has also been investigated for high temperature

applications such as sensors, ion conducting membranes, and electrets due to its proton

conducting capabilities at elevated temperatures.[3-6] Proton mobility within the HAP crystal

structure begins around 300oC, with a conduction mechanism of rotation along the hydroxyl end

groups.

Unfortunately, at elevated temperatures, calcium deficient HAP has been shown to

dehydroxylate according to the proposed equation below:[7]

Ca10-z(HPO4)2z(PO4)6-2z(OH)2 Ca10-z(P2O7)z-s(PO4)6-2z+2s(OH)2(1-s) + (z+s)H2O (2.1)

Upon further heating, calcium deficient HAP will decompose to byproducts such as β-tricalcium

phosphate (β-Ca3(PO4)2 ,β-TCP), completely removing all hydroxyl groups, as well as

stoichiometric HAP:

Ca10-z(P2O7)z-s(PO4)6-2z+2s(OH)2(1-s) + (z+s)H2O (1-z)Ca10(PO4)6(OH)2 + 3zCa3(PO4)2 + (z-s)H2O (2.2)

The dehydroxylation process is accelerated the more calcium deficient HAP is, with

dehydroxylation occuring as low as 500oC.[7] Decomposition can be detrimental for devices

that exploit HAP’s ion conducting abilities since complete removal of the hydroxyl groups will

change its conduction mechanism and properties.[4, 8, 9] Another application where

23

dehydroxylation can be problematic is when HAP coatings are being sintered onto orthopedic

implants since the decomposition byproducts can dissolve in vivo faster than pure HAP.[10]

To improve HAP’s thermal stability, many groups have shown the benefit of substituting

the hydroxyl end groups with fluoride due to fluoride’s higher affinity for protons and its ability

to create a more ordered HAP crystal structure.[10-12] This enhanced thermal stability makes

fluorinated hydroxyapatite (FHAP) more resistant to decomposition at elevated temperatures.

This study shows that incorporating fluoride into a novel electrochemical-hydrothermal

synthesis method can improve the film’s thermal stability at elevated temperatures compared

to unfluorinated samples. To further demonstrate that dehydroxylation is the mechanism that

leads to HAP’s decomposition, unfluorinated samples were heated in a steam atmosphere to

show that introducing steam can shift the decomposition equations toward the original HAP.

Finally, the ability to co-dope samples was demonstrated since multiple dopants such as yttrium,

magnesium, and carbonate are often incorporated into HAP to modify its properties.[12-15] For

this study, yttrium-fluoride co-doped HAP films (YFHAP) were synthesized and shown to have

enhanced thermal stability. Samples only doped with yttrium (YHAP) were synthesized as a

comparison and showed poor thermal stability at elevated temperatures.

2.2. Experimental

Materials

Analytical grade Ca(NO3)2·4H2O (99.0% purity) and Y(NO3)3·6H2O (99.9% purity) were

purchased from Alfa Aesar. (NH4)2HPO4 (>99.0% purity) was purchased from EMD. K2HPO4

(99.99% purity), CaCl2·2H2O (99+% purity), NaCl (≥99.0% purity), tris(hydroxymethyl)-

aminomethane (tris) (99.8+% purity), and disodium ethylenediaminetetraacetate dihydrate

24

(Na2EDTA·2H2O) (ACS reagent, 99.0-101.0% purity) were all obtained from Sigma- Aldrich. NH4F

(ACS reagent, 98% purity) was purchased from Aldrich Chemical Company, Inc. 36.5-38.0%

hydrochloric acid and 28.0-30.0% pure ammonium hydroxide were purchased from Mallinckrodt

Chemicals. Detergent powder was purchased from Alconox. Titanium(Ti) substrates (12.5 mm x

12.5 mm and 0.89 mm thick) and platinum foil (25 mm x 25 mm and 0.127 mm thick) were

purchased from Alfa Aesar. Deionized water was used for all solutions.

Electrochemical deposition of HAP seeds

Titanium substrates were polished with SiC paper (800 grit) to provide surface

roughness. The substrates were then washed with Alconox detergent powder, rinsed with tap

water, attached to a silver wire by tightly wrapping the wire through a premade hole in the

substrate, sonicated in an ethanol/acetone (volume ratio = 50/50) solvent for 30 minutes, and

then rinsed with deionized water.

An electrolyte solution was prepared containing 138 mM NaCl, 50 mM tris, 1.25 mM

CaCl2, and 0.828 mM K2HPO4 in 125 mL deionized water per substrate. The solution was

buffered to pH 7.2 using hydrochloric acid. The electrolyte solution was then heated to 95°C.

The substrate and a platinum plate were held parallel to each other with a fixed distance of

separation of 10 mm. The substrate and platinum plate were connected to the negative and

positive outlets of a direct current power supply (Instek GPS-3030D), respectively, and

immersed in the electrolyte solution. The electrochemical reaction was carried out for 5

minutes at 12.5 mA/cm2 (area relative to the platinum plate), then rinsed off with deionized

water and dried in air.

Hydrothermal growth of HAP

25

The hydrothermal solution consisted of 0.115 M Na2EDTA·2H2O that was stirred into 30

mL of deionized water per sample at 80oC until it was completely dissolved. 0.1 M

Ca(NO3)2·4H2O was then added, followed by 0.06 M (NH4)2HPO4. Hydroxyapatite (HAP) samples

contained no other dopants. For yttrium doped hydroxyapatite (YHAP), 0.01 M Y(NO3)3·6H2O

was added after Ca(NO3)2·4H2O. For fluoride doped hydroxyapatite (FHAP), 0.02 M NH4F was

added after (NH4)2HPO4. For yttrium and fluoride co-doped hydroxyapatite (YFHAP), the

solution contained both Y(NO3)3·6H2O and NH4F. The pH of each solution was raised to 10.0

with ammonium hydroxide and then stirred for about 10 minutes.

The electrolyte solution was then transferred to a Teflon-lined stainless steel acid

digestion autoclave (Parr Item No. 4744, 45 mL internal volume), until filled with approximately

60% solution. The seeded substrate was submersed in the solution with the seed layer facing

down and tilted ~45 degrees relative to the bottom of the autoclave. The autoclave was heated

for 10 hours at 200°C and autogenous pressure. After heating the autoclave was cooled to room

temperature in a fume hood, then the sample was taken out, rinsed with deionized water

several times, and dried in air.

Heat treatment

For each heat treatment temperature, a sample was placed into a tube furnace for 3

hours at the set point temperature of 600, 700, 800, 900, or 1000oC. The ramp rate for heating

and cooling was 2oC/minute with the heating beginning and ending at room temperature. For

samples heated in steam, a peristaltic pump (Gilson, Minipuls 3) delivered a steady supply of

water at a flow rate of 7 mL/minute through the alumina tube samples were in. The pump was

turned on when the temperature of the oven reached 200oC, then turned off again when it

cooled back down to 200oC.

26

Product characterization

Crystal morphology was examined using a scanning electron microscope (SEM, Zeiss-

Auriga) with an accelerating voltage of 3 kV. The composition of HAP membranes was

determined by EDX (Energy-dispersive X-ray spectroscopy, EDAX). Three spots on each sample

were probed at 15 kV and the values were averaged and standard deviations were calculated.

The crystal structure of HAP was determined by X-ray diffraction (XRD) (Philips PW3020) with Cu

K1 radiation (λ=1.540560 Ǻ) from 20-40o with a step rate of 0.02 degrees/second. Phase

identification was made by comparison with the Joint Committee on Powder Diffraction

Standards (JCPDS) files. Relative intensities were calculated using X’Pert Data Viewer software.

2.3. Results and discussion

Figure 2.1 shows a typical seed layer after the electrochemical deposition. The height of

the seed layer is roughly 500-800 nm, while the width of each crystal is about 20-25 nm. The

seed layer provides nucleation sites for further crystal growth during the hydrothermal

synthesis. After the hydrothermal growth, EDX was used to characterize the samples’

compositions as shown in Table 2.1. Yttrium atomic% (at%) was confirmed using an electron

microprobe since the yttrium and phosphorous peaks are very close . Electron microprobe

results agreed well with EDX (results not shown). EDX clearly confirmed the presence of

fluoride, yttrium, and yttrium and fluoride in FHAP, YHAP, and YFHAP, respectively. The calcium

to phosphorous (Ca:P) ratio was also determined since HAP’s thermal stability is influenced by

how close the ratio is to stoichiometric, or 1.67. For samples with yttrium substitution, the

amount of yttrium was considered when calculating the Ca:P ratio because yttrium substitutes

27

for calcium, therefore the (Ca+Y):P was determined. All the samples were verified to have a

slight calcium deficiency.

Figure 2.1. HAP seed layer: a) side view and b) top view.

Table 2.1. EDX composition analysis of as-synthesized HAP, FHAP, YHAP, and YFHAP samples.

Figures 2.2 a and b show the side and top view, respectively, of a typical as-synthesized

HAP sample after the hydrothermal growth, unheated. Each individual crystal has a smooth

surface, with dimensions of roughly 10 µm tall and 1-2 µm wide. Figures 2.2 c and d show very

few morphological changes when heating the HAP sample to 600oC or 700oC in air, although a

few small grains on the top surface of some crystals form at 700oC. When heated to 800oC,

however, grain boundaries form on the entire surface of a majority of the crystals, as shown in

Figure 2.2 e. By 900oC, all of the crystals contain grain boundaries along their entire surfaces.

However, when these samples are heated to 900oC and 1000oC in a steam atmosphere, this

grain boundary formation is prevented, as shown in Figures 2.2 g and h, respectively.

28

Figure 2.2. HAP samples as-synthesized and heated in air and steam: a) as-synthesized side view,

b) as-synthesized top view, c) heated to 600oC in air, d) heated to 700oC in air, e) heated to 800oC in air, f) heated to 900oC in air, g) heated to 900oC in steam, and h) heated to 1,000oC in

steam.

29

Figure 2.3 shows the X-ray diffraction (XRD) pattern of the HAP as-synthesized sample,

or sample with no heating. All the peaks match the standard XRD pattern of HAP (JCPDS 09-

0432) which is from the film or titanium (JCPDS 01-1197) which is due to the substrate. The

largest peak represents the (0 0 2) plane and verifies that the crystals grow preferentially along

the crystallographic c-axis, which is perpendicular to the substrate. For the rest of the XRD

patterns, in order to zoom in and analyze the smaller peaks, the top of the (0 0 2) peak is cut off.

Figure 2.4 shows the XRD results of HAP being heated to various temperatures and confirms

that the morphological changes are associated with the progression of dehydroxylation and the

eventual conversion of HAP to β-TCP. Upon heating the HAP sample, new peaks are observed.

The highest, second highest, and third highest peaks of β-TCP ((0 2 10), (2 2 0), and (2 1 4)

respectively, according to JCPDS 09-0169) become more intense as HAP is heated to higher

temperatures. Titanium dioxide (JCPDS 21-1276) can also be detected at higher temperatures

as the titanium substrate oxidizes. For the samples heated in steam, however, the β-TCP peaks

are suppressed since steam shifts the equilibrium of equation 2.1 toward the reactants, or the

calcium deficient HAP, according to Le Chatelier’s principle.

30

Figure 2.3. XRD of unheated HAP. Peaks were labeled using JCPDS 09-0432 for HAP and JCPDS

01-1197 for Ti.

31

Figure 2.4. XRD of HAP samples as-synthesized, heated to 900oC in steam, heated to 1,000oC in

steam, heated to 600oC in air, heated to 700oC in air, heated to 800oC in air, and heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for HAP, JCPDS 09-0169 for β-TCP, JCPDS 01-

1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.5 shows SEM images of FHAP samples as-synthesized and being heated from

700-900oC in air. The crystals maintain the same morphology after every heating process

investigated, showing an enhanced thermal stability for fluorinated samples. Figure 2.6 shows

the XRD pattern of the FHAP samples being heated. None of the β-TCP peaks are detected upon

heating, confirming the hypothesis that fluoride’s high affinity for protons and its ability to

create a more ordered structure can prevent dehydroxylation.

32

Figure 2.5. FHAP samples as-synthesized and heated in air: a) as-synthesized side view, b) as-

synthesized top view, c) heated to 700oC in air, d) heated to 800oC in air, and e) heated to 900oC in air

33

Figure 2.6. XRD of FHAP samples as-synthesized, heated to 700oC in air, heated to 800oC in air,

and heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for HAP, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.7 shows SEM images of YHAP samples as-synthesized and after being heated to

600-900oC in air. The crystal morphology begins to show signs of changes at 600oC, and above

700oC multiple grain boundaries form on most crystals. Figure 2.8 shows the XRD pattern of

these samples, with enhanced peaks for β-TCP arising upon heating. Therefore, yttrium

substitution alone does not prevent dehydroxylation, as expected.

34

Figure 2.7. YHAP samples as-synthesized and heated in air: a) as-synthesized side view, b) as-synthesized top view, c) heated to 600oC in air, d) heated to 700oC in air, e) heated to 800oC in

air, and f) heated to 900oC in air.

35

Figure 2.8. XRD of YHAP samples as-synthesized, heated to 600oC in air, heated to 700oC in air, heated to 800oC in air, and heated to 900oC in air. Peaks were labeled using JCPDS 09-0432 for

HAP, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Figure 2.9 shows SEM images of YFHAP samples as-synthesized and being heated to 800

and 900oC in air. The crystals maintain the same morphology after heating at both

temperatures. The XRD pattern in Figure 2.10 shows that none of the major β-TCP peaks form,

which demonstrates the ability of fluoride to maintain thermal stability with other dopants.

Therefore, a variety of elements can potentially be doped with fluoride to modify the crystal

properties, yet with an enhanced thermal stability. It should be noted that the high peak

intensity for TiO2 is observed at 900oC due to the fact that the YFHAP samples are roughly half as

thick as the HAP, FHAP, and YHAP samples, making TiO2 easier to detect on YFHAP samples that

are heated.

36

Figure 2.9. YFHAP samples as-synthesized and heated in air: a) as-synthesized side view, b) as-

synthesized top view, c) heated to 800oC in air, and d) heated to 900oC in air.

37

Figure 2.10. XRD of YFHAP samples as-synthesized, heated to 800oC in air, and heated to 900oC

in air. Peaks were labeled using JCPDS 09-0432 for HAP, JCPDS 09-0169 for β-TCP, JCPDS 01-1197 for Ti, and JCPDS 21-1276 for TiO2.

Table 2.3 is a quantitative comparison of the amount of β-TCP in HAP and YHAP at

various temperatures from X-ray diffraction using the equation for relative intensity (R.I.):

(. . ) = !@(#$%#),($$#),'($%())*!@(##$) +100 2.3

The integral intensity is the area under the specified β-TCP peak; either (0 2 10), (2 2 0), or (2 1

4), or the HAP (0 0 2) peak. The HAP and YHAP samples heated in air clearly have increased

amounts of β-TCP as temperature is increased, while the FHAP, YFHAP, and HAP samples heated

in steam had no quantifiable amounts of β-TCP.

38

Table 2.2. Relative intensities of β-TCP formation. Relative intensities were calculated using the ratio of the integral intensity of the (0 2 10), (2 2 0), or (2 1 4) peak of β-TCP with respect to the

integral intensity of the (0 0 2) HAP peak.

The XRD patterns were also used to calculate the a and c lattice parameters for each as-

synthesized sample after the hydrothermal growth. The c lattice parameter was calculated from

the (0 0 2) peak using the hexagonal close-packed (HCP) unit cell relationship given in equation

2.4 below:

%./ =

(0

1/21323// + + /

5/ 2.4

where d is the d-spacing, or distance between adjacent planes in the set of Miller indices (h k l),

and a and c represent the a and c lattice parameters, respectively. Once the c lattice parameter

was known, the a lattice parameter could be determined from the (1 0 2) peak and using

equation 2.4. The volume (V) of the hexagonal unit cell was calculated using equation 2.5:

V = 2.589a2c 2.5

Lattice parameters and unit cell volumes of all the as-synthesized samples are given in Table 2.2.

The samples have similar lattice parameters to standard HAP (JCPDS 9-0432), where a = 9.418

and c = 6.884. Evis et al. has shown that doping Y+3 and F- in HAP does not dramatically impact

the lattice parameters, with increasing F- content only slightly decreasing the a lattice

parameter.[16] The results in Table 2.3 also show that F- substitution does indeed slightly

39

decrease the a lattice parameter. Also, EDX shows that YFHAP contains slightly more F- than

FHAP, which is consistent with the fact that YFHAP has a smaller a lattice parameter than FHAP.

Table 2.3. Lattice parameters of as-synthesized HAP, FHAP, YHAP, and YFHAP samples.

2.4. Conclusions

The novel electrochemical-hydrothermal hydroxyapatite synthesis method presented in

this work has a lot of potential in ion conducting applications such as sensors, electrets, and fuel

cell electrolytes. However, most of these applications require temperatures from 200oC up to

900oC, which may induce dehydroxylation, thus inhibiting the sample’s ability to conduct ions

and also alter other properties such as its mechanical integrity. By incorporating fluoride into

the hydrothermal growth solution, the hydroxyapatite crystals that were synthesized have some

of the hydroxyl ions replaced by fluoride, but without altering the crystal structure, morphology,

and orientation. The fluorinated samples showed enhanced thermal stability at temperatures as

high as 900oC using SEM and XRD analysis. Fluoride can also be co-doped with other ions such

as yttrium to make films with improved thermal stability, while enhancing other properties such

as conductivity.

40

References

1. Koutsopoulos, S., Synthesis and characterization of hydroxyapatite crystals: A review

study on the analytical methods. Journal of Biomedical Materials Research, 2002. 62: p. 600-612.

2. S.Ban and S. Maruno, Hydrothermal-electrochemical deposition of hydroxyapatite.

Journal of Biomedical Materials Research, 1998. 42: p. 387-395. 3. Andres-Verges, M., et al., A new route for the synthesis of calcium-deficient

hydroxyapatites with low Ca/P ratio: Both spectroscopic and electric characterization. Journal of Materials Research, 2000. 15: p. 2526-2533.




Materials Science and Engineering, 2009. 161: p. 115-119. 6. Tanaka, Y., et al., Polarization and microstructural effects of ceramic hydroxyapatite

electrets. Journal of Applied Physics, 2010. 107: p. 1-10. 7. S.Raynaud, et al., Calcium phosphate apatites with variable Ca/P atomic ratio I.

Synthesis, characterisation and thermal stability of powders. Biomaterials, 2002. 23: p. 1065-1072.

8. Maiti, G. and F. Freund, Influence of Fluorine Substitution on the Proton Conductivity of

Hydroxyapatite. J. Chem. Soc., Dalton Trans., 1981: p. 949-955. 9. Yamashita, K., K. Kitagaki, and T. Umegaki, Thermal Instability and Proton Conductivity of

Ceramic Hydroxyapatite at High Temperatures. J. Am. Ceram. Soc., 1995. 78([5]): p. 1191-97.


with different fluorine contents. Biomaterials, 2005. 26: p. 1205-1210. 11. Eslami, H., M. Solati-Hashjin, and M. Tahriri, Effect of fluorine ion addition on structural,

thermal, mechanical, solubility and biocompatibility characteristics of hydroxyapatite

nanopowders. Advances in Applied Ceramics, 2010. 109: p. 200-212. 12. Nasr, S., et al., Effect of fluorine on the thermal stability of the magnesium-substituted

hydroxyapatite. Annales De Chimie-Science Des Materiaux, 2011. 36: p. 159-176.

41

13. Owada, H., et al., Humidity-Sensitivity of Yttrium Substituted Apatite Ceramics. Solid State Ionics, 1989. 35: p. 401-404.

14. Owada, H., K. Yamashita, and T. Kanazawa, High-temperature stability of hydroxyl ions in

yttrium-substituted oxyhydroxyapatites. Journal of Materials Science Letters, 1990. 9: p. 26-28.



16. Basar, B., et al., Improvements in microstructural, mechanical, and biocompatibility

properties of nano-sized hydroxyapatites doped with yttrium and fluoride. Ceramics International, 2010. 36: p. 1633-1643.

42

Chapter 3

Stored Charge Analysis of Yttrium-fluoride Co-doped Hydroxyapatite Membranes

3.1. Introduction

Hydroxyapatite (HAP) is a highly studied material in the orthopedics community as an

implant coating and bone filler due to its similar composition to bone and its osteoconductive

properties.[1-8] Recently, there has been a lot of interest to improve HAP’s osteoconductive

properties by negatively polarizing HAP’s surface with a high electric field (>1 kV/cm) at elevated

temperatures (T > 300oC).[9-16] Negatively charged HAP surfaces have been shown to improve

osteoblast adhesion and stimulate new apatite formation since the negative surface attracts

positively charged proteins and calcium ions via electrostatic attraction,[10, 12, 17, 18],

although a recent study has shown that a significant factor in cell adhesion and growth may be

due to any heat treatment that removes surface carbon contamination.[19] Regardless of the

uncertainty over polarization’s impact on cell growth and adhesion, the ability to polarize HAP

films is well documented.[10, 11, 13, 14, 19] Figure 3.1a shows the atomic environment around

the OH- ion chains, which are parallel to the c-axis and contain Ca2+ triangles surrounding them.

An unpolarized HAP sample will have H+ randomly occupy the upper or lower site of O2- to form

the OH- group, as shown in Figure 3.1b.[11] With a high enough electric field (> 1 kV/cm) at

elevated temperatures ( > 211oC), H+ begins to rotate to either the upper or lower site on O2-,

aligning in the direction of the electric field as shown in Figure 3.1c.[11] Eventually, H+ ions can

migrate to a neighboring O2- that has a H+ vacancy, thus allowing for long range H+ conduction,

43

as shown in Figure 3.1d. If the electric field is applied until the sample returns to room

temperature, the polarization will remain fixed since H+ is not mobile at room temperature.

Figure 3.1. a) Illustration of OH- aligned along the crystallographic c-axis surrounded by Ca2+

triangles, b) H+ randomly positioned on the top or bottom of O2-, c) H+ all occupying the bottom position on O2- due to an applied electric field at elevated temperatures, and d) H+ migration to

a neighboring O2- with an H+ vacancy. (Adapted from Nakamura et al.)[11]

To characterize the amount of polarization stored within HAP, a thermally stimulated

depolarization current (TSDC) test is used. The polarized HAP is heated to a high enough

temperature until the H+ ions redistribute to evenly occupy O2- vacancies in the crystal structure.

The relaxation of polarization via the movement of H+ can be measured as a current, and the

44

total stored charge can be calculated by integrating the area under the current curve as a

function of temperature.

This section will present TSDC results of yttrium-fluoride co-doped hydroxyapatite

(YFHAP) membranes prepared using a two-step process. First, HAP was seeded onto titanium

substrates by either electrochemical deposition or drying a suspension of HAP nanocrystals onto

the substrate. Then, the HAP seeded substrates were subjected to a hydrothermal growth step

in which they were also co-doped with Y3+ and F-. YFHAP was chosen because previous work has

shown that Y3+ and F- in the hydrothermal solution can produce dense films that have

preferential orientation along the crystallographic c-axis. Second, Y3+ doped HAP has been

shown to produce extra O2- vacancies in the crystal structure in order to maintain charge

neutrality, producing HAP with the chemical formula Ca10‑XYX(PO4)6(OH)2‑XOX.[20, 21] These

extra O2- vacancies have been shown to improve H+ conductivity in HAP.[22] The TSDC results of

this section are unique since the samples were shown to have a very large stored charge

without any polarization step. Currently, there are no known HAP samples that have a

measurable stored charge using TSDC analysis without previous polarization by an external

electric field. Second, not only do the unpolarized samples contain a measured stored charge,

but the value is around one order of magnitude higher than the highest stored charge of 4.3

mC/cm2 recorded for polarized carbonated HAP.[13] The large stored charge is speculated to

come from a composition gradient along YFHAP’s crystal length.

3.2. Experimental

Materials

Titanium(Ti) substrates (12.5 mm x 12.5 mm and 0.89 mm thick), platinum foil (25 mm x

25 mm and 0.127 mm thick), analytical grade Ca(NO3)2·4H2O (99.0% purity), and Y(NO3)3·6H2O

45

(99.9% purity) were purchased from Alfa Aesar. (NH4)2HPO4 (>99.0% purity) was purchased

from EMD. K2HPO4 (99.99% purity), CaCl2·2H2O (99+% purity), NaCl (≥99.0% purity),

tris(hydroxymethyl)- aminomethane (tris) (99.8+% purity), and disodium

ethylenediaminetetraacetate dihydrate (Na2EDTA·2H2O) (ACS reagent, 99.0-101.0% purity) were

all obtained from Sigma- Aldrich. NH4F (ACS reagent, 98% purity) was purchased from Aldrich

Chemical Company, Inc. 36.5-38.0% hydrochloric acid and 28.0-30.0% pure ammonium

hydroxide were purchased from Mallinckrodt Chemicals. Detergent powder was purchased

from Alconox. Plasma sprayed HAP samples were provided by APS Materials, Inc, Dayton, OH.

Nanohydroxyapatite (0.4 wt%) dispersed in water was provided by Transparent Materials,

Rochester, NY. Deionized water was used for all solutions.

Electrochemical deposition of HAP seeds (EdepSeed)

Titanium substrates were gently polished with SiC paper (800 grit) to provide surface

roughness. The substrates were then thoroughly washed with Alconox detergent powder,

rinsed with tap water, attached to a silver wire by tightly wrapping the wire through a premade

hole in the substrate, sonicated in an ethanol/acetone (volume ratio = 50/50) solvent for 30

minutes, and then rinsed with deionized water. An electrolyte solution was prepared containing

138 mM NaCl, 50 mM tris, 1.25 mM CaCl2, and 0.828 mM K2HPO4 in 125 mL deionized water per

substrate. The solution was buffered to pH 7.2 using hydrochloric acid. The electrolyte solution

was then heated to 95°C. The substrate and a platinum plate were held parallel to each other

with a fixed distance of separation of 10 mm. The substrate and platinum plate were connected

to the negative and positive outlets of a direct current power supply (Instek GPS-3030D),

respectively, and immersed in the electrolyte solution. The electrochemical reaction was carried

46

out for 5 minutes at 12.5 mA/cm2 (area relative to the platinum plate), then rinsed off with

deionized water and dried in air.

Deposition of nanohydroxyapatite seed layer (NanoSeed)

A suspension with 0.4 wt% nanohydroxyapatite crystals dispersed in water was provided

by Transparent Materials, Rochester, NY. 50 µl of solution was deposited onto the titanium

substrate. The solution was allowed to dry in a fume hood for at least three hours.

Hydrothermal growth of YFHAP

The hydrothermal solution was prepared by first stirring 0.115M Na2EDTA·2H2O into 30

mL of deionized water per sample until it was completely dissolved. 0.1 M Ca(NO3)2 was then

added, followed by 0.06 M (NH4)2HPO4, 0.01M Y(NO3)3,and 0.02M NH4F. The solution pH was

raised to 10.0 with ammonium hydroxide and then stirred for about 10 minutes. The electrolyte

solution was then transferred to a Teflon-lined stainless steel acid digestion autoclave (Parr Item

No. 4744, 45 mL internal volume), until filled with approximately 60% solution. The seeded

substrate was submersed in the solution with the seed layer facing down and tilted ~45 degrees

relative to the bottom of the autoclave. The autoclave was heated for 15 hours at 200°C and

autogenous pressure. After heating, the autoclave was cooled to room temperature in a fume

hood. The sample was then taken out, rinsed with deionized water several times, and dried in

air.

Thermally stimulated depolarization current (TSDC) measurement

After the hydrothermal deposition, the back sides of the samples were scraped to

remove excess deposition of apatite in order to expose the titanium. Samples then underwent a

mild heat treatment to 300oC for 1 hour with a heating rate of 5oC/minute in order to remove

47

any surface contamination. Then, ~10 nm of platinum was sputtered on the samples’ surface to

improve the surface conductivity. Samples were then tested for their stored charge using the

thermally stimulated depolarization current (TSDC) test, which consists of measuring the current

produced by a sample as it is being heated. Samples were placed between two 1 cm x 1 cm

pieces of platinum foil and were held together using a C-clamp. The platinum foil was

connected to platinum wire that was shielded with alumina tubes. The platinum wires were

then attached to a Keithley 6487 picoammeter to measure current while the samples were

heated to 700oC at a rate of 5oC/min.

Polarization

A sample that was prepared with an electrochemically deposited seed layer and then

hydrothermally grown was placed in the same setup as the TSDC measurement. The sample

was heated to 400oC at a rate of 5oC/minute, and then held there for several hours until the

current output was relatively stable, which turned out to be about 67 nA/cm2. Then, new leads

were attached so that the picoammeter could apply a voltage of 30V for 1 hour to the sample.

After 1 hour, the current generated by the sample was measured using the picoammeter.

Sample characterization



determined by EDX (EDAX). Three spots on two different samples were probed at an

accelerating voltage of 15 kV and the values were averaged and standard deviations were

calculated. The crystal structure was determined by X-ray diffraction (XRD) (Philips PW3020)

with Cu K1 radiation (λ=1.540560 Ǻ) from 20-60o with a step rate of 0.02 degrees/second. Phase

identification was made by comparison with the JCPDS files. For zeta potential measurements,

48

approximately 1 mg of HAP sample was scrapped off and placed into 1 mL of DI water. A 90Plus

Particle Size Analyzer by Brookhaven Instruments Corporation was used. 10 runs, with 5

cycles/run were used to measure the zeta potential. The Smoluchowski model was used for

calculating zeta potential values.


Top and side view SEM images of the two seed layers are shown in Figure 3.2. The seed

layer prepared by electrodeposition (EdepSeed) has a film of individual crystal rods pointing

straight up from the titanium substrate. The individual crystals are just under 1 µm long and

about 20-30 nm wide. The seed layer prepared with the nanohydroxyapatite suspension

(NanoSeed) is a very dense film and is approximately 0.5 µm thick.

Figure 3.2. SEM of seed layers: a) EdepSeed top view, b) EdepSeed side view, c) NanoSeed top

view, and d) NanoSeed side view.

49

Figure 3.3 shows SEM images of the seed layers after the hydrothermal growth. The

EdepSeed layer after hydrothermal growth (EdepYFHAP) is a very dense film of crystals that are

tightly packed. The side view shows that densely packed single crystals make up the film

thickness. Similar results occur for the NanoSeed layer after a hydrothermal growth

(NanoYFHAP), but the crystals are not quite as dense and the film is slightly thinner. Figure 3.3

also compares the HAP deposited via plasma spraying (PlasmaHAP), which was provided by APS

Materials, Inc, Dayton, OH. PlasmaHAP mostly consists of closely packed spherical HAP particles

that form a film roughly 30 µm thick.

50

Figure 3.3. SEM of YFHAP samples after hydrothermal growth and PlasmaHAP: a) EdepYFHAP

top view, b) EdepYFHAP side view, c) NanoYFHAP top view, d) NanoYFHAP side view, e) PlasmaHAP top view, and f) PlasmaHAP side view.

Figure 3.4 shows the XRD results for all the samples. The seed layer prepared by

electrodchemical deposition (EdepSeed) has some preferential orientation along the c-axis, as

demonstrated by the fact that the (0 0 2) peak is slightly higher than the next highest peak of (2

51

1 1). However, after the EdepSeed sample undergoes a hydrothermal growth, the (0 0 2) peak

becomes highly pronounced, being orders of magnitude higher than the next highest peak. A

similar phenomenon occurs for the NanoSeed sample. NanoSeed actually has very little

preferred orientation, but after a hydrothermal growth, the crystals become highly oriented

along the c-axis. These results show that the hydrothermal crystal growth of both seed layers

occurs according to the evolutionary selection mechanism proposed by van der Drift, as

demonstrated in Figure 3.5.[23] Initially, the seed layer crystals grow in a variety of directions,

as shown at t = to. After time t = t1, the crystals begin to intersect each other. When a crystal

that is angled relative to the substrate touches a crystal oriented perpendicular to the substrate,

the angled crystal stops growing. By t = t2, the oriented crystal is the only one to continue

growing. Therefore, only crystals that grow perpendicular to the substrate survive, while other

crystals become embedded under the growth layer.[24]

Figure 3.4. XRD patterns of all samples prepared, with peaks labeled according to JCPDS 09-0432

for HAP and JCPDS 01-1197 for Ti.

52

Figure 3.5. Preferentially oriented crystal growth via an “evolutionary selection” mechanism, as

proposed by van der Drift. (Adapted from Gang et al.)[24]

The thermally stimulated depolarization current (TSDC) results in Figure 3.6 show very

large current densities obtained upon heating both the NanoYFHAP and EdepYFHAP samples,

with measurable current beginning at 250 and 270oC, respectively. The NanoYFHAP sample has

dual peaks around 400 and 530oC, with both peaks having maximum current densities around

16,500 nA/cm2. EdepYFHAP on the other hand, has a single peak at 440oC with a maximum

current density of 25,300 nA/cm2. The total stored charge can be calculated by integrating the

area under the curve as a function of temperature:

( )dTTJ1

Q ∫β= (3.1)

where Q is the stored charge (C/cm2), β is the heating rate (oC/minute) and J(T) is the current

density (A/cm2) at temperature T (oC). EdepYFHAP and NanoYFHAP have stored charges of

29.31 mC/cm2 and 47.81 mC/cm2, respectively. The fact that samples prepared with different

seed layers contained large stored charges demonstrates that the stored charge does not come

from a particular seeding step. To verify that the stored charge is not due to the titanium

substrate or stray currents, the TSDC curve for HAP deposited via plasma spraying shows that

almost no current is produced for the unaligned sample.

53

Figure 3.6. TSDC curve of EdepYFHAP, NanoYFHAP, and PlasmaHAP.

One factor that attributes to the sample’s high stored charge is the fact that HAP can

exist in various compositions ranging from stoichiometric HAP, Ca10(PO4)6(OH)2, to calcium-

deficient HAP as expressed by Ca(10-x-y)(HPO4)x(PO4)6-x(OH)2-yy. As the chemical formula shows,

the creation of one Ca2+ vacancy can either cause the creation of a OH- vacancy, , or the

protonation of a PO43- group in order to maintain charge neutrality. This ability to exist in a

variety of compositions allows HAP films to potentially be formed with a composition

concentration gradient.[25, 26] The hydrothermally grown samples in this work were capable of

forming a composition concentration gradient because of the seed layers being negatively

charged, as verified by zeta-potential measurements. EdepSeed has a zeta-potential of -15.60

mV ± 0.51, while NanoSeed’s is -15.01 mV ± - 1.28. Negative zeta-potential values for HAP are

common due to speculation that HAP is thought to have an excess concentration of PO43- groups

54

at its surface.[25-27] When the negative seed layers are placed into the hydrothermal solution,

they first interact with the positive Ca2+ ions, forming a Ca2+ rich apatite layer, as depicted in

Figure 3.7. After the Ca2+ rich layer forms, it will interact with the negative phosphate ions in

solution, forming a Ca2+ deficient apatite layer. Eventually, the HAP crystals reach an

equilibrium composition.[25, 26] A potential profile of the composition change within the

hydrothermally grown YFHAP crystals is shown in Figure 3.8. The variation of HPO42- along the

hydrothermally grown YFHAP crystal length provides a change of H+ concentration within the

crystal structure. Upon heating these samples during the TSDC test, the H+ from HPO42- become

mobile and are redistributed to O2- vacancies to form OH- groups. In fact, it has been

demonstrated that for nonstoichiometric HAP, H+ generated from HPO42- predominantly migrate

rather than H+ from OH- below 400oC.[9, 28] This H+ migration is the current being measured as

the YFHAP films are being heated during the TSDC test. Therefore, the hydrothermally grown

YFHAP samples have a built in stored charge due to the variation of composition along their

crystal length. Another piece of evidence that demonstrates there is a concentration gradient of

H+ within the YFHAP crystals is that if the electrodes of the picoammeter are reversed during

TSDC testing, the sign of the current always becomes negative. This demonstrates that the

concentration gradient of H+ always occurs in the same direction of the YFHAP crystals.

55

Figure 3.7. Schematic of Ca2+ attraction to the negatively charged surface of a seed layer;

initiating the calcium-rich base of a hydrothermally grown crystal.

Figure 3.8. Potential composition concentration profile of a YFHAP crystal grown during

hydrothermal synthesis.

Another factor that allows these stored charge values to be so large is due to the high c-

axis orientation of the YFHAP films. H+ migration has been shown to occur along OH- ions, which

are aligned along the crystallographic c-axis.[9, 11, 22, 29] As confirmed by XRD in Figure 3.4,

56

the YFHAP films are almost perfectly aligned along the c-axis, allowing for easier migration of H+.

Previous work in our group has shown enhanced conductivity of HAP and yttrium-doped HAP

membranes with high c-axis orientation.[20, 30] Another factor that improves to mobility of H+

in the YFHAP samples is the fact that the individual crystals contain very few grain boundaries,

which have been shown to act as traps for H+ conduction.[9]

In order to demonstrate the ability of these coatings to be polarized by an electric field,

the samples first had to be depolarized as much as possible since the intrinsic stored charge in

the films was so large that it overshadows the influence of polarizing the films. Therefore, an

EdepYFHAP sample was first heated to 400oC, with Figure 3.9 showing the resulting TSDC curve.

As expected, a large stored charge can be measured in the unpolarized sample. To depolarize

the sample, it was held at 400oC for several hours until the current measured was significantly

lowered and stayed at a relatively stable value, as shown in Figure 3.10. The inset in Figure 3.10

shows that after being held at 400oC for over 10 hours, the current measured reached a

somewhat stable value of around 67 nA/cm2.

57

Figure 3.9. TSDC curve of unpolarized EdepYFHAP sample heated to 400oC.

Figure 3.10. TSDC curve depolarizing the EdepYFHAP sample by keeping it at 400oC for several

hours. Inset shows a zoomed in shot at the end of the TSDC curve.

58

At this point, the sample was then polarized by an external electric field of 30V for 1

hour. After polarization, a TSDC curve was obtained, as shown in Figure 3.11. Clearly, the

current measured increases above 67 nA/cm2 after polarization. To calculate the total stored

charge induced by polarization, the area under the curve of Figure 3.11 was subtracted from a

straight line at 67 nA/cm2, as shown in equation 3.2, where t stands for time in seconds:

∫ ∫t

=

0

t

0

2dt

cm

nA 67-J(t)dtQ (3.2)

The total stored charge induced by polarization in Figure 3.11 was therefore measured to be

around 0.05 mC/cm2. This value underestimates the stored charge induced by polarization due

to the fact that the current was not measured until the sample reached the original current of

67 nA/cm2. Regardless, this test does validate the ability to induce a polarization within the

films using an external electric field.

Figure 3.11. TSDC curve after polarizing the EdepYFHAP sample at 400oC and 30V for 1 hour.

59

3.4. Conclusions

Two different HAP seed layers were hydrothermally grown to produce YFHAP films that

contained large stored charge values according to TSDC experiments. The fact that two

different seed layer processes were used shows that the stored charge is not the result of a

particular seeding process. Also, the fact that plasma deposited HAP does not have any stored

charge verifies that the large currents were not from the titanium substrate or stray currents

from the setup. Rather, the stored charge is built into the YFHAP crystals due to a composition

concentration gradient along the crystal length. Due to variations of HPO42- in the crystal

structure, there is a variation of H+ throughout the sample. Upon heating, the H+ becomes

redistributed by migrating to O2- vacancies and forming OH-. The fact that the YFHAP crystals

are highly c-axis oriented and contain few grain boundaries along their crystal length facilitates

the redistribution of H+, leading to the large stored charge values.

60

References

1. Zakaria, S., et al., Nanophase hydroxyapatite as a biomaterial in advanced hard tissue engineering: A review. Tissue Eng. Part B Rev., 2013.

2. Palazzo, B., et al., Biomimetic hydroxyapatite–drug nanocrystals as potential bone substitutes with antitumor drug delivery properties. Advanced Functional Materials, 2007.

17(2180-2188).

3. Schnieders, J., et al., Controlled release of gentamicin from calcium phosphate-

poly(lactic acid-co-glycolic acid) composite bone cement. Biomaterials, 2006. 27: p. 4239-4249.

4. Kar, A., K. Raja, and M. Misra, Electrodeposition of hydroxyapatite onto nanotubular TiO2 for implant applications. Surface and Coatings Technology, 2006. 2001: p. 3723-3731.

5. Yang, Y., k. Kim, and J. Ong, A review on calcium phosphate coatings produced using a

sputtering process—an alternative to plasma spraying. Biomaterials, 2005. 26: p. 327-337.

6. Stigter, M., K.d. Groot, and P. Layrolle, Incorporation of tobramycin into biomimetic

hydroxyapatite coating on titanium. Biomaterials, 2002. 23: p. 4143-4153.

7. Kuo, M. and S. Yen, The process of electrochemical deposited hydroxyapatite coatings on biomedical titanium at room temperature. Materials Science and Engineering C, 2002. 20: p.

153-160.

8. Koutsopoulos, S., Synthesis and characterization of hydroxyapatite crystals: A review study on the analytical methods. Journal of Biomedical Materials Research, 2002. 62: p. 600-612.



10. Kumar, D., et al., Polarization of hydroxyapatite: Influence on osteoblast cell

proliferation. Acta Biomaterialia, 2010. 6: p. 1549-1554.



12. Yamashita, K., N. Oikawa, and T. Umegaki, Acceleration and deceleration of bone-like crystal growth on ceramic hydroxyapatite by electric poling. Chem. Mater., 1996. 8: p. 2697-

2700.

13. Nagai, A., et al., Electric polarization and mechanism of B-type carbonated apatite

ceramics. Journal of Biomedical Materials Research Part A, 2011. 99A: p. 116-124.

61

14. Bodhak, S., S. Bose, and A. Bandyopadhyay, Role of surface charge and wettability on early stage mineralization and bone cell–materials interactions of polarized hydroxyapatite. Acta Biomaterialia, 2009. 5: p. 2178-2188.

15. Itoh, S., et al., Enhanced bone ingrowth into hydroxyapatite with interconnected pores

by Electrical Polarization. Biomaterials, 2006. 27: p. 5572-5579.

16. Ueshima, M., S. Nakamura, and K. Yamashita, Huge, millicoulomb charge storage in ceramic hydroxyapatite by bimodal electric polarization. Advanced Materials, 2002. 14: p. 591-595.

17. Kobayashi, T., S. Nakamura, and K. Yamashita, Enhanced osteobonding by negative surface charges of electrically polarized hydroxyapatite. Journal of Biomedical Materials

Research Part A, 2001. 57: p. 477-484.

18. Ohgaki, M., et al., Manipulation of selective cell adhesion and growth by surface charges of electrically polarized hydroxyapatite. Journal of Biomedical Materials Research Part A, 2001. 57: p. 366-373.

19. Wolf-Brandstetter, C., et al., The impact of heat treatment on interactions of contact-poled biphasic calcium phosphates with proteins and cells. Acta Biomaterialia, 2012. 8: p. 3468-

3477.

20. Wei, X. and M.Z. Yates, Yttrium-doped hydroxyapatite membranes with high proton conductivity. Chemistry of Materials, 2012. 24: p. 1738-1743.

21. Owada, H., K. Yamashita, and T. Kanazawa, High-temperature stability of hydroxyl ions in yttrium-substituted oxyhydroxyapatites. Journal of Materials Science Letters, 1990. 9: p. 26-

28.

22. Yamashita, K., et al., Protonic conduction in yttrium-substituted hydroxyapatite ceramics and their applicability to H2-O2 fuel cell. Solid State Ionics, 1990. 40/41: p. 918-921.

23. Drift, A.v.d., Evolutionary selection, a principle governing growth orientation in vapor-

deposited layers. Phil. Res. Rep., 1967. 22: p. 267-288.

24. Gang, L., et al., Growth mechanism of a preferentially oriented mordenite membrane.

Journal of Zhejiang University SCIENCE, 2005. 5: p. 369-372.

25. Kim, H., et al., Process and kinetics of bonelike apatite formation on sintered hydroxyapatite in a simulated body fluid. Biomaterials, 2005. 26: p. 4366-4373.

62

26. Brown, P. and R. Martin, An analysis of hydroxyapatite surface layer formation. Journal

of Physical Chemistry B., 1999. 103: p. 1671-1675.

27. Vandiver, J., et al., Nanoscale variation in surface charge of synthetic hydroxyapatite detected by chemically and spatially specific high-resolution force spectroscopy. Biomaterials,

2005. 26(271-283).


Materials Science and Engineering, 2009. 161: p. 115-119.

29. Yamashita, K., K. Kitagaki, and T. Umegaki, Thermal Instability and Proton Conductivity of Ceramic Hydroxyapatite at High Temperatures. J. Am. Ceram. Soc., 1995. 78: p. 1191-97.

30. Liu, D., K. Savino, and M.Z. Yates, Microstructural engineering of hydroxyapatite membranes to enhance proton conductivity. Advanced Functional Materials, 2009. 19: p. 3941-

3947.

63

Chapter 4

Characterization of Ytterbium and Ytterbium-fluoride Doped Hydroxyapatite Membranes for

Ion Conducting Applications

4.1. Introduction

While hydroxyapatite (HAP, Ca10(PO4)6(OH)2) has been highly studied as a bioceramic

due to its biocompatibility and osteoconductive properties, recent research has focused on

using HAP in new applications such as catalyst supports, gas sensors, ionic exchangers, electrets,

and fuel cell electrolytes due to its ionic conducting abilities.[1-8] HAP belongs to the hexagonal

P63/m space group, with OH- ions forming one-dimensional chains parallel to the c-axis in the

center of Ca2+ triangles.[9] HAP has been shown to be a proton conductor at elevated

temperatures, with the conduction mechanism being proposed to be protons rotating along the

c-axis of the OH- groups.[10] Thermally stimulated depolarization current (TSDC) measurements

indicate that the onset of proton mobility can be as low as 200oC.[4] With proton mobility

occurring at this low of a temperature, HAP has the potential to be an electrolyte for an

intermediate temperature (400-700oC) fuel cell. Intermediate temperature fuel cells have

advantages over traditional low temperature (T < 100oC) fuel cells such as polymer electrolyte

membranes because they do not require expensive rare metal catalysts for the electrodes. High

temperature (T > 800oC) fuel cell electrolytes such as yttria-stabilized zirconia (YSZ) are

problematic because they often require expensive high temperature resistant auxiliary

components, plus the high temperatures often cause mechanical issues such as cracking over

the long term.[11] While intermediate temperatures are ideal for a fuel cell, the ionic

64

conductivity of current intermediate temperature fuel cell electrolytes is too low or they are

chemically unstable for practical applications.[12] While HAP can conduct protons in the

intermediate temperature range, the proton conductivity of a typical HAP pellet sintered in air is

much lower (10-6 S/cm) than the O2- conductivity of current state of the art high temperature

electrolytes such as YSZ (10-2 S/cm at 700oC).[13, 14]

To further enhance HAP’s ionic conductivity, our group has developed a novel method

that deposits c-axis oriented HAP membranes onto metal substrates.[3, 15] The first step is an

electrochemical deposition that seeds the metal surface with submicron HAP crystals. The seed

layer acts as a nucleation site for crystal growth during the second step, which is a hydrothermal

process that increases the size, density, and c-axis orientation of the HAP membrane. The fact

that the crystals are grown preferentially along the c-axis and the individual crystals have very

few grain boundaries means that the protons have a direct path for conduction. These

properties produce proton conductivity values of ~10-4 S/cm at 700oC on palladium

substrates.[3] Not only is the conductivity improved, but the area specific resistance is greatly

reduced compared to sintered pellets of hydroxyapatite because of the 2-3 orders of magnitude

reduction in the HAP membrane thickness. Even with these improvements, the conductivity is

still too low for practical applications as an intermediate temperature fuel cell electrolyte.

Another method to enhance conductivity is to partially dope the HAP crystals with

various cations or anions. Ca2+ can be substituted with monovalent, bivalent, or trivalent cations

such as Na+, Mg2+, and Y3+ respectively, PO43- is typically replaced with a bivalent anion such as

CO32-, and OH- has been exchanged with monovalent anions such as F-.[1, 16-18] When yttrium

is substituted, an apatite with the chemical formula Ca10-xYx(PO4)6(OH)2-xOx is created.[6, 7, 19-

21] Y3+ substitution for Ca2+ converts some OH- groups to O2- in order to maintain charge

65

neutrality. It was shown that as yttrium substitution increases up to x = 0.65, conductivity

increases to a maximum value of ~10-4 S/cm at 800oC. However, when x > 0.65, the apatite

transitions from a pure proton conductor, to a mixed proton/O2- ion conductor, and eventually

to a pure O2- ion conductor, with the conductivity falling to ~10-7 S/cm at 800oC. Previous work

in our group has shown that with yttrium substitution, conductivity values approaching 10-2

S/cm at 700oC can be obtained.[22] Carbonate substitution has also been shown to improve

conductivity, giving values approaching 10-3 S/cm at 700oC.[1] As with yttrium substitution, a

similar conversion to O2- ion conduction is observed with increased carbonate substitution.

One drawback to using hydroxyapatite for high temperature ion conducting applications

is their potential for dehydroxylation. Nonstoichiometric hydroxyapatite can begin

dehydroxylation as low as 500oC.[13] Since the proton conduction mechanism is dependent on

OH- ions, conductivity decreases as dehydroxylation occurs. Conductivity values can decrease by

103 S/cm after only 24 hours for HAP pellets and films that have become dehydroxylated.[13]

To prevent dehydroxylation, fluoride substitution has been shown to improve the thermal

stability of HAP.[23] Proton conduction was even shown to increase with the addition of up to

50% F- substitution, but decreased with higher levels of substitution due to the strong

electronegativity of F- and their affinity for protons.[8]

While a variety of elements have been doped into the HAP crystal structure to influence

biological, mechanical, and ionic properties, very little is known about how elements from the

lanthanide series impact the crystal growth and structure of hydroxyapatite membranes.

Doping with elements from the lanthanide series may be of interest since they are trivalent

cations. As mentioned with yttrium, trivalent cations can convert OH- to O2-, and potentially

improve HAP’s conductivity. Therefore, preliminary experiments in this work were done to

66

introduce various elements from the lanthanide series into the HAP crystal structure using a

novel electrochemical-hydrothermal technique. Since it was desired to test conductivity and

fuel cell performance measurements on the doped HAP films, it was necessary to produce highly

dense films in order to prevent hydrogen gas crossover during measurements. It was found that

ytterbium (Yb3+) doped HAP (YbHAP) was the most suitable dopant to form a fully dense and

smooth membrane. The effect of Yb3+ concentration on the crystal morphology, crystal

structure, and composition was further evaluated. Also, a co-doped ytterbium-fluoride HAP

(YbFHAP) sample was synthesized and characterized since fluorinated HAP has been shown to

have enhanced thermal stability at elevated temperatures, which is critical during conductivity

and fuel cell testing.

To test the conductivity and fuel cell performance of the YbHAP and YbFHAP

membranes, it was necessary to deposit the coating on a substrate that could act as a support

structure for the membrane, while also being the anode for hydrogen oxidation. Previous work

has shown that palladium can serve as a hydrogen membrane at elevated temperatures.[12, 24]

When exposed to palladium at temperatures above 300oC, hydrogen will dissociate into protons

and electrons, with the protons traveling through the palladium. In this work, a thin palladium

film (~10-20 µm) was electrolessly deposited onto one side of a porous stainless steel

substrate.[25] The porous stainless steel provides mechanical support for the thin palladium

film, as compared to using palladium foil that could support itself (80-100 µm). The decreased

thickness of palladium reduces the resistance of protons traveling through the palladium

membrane. The reduced thickness also minimizes the amount of palladium used, which lowers

cost. After the YbHAP and YbFHAP electrolyte was then deposited onto the palladium anode, a

cathode catalyst paste of La0.6Sr0.4Co0.2Fe0.8O3-x (LSCF) (~40 µm) was brush coated on top of the

67

YbHAP or YbFHAP membrane to perform oxygen reduction. This entire assembly is called a

“hydrogen membrane fuel cell”, as represented in Figure 4.1.[12, 24]

Figure 4.1. Hydrogen membrane fuel cell, with palladium anode, YbHAP or YbFHAP electrolyte,

and LSCF cathode.

4.2. Experimental

Materials


25 mm and 0.127 mm thick), analytical grade Ca(NO3)2·4H2O (99.0% purity), La(NO3)3·6H2O

(99.9% purity), Eu(NO3)3·6H2O (99.9% purity), Ce(NO3)3·6H2O (99.5% purity), Nd(NO3)3·6H2O

(99.9% purity), and Yb(NO3)3·6H2O (99.9% purity) were purchased from Alfa Aesar. (NH4)2HPO4

(>99.0% purity) was purchased from EMD. K2HPO4 (99.99% purity), CaCl2·2H2O (99+% purity),

NaCl (≥99.0% purity), tris(hydroxymethyl)- aminomethane (tris) (99.8+% purity), analytical

hydroxyapatite powder, and disodium ethylenediaminetetraacetate dihydrate (Na2EDTA·2H2O)

(ACS reagent, 99.0-101.0% purity) were all obtained from Sigma- Aldrich. NH4F (ACS reagent,

68

98% purity) was purchased from Aldrich Chemical Company, Inc. La0.6Sr0.4Co0.2Fe0.8O3-x (LSCF)

paste was obtained from NexTech Materials. Platinum paint came from SPI Supplies. 36.5-

38.0% hydrochloric acid and 28.0-30.0% pure ammonium hydroxide were purchased from

Mallinckrodt Chemicals. Detergent powder was purchased from Alconox. Sintered porous

stainless steel discs (0.5 µm pore size, 19.05 mm diameter, 1.59 mm thick) were obtained from

Mott Corporation. Deionized water was used for all solutions.

Preparation of palladium on porous stainless steel (PSS)

Samples prepared for conductivity and fuel cell testing were deposited onto a PSS

substrate that was coated with a thin layer of palladium using a reported electroless deposition

process.[25] The PSS substrate was first cleaned with an alkaline solution that consisted of of 5

g Na3PO4, 16.25 g Na2CO3, 11.26 g NaOH, and 1.25 g Alconox powder in 250 mL of DI water. The

substrate was sonicated for 1h in a water bath at 60oC. Then the substrates were removed,

rinsed off with tap water, and sonicated in DI water for another hour. The sample was then

sonicated for 15 minutes in isopropanol and allowed to dry by placing samples in an oven for 30

minutes at 140oC. Finally, the substrate was placed in 1M HCl for 10 minutes, followed by

sonicating in DI water for 10 minutes.

Two beakers were then prepared for the surface activation. Beaker one contained 1 g/L

SnCl2*2H20 and 1 ml/L concentrated HCl in DI water, while beaker two contained 0.1 g PdCl2 and

1 ml concentrated HCl in 100 ml DI water. Both solutions were stirred until the chemicals fully

dissolved. Then the cleaned PSS substrates were immersed in beaker one for 5 minutes at room

temperature, rinsed with DI water, then immersed in beaker two for 4 minutes at room

temperature. The samples were rinsed with 0.01M HCl, then rinsed with DI water, and placed

back in beaker one for 5 minutes. This process was repeated five more times until a uniform,

69

brownish gray surface appeared. The samples were then dried for 2 hour at 140oC. One of the

activated sides was lightly scraped off using a box cutter blade since the palladium only needs to

be deposited on one side.

A plating solution was then prepared with 4 g/L Pd(NH3)4Cl2*H2O, 198 ml/L NH4OH

(28%), 40.1 g/L Na2EDTA, and 6 ml/L of 1M H2NNH2, giving a pH ~7.4. Samples were attached

with silver wire through the hole in the substrate so that the sides with the scraped off

activation layer were touching. The samples were placed in a volume of plating solution

according to the ratio Vsolution/Splating area ~3.5 cm3/cm2 for 90-120 minutes at 60oC, then rinsed

with 60oC water. The substrates were then placed in fresh plating solution for another 90-120

minutes. This process was repeated (usually 5-6 times) until a dense palladium layer of about 20

µm was formed.

The substrates were then annealed at a temperature of 640oC for 2 hours using a ramp

rate of 5oC/minute. Hydrogen was applied when the sample reached 300oC at a rate of

200cc/minute, then turned off before the sample cooled down to 300oC to avoid palladium

embrittlement.


Samples for XRD and EDX characterization were deposited on titanium substrates, while

samples prepared for conductivity and fuel cell performance were deposited onto PSS-Pd. PSS-

Pd and titanium substrates were gently polished with SiC paper (800 grit) to provide surface





70





to the negative and positive outlets of a direct current power supply (Instek GPS-3030D),




Hydrothermal growth

The hydrothermal solution was prepared by first stirring Na2EDTA·2H2O into 30 mL of

deionized water per sample until it was completely dissolved. The concentration of

Na2EDTA·2H2O was dependent on the concentrations of Ca2+ and Yb3+ (or La3+, Ce3+, Nd3+, and

Eu3+) in the solution according to the equation [EDTA] = [Ca2+] + 1.5[Yb3+]. 0.1 M Ca(NO3)2·4H2O

was then added, followed by 0.06 M (NH4)2HPO4. For La3+, Eu3+, Ce3+, and Nd3+ samples, 0.01M

La(NO3)3·6H2O, Eu(NO3)3·6H2O, Ce(NO3)3·6H2O, and Nd(NO3)3·6H2O of was added to the

hydrothermal solution, respectively. For Yb3+ doped HAP, Yb(NO3)3·6H2O was added in amounts

ranging from 0.05M to 0.03M (YbHAP(0.005M) to YbHAP(0.03M)) for composition and XRD

analysis. A co-doped ytterbium-fluoride sample (YbFHAP) was also prepared, which contained

0.01M Yb(NO3)3·6H2O and 0.02M NH4F. A fluorinated sample (FHAP) was prepared with the

same composition as YbFHAP, except without the ytterbium. A YbFHAP sample was prepared

on PSS-Pd for conductivity and fuel cell performance tests, as well as a YbHAP(0.01M) sample.

The pH of each solution was raised to 10.0 with ammonium hydroxide and then stirred for about

71

10 minutes. The electrolyte solution was then transferred to a Teflon-lined stainless steel acid

digestion autoclave (Parr Item No. 4744, 45 mL internal volume), until filled with approximately

60% solution. The seeded substrate was submersed in the solution with the seed layer facing

down and tilted ~45 degrees relative to the bottom of the autoclave. The autoclave was heated

for 15 hours at 200°C and autogenous pressure. After heating, the autoclave was cooled to

room temperature in a fume hood. The sample was then taken out, rinsed with deionized water

several times, and dried in air. For samples with multiple hydrothermal growths, a new

hydrothermal solution was prepared and the hydrothermal process was repeated for a specified

number of times.

Membrane characterization



determined by EDX (EDAX). Three spots on each sample were probed at an accelerating voltage

of 15 kV and the values were averaged and standard deviations were calculated. The crystal

structure was determined by X-ray diffraction (XRD) (Philips PW3020) with Cu K1 radiation

(λ=1.540560 Ǻ) from 20-40o with a step rate of 0.02 degrees/second. Phase identification was

made by comparison with the JCPDS files.

Fuel cell conductivity and performance test

A 50-100 µm thick LSCF cathode with dimensions of 0.5 x 0.5 cm was brush coated onto

a YbHAP or YbFHAP membrane that was deposited on PSS-Pd. The LSCF was allowed to dry for 1

hour at 80oC. A platinum mesh with platinum wires attached for lead connections was then

pasted to the PSS and LSCF using platinum paint and LSCF paste on each side, respectively,

followed by drying for 1 hour at 80oC. The sample was then attached to the end of a 20 mm

72

diameter alumina tube using a ceramic adhesive (Ceramabond 569, Aremco Products, Inc.), with

the YbHAP or YbFHAP membrane facing outward. The ceramic tube was then attached to a

ProboStat test assembly for hydrogen gas input and removal, as well as connecting to the

Zahner Zennium for electrochemical impedance spectroscopy (EIS) and fuel cell current-voltage

(I-V) measurements. The sample was heated to 400, 500, 600, and 700oC at a ramp rate of

2oC/minute for EIS and I-V measurements. The sample was maintained at each temperature for

15 minutes before measurements in order to allow the sample to reach thermal equilibrium.

Room temperature hydrogen gas was used at a rate of 200 cc/minute for each measurement,

which was controlled by a Scribner Associates 850C test stand. EIS measurements were carried

out using the two-point probe alternating current impedance spectroscopy method over a

frequency range of 1 MHz to 0.1 Hz with an amplitude of 50 mV. I-V measurements were done

in potentiostatic mode at a rate of 5mV/s.


All samples were prepared by first electrochemically depositing a HAP seed layer, as

shown in Figure 2.1 of chapter 2. Figure 4.2 shows samples that were then prepared with

various ions from the lanthanide series, in particular La3+, Ce3+, Nd3+, Eu3+, and Yb3+, after one 15

hour hydrothermal growth. EDX verified the presence of each element in the crystal structure.

While all samples contain a relatively similar crystal structure, Yb3+ doped HAP clearly provides

the most suitable morphology for producing a potentially dense film for conductivity

characterization due to its smooth surface and densely packed crystals. Therefore, Yb3+ was

chosen to investigate and characterize in more detail.

73

Figure 4.2. SEM of a) La3+, b) Ce3+, c) Nd3+, d) Eu3+, and e) Yb3+ doped HAP.

Figure 4.3 shows top view SEM images of samples grown as a function of Yb3+

concentration after one hydrothermal growth. Samples with no Yb3+ (HAP) during the

hydrothermal growth have many gaps between the crystals, while samples with 0.005M, 0.01M,

and 0.015M Yb3+ (YbHAP(0.005M), YbHAP(0.01M), and YbHAP(0.015M)) show crystals with a

74

smooth surface and are densely packed. The YbHAP(0.02M) sample shows that the crystals

begin to have more gaps between them, while the YbHAP(0.03M) sample seems to change from

a hexaganol surface to pyramidal, with many gaps at the surface.

To understand the morphological changes as Yb3+ concentration increases, composition

analysis was done using EDX and the results are shown in Table 4.1. The ratio of Ca2+ to Yb3+

actually incorporated into the crystals (CaEDX/YbEDX) almost matches the ratio of Ca2+ to Yb3+ in

the hydrothermal solution (CaHydrothermal/YbHydrothermal) for YbHAP(0.005M) and YbHAP(0.01M).

For YbHAP(0.015M) and YbHAP(0.02M), even though 50% and 100% more Yb3+ was

incorporated into the hydrothermal solution compared to YbHAP(0.01M), respectively, only

slightly more Yb3+ was incorporated into the crystal structure. However, a dramatic increase of

Yb3+ incorporated into the crystals occurs for YbHAP(0.03M). This dramatic change of

composition correlates to the morphological changes that occur for YbHAP(0.03M) in Figure 4.3

f.

75

Figure 4.3. SEM top view images of crystal layers after one hydrothermal growth as a function of

Yb concentration: a) [Yb3+] = 0.0M, b) [Yb3+] = 0.005M, c) [Yb3+] = 0.01M, d) [Yb3+] = 0.015M, e) [Yb3+] = 0.02M, and f) [Yb3+] = 0.03M.

76

Table 4.1. EDX results of HAP, YbFHAP, and YbHAP samples with various Yb3+ concentrations.

CaHydrothermal/YbHydrothermal is the Ca2+/Yb3+ atomic ratio in the starting hydrothermal solution. CaEDX/YbEDX is the atomic ratio of the synthesized apatite membrane.

To further characterize the samples as a function of Yb3+ concentration, XRD analysis

was done on each sample, as shown in Figure 4.4. The XRD spectrums of analytical grade HAP

powder and the HAP electrochemical seed layer are shown as a comparison. All the peaks for

the analytical HAP powder are consistent with the standard (JCPDS 09-0432), with the two

highest peaks being the (2 1 1) and (1 1 2) planes. The HAP seed layer has the greatest peak

intensity for the (0 0 2) plane, with significant peaks from the (2 1 1) and (1 1 2) planes also. The

titanium (Ti) (0 0 2) peak can be detected easily due to the thin nature of the film (JCPDS 01-

1197). For the HAP sample with no Yb3+ after one hydrothermal growth, a dramatic increase of

the (0 0 2) peak is observed (the peak is actually off the chart and is around three orders of

magnitude higher than the next highest peak). The large (0 0 2) peak means that the crystals

have preferred c-axis orientation normal to the substrate. For YbHAP(0.005M), YbHAP(0.01M),

and YbHAP(0.015M), almost all other peaks besides the (0 0 2) plane become indistinguishable,

representing almost perfect c-axis orientation. For YbHAP(0.02M), secondary phases begin to

form, in particular Na3Yb2(PO4)3 (JCPDS card 23-0534) and Ca3Yb2O6 (JCPDS card 19-0263). The

crystals also have an enhanced (2 1 1) orientation. For YbHAP(0.03M), the (0 0 2) peak becomes

less intense than the (2 1 1) peak, which shows decreased c-axis orientation. The intensity of

the secondary phase’s peaks also increases. The XRD data is consistent with the morphological

77

and composition changes that begin when 0.02M Yb3+ or more is added. Clearly there is a limit

to how much Yb3+ can be incorporated into the crystal structure before the orientation and

composition deviate greatly from undoped HAP. Lattice parameters were calculated the same

way as in chapter 2. Yb3+ is only slightly smaller than Ca2+, so Yb3+ substitution, as expected, did

not dramatically change the unit cell volume, as shown in Table 4.2.

Figure 4.4. XRD patterns of HAP powder, HAP seed layer, and HAP, YbHAP, and YbFHAP after

one hydrothermal growth on titanium: (a) Analytical grade HAP powder, (b) HAP seed layer, (c) YbFHAP with [Yb] = 0.01M and [F] = 0.02M, (d) HAP, (e) YbHAP(0.005M), (f) YbHAP(0.01M), (g)

YbHAP (0.02M), and (h) YbHAP(0.03M). XRD peaks are identified from the hydroxyapatite standard pattern (JCPDS 09-0432), titanium (Ti) is identified with (JCPDS 01-1197), while

secondary phases Na3Yb2(PO4)3 and Ca3Yb2O6 are determined from (JCPDS 23-0534) and (JCPDS 19-0263), respectively.

Table 4.2. Lattice parameter calculations.

Previous studies using HAP as a liquid chromatography packing have shown that HAP

has crystal c-surfaces (the crystal facets parallel to the titanium substrate) that are negatively

78

charged, while its a-surface crystal facets are positively charged.[26] Since the positively

charged Yb3+ will most likely preferentially adsorb onto the negatively charged c-surface, the

crystal growth onto that surface will slow due to the reduced interfacial free energy of that

facet.[27] Therefore, it becomes more favorable for the crystal to grow wider. With too much

Yb3+ doping, the change in crystal morphology may be due to the fact that calcium substitution

can occur at two different sites. Cations such as Yb3+ can substitute at calcium sites called Ca(I)

or Ca(II). There are four Ca(I) sites per unit cell, and they are in columns parallel to the c-axis

surrounded by nine phosphate oxygen atoms. There are six Ca(II) sites per unit cell, which are

surrounded by seven oxygen atoms.[28, 29] It has been shown that under low cation

substitution, the cations will be evenly distributed among the two sites, but at high substitution,

they will preferentially go to the Ca(II) site.[30] This transition of Yb3+ substitution to the Ca(II)

site manifests itself in a morphological and composition change during the hydrothermal

growth.

In order to make a fully dense membrane to provide a gas tight layer for conductivity

and fuel cell testing, the hydrothermal step can be repeated. Figure 4.5 shows a YbHAP(0.01M)

sample after one, two, three, and four hydrothermal growths. By three hyrdrothermal growths,

the sample has very few gaps as viewed from the top. The thickness after the various growths

ranges from ~10-30 µm, as seen in the side view images of Figure 4.5. The side view also

confirms that the crystals are highly packed and have no gaps through the crystal membrane.

As a comparison, Figure 4.6 shows a HAP sample with no Yb3+ after four hydrothermal growths.

The top view shows that many gaps between the crystals still exist; meaning that under the

current growth conditions a dense film cannot be obtained from a pure hydroxyapatite sample.

79

Figure 4.5. Top and side view SEM images of YbHAP(0.01M) after 1, 2, 3, and 4 hydrothermal

growths of 15 hours each: a) 1 growth top view, b) 1 growth side view, c) 2 growths top view, d) 2 growths side view, e) 3 growths top view, f) 3 growths side view, g) 4 growths top view, h) 4

growths side view.

80

Figure 4.6. Top and side view SEM images of HAP after four hydrothermal growths: a) HAP top

view, b) HAP side view.

By comparison, a sample that was co-doped with 0.01M Yb3+ and 0.02M F- (YbFHAP)

creates an extremely smooth and densely packed membrane after only one hydrothermal

growth, as seen in Figures 4.7 a and b. The sample is also very thin, only ~5 µm thick. This

morphology is ideal for a fuel cell membrane. XRD data in Figure 4.4 confirms a highly oriented

crystal structure similar to YbHAP(0.01M). EDX data also shows a similar composition to

YbHAP(0.01M), with the addition of fluoride clearly detected. By comparison, a sample that was

only doped with fluoride (FHAP) after four hydrothermal growths is shown in Figures 4.5 c and

d. Under these growth conditions, a sample with many gaps is formed; thus the film is not

suitable for conductivity and fuel cell testing.

81

Figure 4.7. Top and side view SEM images of ytterbium-fluoride co-doped sample (YbFHAP) with [Yb] = 0.01M and [F] = 0.02M, after 1 hydrothermal growth and a fluoride doped sample (FHAP)

with [F] = 0.02M, after 4 hydrothermal growths: a) YbFHAP top view, b) YbFHAP side view, c) FHAP top view, d) FHAP side view.

Figure 4.8 shows the porous stainless steel (PSS) substrate and the palladium layer that

was electrolessly deposited onto PSS. The PSS is porous enough to allow hydrogen gas to reach

the palladium layer, while the palladium is dense enough to provide a continuous surface to

grow the apatite membrane on. For conductivity and fuel cell testing, YbHAP(0.01M) samples

with three hydrothermal growths were grown onto the palladium, as well as YbFHAP samples

with one hydrothermal growth.

82

Figure 4.8. SEM images of porous stainless steel and palladium (Pd): a) Porous stainless steel,

zoomed in, b) Porous stainless steel, zoomed out, c) Pd on porous stainless steel, zoomed in, d) Pd on porous stainless steel, zoomed out.

Figures 4.9 and 4.10 show Nyquist plots for YbHAP(0.01M)-PSS-Pd-LSCF and YbFHAP-

PSS-Pd-LSCF, respectively, at temperatures from 400 to 700oC. The membrane bulk resistance

was determined using curve-fitting software to determine the point where the arc intersects the

x-axis. It is clear that for both samples, as temperature increases, the bulk resistance decreases.

In Figure 4.11, an Arrhenius plot is created using the equation σ = L/(R*A), where σ is

conductivity (Siemens/cm , S/cm), L is thickness (cm) of the apatite membrane, A is the area

(cm2) of the smallest electrode, and R is the bulk resistance (Ω). Both samples had areas of 0.25

cm2, while the thickness was approximated as 0.001 cm for YbHAP and 0.0005 cm for YbFHAP

from SEM images. Based on the Arrhenius plot, it can be seen that the YbHAP(0.01M)-PSS-Pd-

83

LSCF sample has a slightly higher conductivity throughout the tested temperatures, with

maximum conductivities of ~4x10-5 S/cm and 6.67x10-6 S/cm at 700oC for YbHAP(0.01M)-PSS-Pd-

LSCF and YbFHAP-PSS-Pd-LSCF, respectively. Based on these preliminary results, the amount of

fluoride incorporated into the crystals could be hindering proton mobility when compared to

the unfluorinated samples. Maiti et al. showed that having more than half of the hydroxyl ions

replaced with fluoride can inhibit proton mobility due to fluoride’s high electronegativity and

affinity for the positively charged proton.[8]

Activation energies of 0.585 and 0.588 eV were calculated for YbHAP and YbFHAP,

respectively, from the Arrhenius plot. These values are much lower than the values of 1.0 eV for

yttrium doped HAP as recorded in the literature, and are comparable to other proton

conducting ceramics used for fuel cells.[6, 31] These results suggest that increased conductivity

and decreased activation energy values are a result of minimizing grain boundary resistance,

which act as traps for proton conduction. Also, the aligned orientation of the films along the c-

axis may provide a more direct path for proton transport through the membrane.

84

Figure 4.9. Nyquist plots of YbHAP-PSS-Pd-LSCF from 400-700oC.

Figure 4.10. Nyquist plots of YbFHAP-PSS-Pd-LSCF from 400-700oC.

85

Figure 4.11. Arrhenius plot of YbHAP(0.01M)-PSS-Pd-LSCF and YbFHAP-PSS-Pd-LSCF.

Area specific resistance (ASR) measurements as a function of temperature are shown in

Table 4.3. ASR is calculated by multiplying the bulk resistance, as determined from the EIS

curves, times the electrode area, 0.25 cm2. An ASR of around 0.25 Ω*cm2 is often desired for

operational fuel cells.[32] One of the main factors causing the ASR of these samples to be so

high is the interfacial resistance at the LSCF-apatite interface. Typically, when depositing a

cathode such as LSCF on a membrane using screen printing and brush painting, a sintering step

of around 1,100oC for at least 2 hours is done to assure good contact between the cathode and

membrane.[33] However, this step was avoided for these tests in order to prevent cracking of

the apatite membrane. Therefore, the only sintering that was done occurred as the sample was

86

being heated for conductivity and performance testing. The conductivity values obtained may

therefore drastically underestimate the true conductivity values of the membrane since the

resistance obtained in the Nyquist plot did not separate out the polarization losses from the

anode and cathode, as well as the interfacial resistance between the electrodes and the

membrane. Using our synthesis technique to create yttrium doped membranes and only

sputtering platinum as the cathode resulted in conductivity values approaching 10-2 S/cm at

700oC.[22] The fact that the ytterbium doped samples have such a significantly lower

conductivity value may therefore be due to the significant contribution of resistance from the

electrodes. In the future, low temperature sintering techniques can be applied to improve the

cathode-membrane contact.[34]

Table 4.3. Area specific resistance of YbHAP(0.01M)-PSS-Pd-LSCF and YbFHAP-PSS-Pd-LSCF.

Another reason for the high ASR may have been the deposition process itself for

applying the cathode. A simple brush coating method was used to apply the LSCF slurry to the

membrane. This method, while simple, does not provide the ideal morphology for a cathode

structure. Many parameters need to be considered when applying the cathode such as

thickness, composition, porosity, thermal expansion matching with the membrane, and

membrane poisoning from the cathode.[35] None of these parameters were optimized for

these experiments, thus reducing overall conductivity values and fuel cell performance. Also,

cathode structures such as functionally graded cathodes can be investigated in order to improve

the cathode performance.[36]

87

Aside from the cathode, the electrolyte itself may not have the proper amount of

doping to improve fuel cell performance. Previous works have shown that hydroxyapatite

doped with yttrium tends to increase in conductivity up to a certain percentage, then

dramatically decreases.[6] The same type of trend occurs for fluoride, where roughly 50%

substitution of fluoride for hydroxyl groups tends to produce the maximum conductivity, but

more or less will decrease conductivity.[8] Optimizing the amounts of dopants in a co-doped

sample would require careful composition analysis to be able to correlate small changes in

composition with conductivity.

Figures 4.12 and 4.13 show current-voltage (I-V) and power curves for YbHAP(0.01M)-

PSS-Pd-LSCF and YbFHAP-PSS-Pd-LSCF, respectively. Both samples have an open circuit

potential (OCP), or the voltage with zero current being drawn, close to the theoretical Nernst

potential values at 500 and 700oC, which are 1.06 and 1.01V, respectively. For YbHAP(0.01M)-

PSS-Pd-LSCF, the OCP values were 1.06 and 0.92V at 500 and 700oC, respectively, while for

YbFHAP(0.01M)-PSS-Pd-LSCF, the OCP values were 0.944 and 0.843V at 500 and 700oC,

respectively. The high OCP indicates that the samples have a gas tight layer, since any leakage

due to a porous or cracked membrane would cause a significant drop in the OCP. Also, the high

OCP indicates that there is no significant contribution from electronic conduction. The power

curves for the YbHAP(0.01M)-PSS-Pd-LSCF sample indicate that as temperature increases, the

power density increases. The maximum power density is 0.45 mW/cm2 at 500oC and 1.92

mW/cm2 at 700oC. As for the YbFHAP-PSS-Pd-LSCF sample, the maximum power density is 0.06

mW/cm2 at 500oC and 1.09 mW/cm2 at 700oC. As discussed before regarding ASR values, the

fact that these power densities are roughly 2-3 orders of magnitude lower than current high

88

temperature fuel cells such as YSZ at 700oC can partially be attributed to the lack of optimizing

the cathode deposition process, as well as controlling the membrane dopant amounts.

Figure 4.12. I-V and power curves of YbHAP(0.01M)-PSS-Pd-LSCF from 400-700oC.

89

Figure 4.13. I-V and power curves of YbFHAP-PSS-Pd-LSCF from 400-700oC.

4.4. Conclusions

The ion conducting abilities of hydroxyapatite offers many potential applications as a

sensor, electret, or fuel cell electrolyte. The electrochemical-hydrothermal method described in

this work improves the conductivity of hydroxyapatite by minimizing grain boundary resistance,

optimizing orientation, and can be further enhanced by controlling the composition, particularly

with dopants. Ytterbium and ytterbium-fluoride doped HAP membranes were characterized.

SEM, EDX, and XRD data shows that the crystal membranes can only be doped with a limited

amount of ytterbium until the crystal morphology, composition, and structure changes. With

the proper amount of doping, a dense hydroxyapatite film with the ideal structure for a fuel cell

electrolyte was created. A protocol for testing the fuel cell characteristics of these

hydroxyapatite membranes was investigated. The high open circuit potentials demonstrated

that the membranes were gas tight, as well as being electronically insulating. While the

90

conductivity and performance results were lower than traditional fuel cell electrolytes, there are

many areas for improvement, particularly in applying and optimizing the cathode structure.

91

References





3. Liu, D., K. Savino, and M.Z. Yates, Microstructural engineering of hydroxyapatite

membranes to enhance proton conductivity. Advanced Functional Materials, 2009. 19: p. 3941-

3947.



5. Laghzizil, A., et al., Comparison of Electrical Properties between Fluoroapatite and

Hydroxyapatite Materials. Journal of Solid State Chemistry, 2001. 156: p. 57-60.

6. Yamashita, K., et al., Protonic conduction in yttrium-substituted hydroxyapatite ceramics

and their applicability to H2-O2 fuel cell. Solid State Ionics, 1990. 40/41: p. 918-921.

7. Owada, H., et al., Humidity-Sensitivity of Yttrium Substituted Apatite Ceramics. Solid State Ionics, 1989. 35: p. 401-404.

8. Maiti, G. and F. Freund, Influence of Fluorine Substitution on the Proton Conductivity of

Hydroxyapatite. J. Chem. Soc., Dalton Trans., 1981: p. 949-955.

9. Horiuchi, N., et al., Proton conduction related electrical dipole and space charge

polarization in hydroxyapatite. Journal of Applied Physics, 2012. 112: p. 1-6.

10. Gittings, J., et al., Electrical characterization of hydroxyapatite-based bioceramics. Acta Biomaterialia, 2009. 5: p. 743-754.

11. Steele, B. and A. Heinzel, Materials for fuel-cell technologies. Nature, 2001. 414: p. 345-

352.

12. Ito, N., et al., New intermediate temperature fuel cell with ultra-thin proton conductor

electrolyte. Journal of Power Sources, 2005. 152: p. 200-203.

13. Yamashita, K., K. Kitagaki, and T. Umegaki, Thermal Instability and Proton Conductivity of

Ceramic Hydroxyapatite at High Temperatures. J. Am. Ceram. Soc., 1995. 78: p. 1191-97.

14. Kwon, O. and G. Choi, Electrical conductivity of thick film YSZ. Solid State Ionics, 2006. 177: p. 3057-3062.

92

15. Liu, D., K. Savino, and M.Z. Yates, Coating of hydroxyapatite films on metal substrates by


16. Wei, X., et al., Carbonated hydroxyapatite coatings with aligned crystal domains. Crystal Growth & Design, 2012. 12: p. 3474-3480.

17. Hammari, L.E., et al., Chrystallinity and Fluorine Substitution Effects on the Proton

Conductivity of Porous Hydroxyapatites. J. Solid State Chemistry, 2004. 177: p. 134-138.


yttrium. I. Structure and microstructure. Journal of Biomedical Materials Research, 2001. 59: p. 305-311.

19. Yamashita, K., et al., Effects of sintering ambient H2O vapour on the protonic conduction

properties of ceramic hydroxyapatite. Journal of Materials Science Letters, 1990. 9: p. 4-6.

20. Owada, H., K. Yamashita, and T. Kanazawa, High-temperature stability of hydroxyl ions in

yttrium-substituted oxyhydroxyapatites. Journal of Materials Science Letters, 1990. 9: p. 26-28.

21. Yamashita, K., et al., Trivalent-Cation Substituted Calcium Oxyhydroxyapatite. J. Am.

Ceram. Soc., 1986. 69(8): p. 590-94.

22. Wei, X. and M.Z. Yates, Yttrium-doped hydroxyapatite membranes with high proton

conductivity. Chemistry of Materials, 2012. 24: p. 1738-1743.


with different fluorine contents. Biomaterials, 2004. 26: p. 1205-1210.

24. Ito, N., et al., Electrochemical analysis of hydrogen membrane fuel cells. Journal of

Power Sources, 2008. 185: p. 922-926.

25. Mardilovich, P., et al., Defect-free palladium membranes on porous stainless-steel

support. AIChE Journal, 1998. 44: p. 310-322.

26. Kawasaki, T., Hydroxyapatite as a liquid chromatographic packing. Journal of

Chromatography, 1991. 544: p. 147-184.

27. Murphy, C., Nanocubes and nanoboxes. Science, 2002. 298: p. 2139-2141.

28. Wei, X., et al., Fully dense yttrium-substituted hydroxyapatite coatings with aligned

crystal domains. Crystal Growth & Design, 2012. 12: p. 217-223.

93

29. Feki, H., J. Savariault, and A. Salah, Structure refinements by the Rietveld method of

partially substituted hydroxyapatite: Ca9Na0.5(PO4 )4.5(CO3)1.5(OH)2. J. Alloys and Compounds, 1999. 287: p. 114-120.

30. Laghzizil, A., et al., Mixed ionic conductivities in sodium fluoroapatites. Solid State Ionics,

1993. 67: p. 137-143.

31. Kreuer, K., Proton-conducting oxides. Annu. Rev. Mater. Res., 2003. 33: p. 333-359.

32. Barbir, F., PEM Fuel Cells: Theory and Practice, ed. R. Dorf. 2005: Elsevier Academic

Press.

33. Horita, T., et al., Oxygen reduction mechanism at porous La1-xSrxCoO3-d

cathodes/La0.8Sr0.2Ga0.8Mg0.2O2.8 electrolyte interface for solid oxide fuel cells. Electrochimica

Acta, 2001 46: p. 1837-1845.

34. Simner, S., et al., Enhanced low temperature sintering of (Sr, Cu)-doped lanthanum

ferrite SOFC cathodes. Solid State Ionics, 2004. 175: p. 79-81.

35. Dusastre, V. and J. Kilner, Optimisation of composite cathodes for intermediate

temperature SOFC applications. Solid State Ionics, 1999. 126: p. 163-174.

36. Hart, N., et al., Functionally graded composite cathodes for solid oxide fuel cells. Journal of Power Sources, 2002. 106: p. 42-50.

94

Chapter 5

Electrochemical Reduction of Silver Nanoparticles onto Hydroxyapatite Films

5.1. Introduction

The orthopedic implant market in 2012 was valued at just over $30.5 billion, with over

2.6 million orthopedic implants inserted annually in the United States alone.[1, 2]. Implants are

typically composed of titanium or titanium alloys due to its biocompatability, but much work has

been done to enhance the integration process by coating implant surfaces with hydroxyapatite

(HAP, Ca5(PO4)3OH).[3-6] HAP’s similar composition to bone provides a bioactive surface to

form better tissue ingrowth between the patient’s body and the orthopedic implant, helping to

speed up the recovery process and prevent future prosthetic loosening.[7, 8] However, titanium

or hydroxyapatite coated implants do very little in regards to preventing infections. Over

100,000 implants are infected each year in the United States alone, with medical costs

exceeding $3 billion to treat these infections.[9] Even in the most sterile surgical environments,

bacteria from the surgical equipment, clothing from medical staff, or a patient’s own skin can

still adhere to an implant.[10] Infected implants can be devastating to a patient and may even

require a secondary surgery to clean or replace the implant.[11] In worst case scenarios,

infected implants can lead to amputation or even lethal sepsis. Therefore, there is a high

urgency to not just treat infections, but prevent them from occurring in the first place. If

bacteria adhesion occurs and a bacteria biofilm forms on an implant before tissue regeneration

95

occurs, the biofilm can be ten to one thousand times more resistant to antimicrobial agents

than free-floating bacteria.[12]

While systemic antibiotic delivery can help prevent infections, the large dosage of drugs

used increases the likelihood of a patient suffering negative side effects.[13] Also, continual

administration of antibiotics may cause antibiotic-resistant bacteria to form.[14, 15] Therefore,

much research has been done to apply antibiotics locally to the surgical site. Not only can local

drug delivery reduce the amount of antibiotic administered, it is also more likely to prevent an

infection due to its close proximity to the surgical site. To locally administer drugs, many

surgeons have used bone cements made of polymethylmethacrylate (PMMA) that are loaded

with antibiotics.[16, 17] However, PMMA does nothing to help stimulate osteointegration of

the implant with the body and may even need to be removed in a subsequent surgery.[18, 19]

Many groups have also tried to load drugs into HAP coated titanium implants.[20-22] However,

the amount of antibiotic loaded and the release profile usually relies on a bulk diffusion

mechanism or surface roughness characteristics, limiting the parameters to enhance the

amount of loading and extend the release time of antibiotics, particularly for smaller

antibiotics.[21, 22] Electrostatically charged hydrogels have been synthesized in an effort to

extend the drug release, but they have very little benefit in terms of helping the implant

integrate into the body.[23]

Silver ions have long been known to have antimicrobial properties, while having minimal

toxicity to human cells at concentrations below 10mg/l.[10, 24, 25] Silver ions are effective

against a broad range of pathogens such as P. aeruginosa, E. coli, S. epidermidis, and S. aureus,

and these bacteria have been shown to not form a resistance to silver ions.[10, 26] Ionized

silver has been shown to disrupt the function and structure of bacteria cell membranes, leading

96

to cell distortion and death. Also, silver ions disrupt critical cell metabolic proteins and enzymes

by binding to thiol groups, as well as bind to DNA and RNA to interfere with replication.[24, 26-

28]

In this section, a novel and simple method for synthesizing a bioactive HAP coating with

antimicrobial silver nanoparticles is presented (AgHAP). The initial HAP layer was deposited

onto a titanium substrate using an electrochemical method. By shortening the electrochemical

deposition time, a uniform and thin film of HAP crystals was formed. Then, an electrolyte

solution containing silver ions and salt was used to electrochemically reduce silver nanoparticles

directly onto the HAP crystals. The HAP and AgHAP films were characterized and tested for

antimicrobial capabilities.

5.2. Experimental

Materials


25 mm and 0.127 mm thick), AgNO3, MgCl2·6H2O (99.0-102% purity) and NaHCO3 (99.7-100.3%

purity) were purchased from Alfa Aesar. K2HPO4 (99.99% purity), K2HPO4·3H2O (99% purity),

CaCl2·2H2O (99+% purity), NaCl (≥99.0% purity), tris(hydroxymethyl)- aminomethane (tris)

(99.8+% purity), and tryptic soy broth were all obtained from Sigma- Aldrich. CaCl2 (99.5%

purity) and Na2SO4 (100.1% purity) was purchased from J.T. Baker. KCl (99.7% purity), 36.5-

38.0% hydrochloric acid, and 28.0-30.0% pure ammonium hydroxide were purchased from

Mallinckrodt Chemicals. Detergent powder was purchased from Alconox. Deionized (DI) water

was used for all solutions.


97

Titanium substrates were gently polished with SiC paper (800 grit) to provide surface









to the negative and positive electrodes of a direct current power supply (Instek GPS-3030D),




Electrochemical reduction of Ag nanoparticles

For each sample coated with Ag nanoparticles, an 80 mL solution containing 1.25 mM

Ag(NO3) and 1.25 mM NaCl was prepared. The electrolyte solution was then heated to 95°C.

The HAP coated substrate and a platinum plate were held parallel to each other with a fixed

distance of separation of 10 mm. The substrate and platinum plate were connected to the

negative and positive electrodes of a direct current power supply, respectively, and immersed in

the electrolyte solution. The electrochemical reaction was carried out for 90 seconds at 12.5

mA/cm2 (area relative to the platinum plate), then rinsed off with deionized water and dried in

air.

Bacteria growth testing

98

A solution of Staphylococcus aureus (S. aureus) (ATCC 25923) bacteria in tryptic soy

broth was grown overnight with shaking at 37oC. The bacteria concentration was then diluted

with tryptic soy broth until the solution’s absorbance value at 490 nm was 0.1. Bacteria growth

curves were produced by placing n = 6 HAP and AgHAP samples into wells of a 24 well plate.

Each well was then filled with 2 mL of bacteria suspension, as well as six control wells with just

bacteria suspension. The samples were then incubated at 37oC. Bacteria growth was measured

at various times by taking 200 μL of bacteria solution from each sample and placing it into a 96

well microtitre plate using light scattering at 490 nm. (BioTek FLx800 Fluorescence Microplate

Reader). The 200 μL solution was then placed back into the solution of their respective samples

after each measurement.

Simulated body fluid setup

N = 3 HAP, AgHAP, and titanium substrates were placed facing up in plastic tubes

containing 15 mL of simulated body fluid that was prepared as described by Kokubo, but with

1.5 times the concentrations.[29]. The chemicals used, their amounts, and order in which they

were added is listed in Table 5.1. The final pH was adjusted to 7.4 with tris. The solution was

kept at 37oC and the samples were left for 24 hours.

Table 5.1. Recipe for simulated body fluid.

99

Sample characterization



determined by EDX (EDAX). Three spots on two different samples were probed at an

accelerating voltage of 15 kV and the values were averaged and standard deviations were

calculated. The crystal structure was determined by X-ray diffraction (XRD) (Philips PW3020)

with Cu K1 radiation (λ=1.540560 Ǻ) from 20-60o with a step rate of 0.02 degrees/second. Phase

identification was made by comparison with the JCPDS files. For zeta potential measurements,

approximately 1 mg of HAP sample was scrapped off and placed into 1 mL of DI water. A 90Plus

Particle Size Analyzer by Brookhaven Instruments Corporation was used. 10 runs, with 5

cycles/run were used to measure the zeta potential. The Smoluchowski model was used for

calculating zeta potential values.


Figure 5.1 shows SEM images of the HAP and AgHAP samples. Individual HAP crystals

are ~ 800 nm long and ~30 nm wide. For AgHAP, the Ag nanoparticles are evenly distributed

over the entire HAP coating and along each individual HAP crystal. The size of the Ag

nanoparticles varies from 5 to 50 nm. The variation in nanoparticle size can potentially be

advantageous to control the release of Ag+ since the rate of ion release is proportional to

particle surface area.[30] A range of particle sizes can insure that there will be some Ag+

released quickly from the smaller particles, while the larger particles will release Ag+ more

slowly. Also, according to the Kelvin equation, particles in the nano-size range have a higher

solubility limit than the bulk, allowing for a higher amount of Ag+ delivered if desired.[31]

100

Figure 5.1. SEM of a) HAP side view, b) HAP top view, c) AgHAP side view, d) AgHAP top view,

and e) AgHAP zoomed in.

101

EDX results in Table 5.2 and Figure 5.2 confirm the presence of Ag in the AgHAP samples

and give an approximate amount of 1.50 At%. While there is a measurable amount of NaCl on

the HAP sample, most of it is removed after the Ag electrochemical deposition.

Table 5.2. EDX of HAP and AgHAP.

Figure 5.2. EDX spectrum of AgHAP.

XRD in Figure 5.3 verifies that the silver electrodeposition does not change the crystal

structure of the HAP. The peaks for HAP and AgHAP match standard HAP and titanium

102

according to JCPDS 09-0432 and 01-1197, respectively. HAP has an enhanced (0 0 2) peak,

showing some preferential orientation along the c-axis. The silver could not be detected using

XRD since the highest peak for silver is at 38.117o according to JCPDS 4-0783, which overlaps

with the (0 0 2) peak for titanium.

Figure 5.3. XRD of HAP and AgHAP.

The formation of silver nanoparticles onto HAP crystals can occur via a variety of

mechanisms. With the applied potential being much larger than the reduction potential of the

metal, nonequilibrium conditions are relevant here.[32] The silver ions are reduced in solution

according to the equation:[33]

Ag+ + e- Ag (1)

The positively charged Ag+ is attracted to the titanium substrate since titanium is attached to the

negative electrode during electrochemical deposition. As the Ag+ forms Ag nanoparticles, the

nanoparticles may become lodged into the HAP crystals due to Ag+’s electrostatic attraction

toward the titanium. Even the HAP crystals themselves inherently have a negative surface

103

charge. The HAP crystals have a zeta potential of -15.60 ± 0.51. This is a common result since

HAP is thought to have an excess concentration of PO43- groups at its surface.[34-36]

Another mechanism could occur via deprotonation of the HAP hydroxyl group by a base

B:

HAP−OH + B HAP−O- + BH (2)

This step would be followed by an electrophilic attack of Ag+:

HAP−O- + Ag+ HAP−Ag (3)

Also, a very similar mechanism could occur directly onto the titanium substrate:[37]

Ti−OH + B Ti−O- + BH (4)

Ti−O- + Ag+ Ti−Ag. (5)

Finally, some Ag+ can directly substitute for Ca2+ in the HAP crystal structure to form Ca5-

xAgxHx(PO4)3OH, where the extra hydrogen is necessary to maintain charge neutrality.

Figure 5.4 compares HAP, AgHAP, and a piece of titanium after being immersed in

simulated body fluid (SBF) for 24 hours. This is a simple, acellular bioactivity test used to

evaluate a coating’s ability to stimulate osseointegration. Both the HAP and AgHAP samples

clearly have a thick new apatite layer formed on and in between the crystals. The thickness of

the new apatite layer is approximately the same for both samples, showing that the silver does

not impede the apatite-forming ability of the HAP coating. EDX results in Table 5.3 verify that

the composition of the new layer is indeed that of apatite. Titanium, on the other hand, clearly

shows its lack of bioactivity. Most of the titanium surface was exposed except for a few small

NaCl and CaCl2 deposits.

104

Figure 5.4. SEM of new HAP growth after being immersed in SBF for 24 hours, a) HAP after SBF immersion, b) AgHAP after SBF immersion, c) representative top view of samples in images a

and b, and d) Ti after SBF immersion.

Table 5.3. EDX of new apatite layer after immersion in SBF for HAP, AgHAP, and Ti.

Figure 5.5 shows the growth profile of S. aureus bacteria and when it is exposed to the

HAP and AgHAP samples. There is a clear distinction between the two samples, showing the

ability of the silver ions released from the AgHAP samples to retard bacteria growth. Based on

105

the SBF results, more Ag can potentially be deposited onto the HAP surface in order to further

reduce bacteria growth, while also not hindering new apatite formation. Future tests should be

done to determine the cytotoxicity of increased amounts of Ag nanoparticles.

Figure 5.5. Growth of S. Aureus bacteria when exposed to AgHAP, HAP, and in suspension.

5.4. Conclusions

A novel and simple method to electrochemically reduce Ag nanoparticles onto bioactive

HAP was developed and characterized. The electrochemical reduction is a fast process that

uniformly deposits Ag nanoparticles over the entire HAP coating, while not requiring any harsh

reducing agents. The AgHAP coating has comparable bioactivity to the HAP coating according to

SBF results, but also has antimicrobial properties.

106

References

1. Hetrick, E. and M. Schoenfisch, Reducing implant-related infections: active release

strategies. Chemical Society Reviews, 2006. 35: p. 780-789.

2. Orthopedic Implants A Global Market Overview. 2011.

3. S.Ban and S. Maruno, Hydrothermal-electrochemical deposition of hydroxyapatite. Journal of Biomedical Materials Research, 1998. 42: p. 387-395.


yttrium. I. Structure and microstructure. Journal of Biomedical Materials Research, 2001. 59: p.

305-311.

5. Suchanek, W. and M. Yoshimura, Processing and properties of hydroxyapatite-based

biomaterials for use as hard tissue replacement implants. Journal of Materials Research, 1998.

13: p. 94-117.

6. Zakaria, S., et al., Nanophase hydroxyapatite as a biomaterial in advanced hard tissue

engineering: A review. Tissue Eng. Part B Rev., 2013.

7. Kuo, M. and S. Yen, The process of electrochemical deposited hydroxyapatite coatings on

biomedical titanium at room temperature. Materials Science and Engineering C, 2002. 20: p.

153-160.

8. Kar, A., K. Raja, and M. Misra, Electrodeposition of hydroxyapatite onto nanotubular

TiO2 for implant applications. Surface and Coatings Technology, 2006. 2001: p. 3723-3731.

9. Darouiche, R., Treatment of infections associated with surgical implants. New England Journal of Medicine, 2004. 350: p. 1422-1429.

10. Simchi, A., et al., Recent progress in inorganic and composite coatings with bactericidal

capability for orthopedic applications. Nanomedicine, 2011. 7: p. 22-39.

11. Gristina, A., Biomaterial-centered infection: microbial adhesion versus tissue integration.

Science, 1987. 237: p. 1588-1595.

12. Davis, D., Understanding biofilm resistance to antibacterial agents. Nature Reviews Drug

Discovery, 2003. 2: p. 114-122.

13. Beringer, P., A. Wong-Berringer, and J. Rho, Economic aspects of antibacterial adverse

effects. PharmacoEconomics, 1998. 13: p. 35-49.

107

14. Neut, D., et al., Residual gentamicin-release from antibiotic-loaded

polymethylmethacrylate beads after 5 years of implantation. Biomaterials, 2003. 24: p. 1829-1831.

15. Eiff, C.V., et al., Development of gentamicin resistant small colony variants of S. aureus

after implantation of gentamicin chains in osteomyelitis as a possible cause of recurrence. Z

Orthop Ihre Grenzgeb, 1998. 136: p. 268-271.

16. Brauer, S.H.G. and G. Dickson, A characterization of polymethylmethacrylate bone

cement. Journal of Bone Joint Surgery, 1975. 55A: p. 380-391.

17. Wahlig, H., et al., Experimental and pharmacokinetic studies with gentamicin PMMA

beads. Zentralbl Chir, 1979. 104: p. 923-933.

18. Kim, S., et al., The characteristics of a hydroxyapatite-chitosan-PMMA bone cement.

Biomaterials, 2004. 25: p. 5715-5723.

19. Schnieders, J., et al., Controlled release of gentamicin from calcium phosphate-

poly(lactic acid-co-glycolic acid) composite bone cement. Biomaterials, 2006. 27: p. 4239-4249.

20. Stigter, M., et al., Incorporation of different antibiotics into carbonated hydroxyapatite

coatings on titanium implants, release and antibiotic efficacy. Journal of Controlled Release,

2004. 99: p. 127-137.

21. Stigter, M., K.d. Groot, and P. Layrolle, Incorporation of tobramycin into biomimetic

hydroxyapatite coating on titanium. Biomaterials, 2002. 23: p. 4143-4153.

22. Liu, W., et al., Gentamicin-loaded strontium-containing hydroxyapatite bioactive bone

cement—An efficient bioactive antibiotic drug delivery system. Journal of Biomedical Materials

Research B: Applied Biomaterials, 2010. 95B: p. 397-406.

23. Serizawa, T., D. Matsukuma, and M. Akashi, Loading and release of charged dyes using

ultrathin hydrogels. Langmuir, 2005. 21: p. 7739-7742.

24. Chen, W., et al., In vitro anti-bacterial and biological properties of magnetron co-

sputtered silver-containing hydroxyapatite coating. Biomaterials, 2006. 27: p. 5512-5517.

25. Vik, H., et al., Neuropathy caused by silver adsorption from arthroplasty cement. Lancet,

1985. 1: p. 872.

26. Rai, M., A. Yadav, and A. Gade, Silver nanoparticles as a new generation of

antimicrobials. Biotechnology Advances, 2009. 27: p. 76-83.

108

27. Campoccia, D., L. Montanaro, and C. Arciola, A review of biomaterials technologies for

infection-resistant surfaces. Biomaterials, 2013. 34: p. 8533-8554.

28. Liu, X., et al., Synthesis of silver-incorporated hydroxyapatite nanocomposites for

antimicrobial implant coatings. Applied Surface Science, 2013. 273: p. 748-757.

29. Kokubo, T., et al., Solutions able to reproduce in vivo surface-structure changes in

bioactive glass-ceramic A-W. J. Biomed. Mater. Res, 1990. 24: p. 721-734.

30. Ionita, D., et al., Merit and demerit effects of silver nanoparticles in the bioperformance

of an electrodeposited hydroxyapatite: nanosilver composite coating. J. Nanopart. Res., 2012. 14: p. 1152-1162.

31. Muller, R., C. Jacobs, and O. Kayser, Nanosuspensions as particulate drug formulations in

therapy. Rationale for development and what we can expect for the future. Advanced Drug

Delivery Reviews, 2001. 47: p. 3-19.

32. Kaniyankandy, S., et al., Electrodeposition of silver nanodendrites. Nanotechnology,

2007. 18: p. 1-6.

33. Lu, X., et al., Nano-Ag-loaded hydroxyapatite coatings on titanium surfaces by

electrochemical deposition. J. R. Soc. Interface, 2011. 8: p. 529-539.

34. Kim, H., et al., Process and kinetics of bonelike apatite formation on sintered

hydroxyapatite in a simulated body fluid. Biomaterials, 2005. 26: p. 4366-4373.

35. Vandiver, J., et al., Nanoscale variation in surface charge of synthetic hydroxyapatite

detected by chemically and spatially specific high-resolution force spectroscopy. Biomaterials, 2005. 26(271-283).

36. Brown, P. and R. Martin, An analysis of hydroxyapatite surface layer formation. Journal

of Physical Chemistry B., 1999. 103: p. 1671-1675.

37. Santillan, M., N. Quaranta, and A. Boccaccini, Titania and titania-silver nanocomposite

coatings grown by electrophoretic deposition from aqueous suspensions. Surface & Coatings Technology, 2010. 205: p. 2562-2571.

109

Chapter 6

Conclusions and Future Work

6.1. Conclusions

This thesis investigated a novel synthesis technique to deposit a highly oriented, crystalline film

of hydroxyapatite onto various metal substrates. Various properties of the coating were characterized

and the coatings were synthesized with many applications in mind such as a fuel cell electrolyte and an

orthopedic coating with antimicrobial properties. A summary of the main conclusions from these

studies are as follows:

(a) The thermal stability of novel hydroxyapatite membranes was investigated in order to determine

their feasibility in high temperature applications. It was found that substituting fluoride into the crystal

structure prevented the hydroxyapatite coating from decomposing into other calcium phosphate phases

such as β-TCP at temperatures up to at least 900oC, whereas unfluorinated samples began

decomposition around 600oC. A yttrium-fluoride co-doped sample demonstrated the potential to

further modify properties of the film by adding additional dopants, while also maintaining a high

thermal stability. The co-doped films also produced a highly dense and thin film after only one

hydrothermal growth, which may be beneficial for applications requiring a dense film such as a fuel cell

electrolyte. Finally, a steam atmosphere was also shown to retard thermal decomposition since it shifts

the decomposition equation to favor hydroxyapatite.

(b) Much research has been done studying the influence of polarizing hydroxyapatite, particularly with a

negative surface charge, in order to improve osteointegration. In this thesis, yttrium-fluoride co-doped

hydroxyapatite membranes on titanium were prepared using a hydrothermal growth and two different

110

seeding steps. Both samples had a very large stored charge according to thermally stimulated

depolarization current measurements, even without polarization. The large stored charge is speculated

to come from a composition concentration gradient within the sample. Also, the c-axis orientation of

the crystals relative to the substrate and the fact that they had no grain boundaries along the crystal

allowed the current measured from the samples to be high.

(c) SEM, EDX, and XRD analysis characterized hydroxyapatite films doped with various amounts of

ytterbium. The results showed that there is a range of ytterbium that can be incorporated into the

crystal structure until a phase change occurs. Once the phase change occurs, the membrane loses its

unique crystal orientation and morphology. Ytterbium and ytterbium-fluoride co-doped hydroxyapatite

membranes were then tested as fuel cell electrolytes. The results showed that the membranes were

ionically conductive, electronically insulating, and gas tight. While the conductivity and power output of

the fuel cells were low compared to state of the art fuel cell electrolytes, much can be done to improve

performance such as optimizing the cathode composition and deposition process, as well as controlling

the amount of dopant used.

(d) A hydroxyapatite film that was electrochemically deposited onto titanium was uniformly coated with

silver nanoparticles using a second electrochemical deposition step. The process for reducing silver

nanoparticles on hydroxyapatite is fast, simple, and can be done on a variety of substrate shapes. The

apatite coating with and without silver nanoparticles were shown to have comparable apatite forming

abilities in a simulated body fluid, while the silver coated samples have a superior capability to slow

down bacteria growth.

111

6.2. Future Work

The unique morphology, orientation, composition, and electrical properties of the

hydroxyapatite membranes investigated in this thesis offers a wide range of potential applications such

as a fuel cell electrolyte, catalyst support, sensor, electret, or orthopedic implant coating. Some future

projects are suggested to continue investigating the unique properties of this coating.

(a) Characterization of coated screws for clinical applications

The versatility of the electrochemical seeding step allows a hydroxyapatite film to be deposited

onto a variety of shapes such as rods, screws, or even porous materials. Figure 6.1 shows preliminary

results of a coating that was electrochemically deposited onto a titanium dental screw. Rather than

using a flat platinum counter electrode for the deposition, a piece of conductive graphene was used to

surround the screw. The graphene surrounding the screw ensured that a uniform deposition would

occur. Using these screws, in vivo analysis can be done to properly understand how these coatings

behave in the body and how they compare to uncoated screws and state of the art plasma sprayed

hydroxyapatite screws.

Figure 6.1. Electrochemically deposited hydroxyapatite coating on a titanium dental screw.

112

(b) Characterization of various metals reduced on hydroxyapatite films

The work in chapter 5 can be expanded in numerous ways. First, more analysis can be done to

understand if the silver nanoparticles not only reduce bacteria growth, but also prevent bacterial

adhesion on HAP’s surface using a Calgary plate. The Calgary plate would allow for carefully placing the

samples upside down in a bacteria suspension so just the coating surface is in the suspension. Upon

drying the samples, SEM observations can give a better idea about bacterial adhesion on the coatings.

Also, more tests should also be done to determine how effective the coatings are against other types of

bacteria. In vitro studies can also be performed using mesenchymal stem cells in order to understand

how the coatings interact with actual human cells.[1] It will be important to understand what amount of

silver can be tolerated before the body’s cells will be negatively impacted. In addition to silver, other

metals can be investigated. Cu2+

has also been shown to have antimicrobial properties.[2] Figure 6.2

shows an SEM image of copper nanoparticles electrochemically reduced onto a hydroxyapatite coating.

Other charged antibiotics or charged species can also be investigated to cover the coating. Also,

hydroxyapatite has been used as a catalyst support with palladium nanoclusters for the selective

oxidation of alcohols.[3] Other catalysts can be deposited and tested in which the HAP coating acts as a

catalyst support.

Figure 6.2. Electrochemical reduction of copper nanoparticles onto the HAP seed layer.

113

References

1. Bodhak, S., S. Bose, and A. Bandyopadhyay, Bone cell–material interactions on metal-ion doped

polarized hydroxyapatite. Materials Science and Engineering C, 2011. 31: p. 755-761.

2. Bir, F., et al., Electrochemical depositions of fluorohydroxyapatite doped by Cu2+, Zn2+, Ag+ on

stainless steel substrates. Applied Surface Science, 2012. 258: p. 7021-7030.

3. Mori, K., et al., Hydroxyapatite-supported palladium nanoclusters: A highly active heterogeneous

catalyst for selective oxidation of alcohols by use of molecular oxygen. J. Am. Chem. Soc., 2004. 126: p.

10657-10666.

Microstructured Hydroxyapatite Membranes for Ion … · 2016-03-25 · Microstructured...

Documents

Transcript of Microstructured Hydroxyapatite Membranes for Ion … · 2016-03-25 · Microstructured...