ELECTROCHEMICAL SENSORS
FOR ORGAN-ON-CHIPS
RuO2 for amperometric and potentiometric sensors
Chair / secretary:
prof. dr. J.N. Kok University of Twente Supervisors:
dr. ir. W. Olthuis University of Twente prof. dr. ir. A. van den Berg University of Twente Committee Members:
dr. F. Bedioui Centre National de la Recherche Scientifique prof. dr. G. Urban University of Freiburg
dr. A. D. van der Meer University of Twente prof. dr. ir. S. le Gac University of Twente prof. dr. ir. J. E. ten Elshof University of Twente
The research presented in this thesis was carried out at the BIOS, Lab on a Chip group at the MESA+ Insititute for Nanotechnology and the Technical Medical Centre of the University of Twente, Enschede, the Netherlands. The research was funded by the European Research Council (ERC) Advanced Grant awarded to Prof. dr. ir. A. van den Berg (grant agreement no 669768, VESCEL project).
Cover design: Esther Tanumihardja
Printed by: Ridderprint | www.ridderprint.nl | on biotop paper Lay-out: Esther Tanumihardja
ISBN: 978-90-365-5103-8 DOI: 10.3990/1.9789036551038
© 2020 Esther Tanumihardja, The Netherlands. All rights reserved. No parts of this thesis may be reproduced, stored in a retrieval system or transmitted in any form or by any means without permission of the author.
ELECTROCHEMICAL SENSORS
FOR ORGAN-ON-CHIPS:
RUO
2FOR AMPEROMETRIC AND POTENTIOMETRIC
SENSORS
DISSERTATION
to obtain
the degree of doctor at the University of Twente,
on the authority of the rector magnificus,
prof. dr. ir. A. Veldkamp,
on account of the decision of the Doctorate Board
to be publicly defended
on Friday, 19 February 2021 at 16.45 hours
by
Esther Tanumihardja
born on the 21
stof October, 1991
dr. ir. W. Olthuis, promotor
TABLE OF CONTENTS
CHAPTER 1
INTRODUCTION AND OUTLINE
1.1. Organ-on-chips: origin and advances ... 2
1.2. Organ-on-chip sensors ... 4
1.2.1. Chemical readout technologies ... 4
1.2.2. Requirements for organ-on-chip sensors ... 6
1.3. Research aim ... 7
1.4. Framework ... 7
1.5. Thesis outline ... 7
1.6. References ... 8
CHAPTER 2
FABRICATION OF RUTHENIUM OXIDE NANORODS
2.1. Introduction ... 14 2.2. Fabrication process ... 14 2.2.1. Precursor preparation ... 14 2.2.2. Chip fabrication ... 15 2.2.3. Chip modification ... 15 2.2.4. Characterisation... 162.3. RuOx nanorods growth ... 16
2.3.1. Heating temperature... 17 2.3.2. Substrate ... 18 2.3.3. Heating time ... 19 2.4. Electrochemical characteristics ... 20 2.5. Conclusion ... 21 2.6. Acknowledgements ... 21 2.7. References ... 21
RUTHENIUM OXIDE AS pH SENSOR FOR ORGAN-ON-CHIP
3.1. Introduction ... 26
3.2. Materials and Methods ... 28
3.2.1. Electrode fabrication ... 28
3.2.2. Experimental setup ... 28
3.2.3. Measurement protocol ... 28
3.3. Results ... 29
3.3.1. Fabrication and characterization ... 29
3.3.2. pH sensitivity and selectivity ... 30
3.3.3. Drift and aging ... 31
3.3.4. Response time ... 32
3.3.5. Oxygen sensitivity ... 33
3.4. Discussion ... 34
3.4.1. pH response and selectivity ... 34
3.4.2. Drift and aging ... 35
3.4.3. Response time and oxygen sensitivity... 36
3.5. Conclusions and outlook ... 36
3.6. Acknowledgements ... 37
3.7. References ... 37
CHAPTER 4
MEASURING BOTH pH AND O
2WITH A SINGLE ON-CHIP SENSOR IN CULTURES OF
hPSC-CMs TO TRACK INDUCED CHANGES IN CELLULAR METABOLISM
4.1. Introduction ... 424.2. Materials and methods ... 44
4.2.1. Electrodes and setup ... 44
4.2.2. pH and O2 sensing characterisation ... 44
4.2.3. Cell culture and measurement ... 45
4.3. Results and discussion ... 47
4.3.1. Fabrication results ... 47
4.3.2. pH sensing characteristics ... 48
4.3.3. O2 sensing characteristics ... 48
4.3.4. Measurement of induced changes in hPSC-CMs metabolism ... 50
4.3.5. pH sensing of hPSC-CMs metabolism... 51
4.4. Conclusion ... 55
4.5. Acknowledgements ... 56
4.6. References ... 56
CHAPTER 5
ON-CHIP ELECTROCATALYTIC NO SENSING USING RUTHENIUM OXIDE NANORODS
5.1. Introduction ... 605.2. Experimental ... 62
5.2.1. DEANONOate solution preparation ... 62
5.2.2. EC/MS experiments ... 63
5.2.3. Preparation of RuOx electrode ... 64
5.2.4. Electrocatalytic experiments ... 64
5.2.5. Cell culture and biological experiments... 65
5.3. Results and discussion ... 67
5.3.1. NO formation from DEANONOate... 67
5.3.2. RuOx nanorods fabrication results... 69
5.3.3. Electrocatalytic NO oxidation on RuOx ... 70
5.3.4. Signal from cells-generated NO ... 73
5.4. Conclusions and future work ... 77
5.5. Acknowledgements ... 78
5.6. References ... 78
CHAPTER 6
MONITORING CONTRACTILE CARDIOMYOCYTES VIA IMPEDANCE USING
MULTIPURPOSE THIN FILM RUTHENIUM OXIDE ELECTRODES
6.1. Introduction ... 846.2. Experimental ... 85
6.2.1. Setups ... 85
6.2.2. Cells seeding and culture... 86
6.2.3. Impedance and visual recording ... 87
6.3. Results and discussion ... 87
6.3.1. Measurement frequency ... 87
6.3.2. Measurement in a well plate with Pt wires ... 88
6.3.3. Measurement in chip holder with RuOx electrode ... 92
6.6. References ... 94
CHAPTER 7
TOWARDS ON-CHIP INTEGRATED RUTHENIUM OXIDE MULTI SENSORS
7.1. Introduction ... 1007.1.1. The demand for miniaturised integrated sensors ... 100
7.1.2. Theory of miniaturised electrochemical sensors ... 100
7.1.3. Study design ... 101
7.2. Experimental ... 102
7.2.1. Design ... 102
7.2.2. RuOx fabrication ... 104
7.2.3. Setup and testing ... 106
7.3. Results and discussion ... 109
7.3.1. Fabrication results ... 109 7.3.2. Potentiometric pH sensing ... 111 7.3.3. Amperometric NO sensing... 112 7.4. Conclusion ... 113 7.5. Acknowledgements ... 114 7.6. References ... 114
CHAPTER 8
SUMMARY AND OUTLOOK
8.1. Summary ... 118 8.2. Outlook... 119 8.2.1. RuOx fabrication ... 119 8.2.2. Biologically-relevant applications ... 120 8.3. References ... 121APPENDICES
A. Supplementary Information for Chapter 2 ... 125B. Supplementary Information for Chapter 3 ... 130
C. Supplementary Information for Chapter 4 ... 132
E. Supplementary Information for Chapter 6 ... 147
F. Supplementary Information for Chapter 7 ... 150
Samenvatting ... 151
Scientific output... 152
Funding and contribution ... 154
CHAPTER 1
INTRODUCTIONS AND OUTLINE
The work presented in this thesis was aimed to develop sensors for organ-on-chips devices. In this chapter, the provenance and significance of organ-on-chips are elaborated, providing the bigger context of the work’s relevance. General background of the different sensing approaches is discussed, as well as the requirements for successful organ-on-chip sensors. Lastly, the framework and outline of this thesis are detailed.
2
1.1.
Organ-on-chips: origin and advances
Therapeutic drugs undoubtedly play a vital role in global health. In the last century, drugs have played a significant role in prolonging life expectancy as well as improving life quality globally [1]–[3]. As we continue to rely on therapeutic drugs, more and more efforts have been put into drug discovery and development. However, in contrast to the technological and scientific advances in the field, the process of discovering and developing new drugs is increasingly less productive and more costly [4]. This decline was already recognised in the early 1950s, and many facets to it have been identified since [4]–[6]. Part of the problem was the lack of adequate models on which the drug candidates can be accurately tested [6], [7]. This problem has been well-established since the 1980s, yet, it is still the case in today’s drug research and development process.
Before testing on human volunteers (during the clinical trial phase), a drug candidate must first be tested on cell cultures (in vitro) or animals (in vivo) models during the pre-clinical phase. Only a few compounds, which efficacy and safety have been proven during the pre-clinical phase, then undergo the expensive and lengthy clinical trial phase. Despite the heavy attrition of drug candidates during the pre-clinical phase, only 12% of the drug candidates are ultimately approved for use by the Food and Drugs Administration [8], [9]. This confirms the inability of the models currently used in pre-clinical testing to predict human (patho)physiology.
The two available models have their pros and cons (Table 1.1). The established in vivo animal model has provided us with many insights regarding complex (mammalian) (patho)physiology. However, many diseases and toxicity are proven to be species-specific [6], [10], contesting the predictive value of the animal’s response for human clinical trial. Exacerbated by the ethical concerns of animals affliction (that, arguably, is often fruitless), momentous efforts have been made to replace, reduce, and refine (the 3Rs) the animal models [11]. The (2D) in vitro models were developed to achieve a more ethical model with higher throughput and better species specificity than the in vivo models, achieved by reducing the model down to a layer of cultured human cells. However, the same simplicity that makes them so attractive also deters them from replicating vital aspects of human physiology realistically [7], [12]. When cultured in the absence of cues from their in vivo (micro)environment, differentiated cells were reported to lose their differentiated functions over time [13]. Culturing the cells in 3D construction (typically in gels) can provide the cells with some of these cues, however, not wholly [14].
Organ-on-chips: origin and advances
3
Table 1.1. Comparison of the different models used in the pre-clinical phase of drug discovery.
Model Complexity Species specificity Throughput
In vivo animal model Complex Cross-species Low 2D in vitro Simple Species-specific High Organ-on-chip Simple to complex Species-specific High
Organ-on-chips, also called microphysiological models, were developed to offer an in vitro model capable of replicating human physiology realistically in a chip [15], [16]. In this model, human cells are cultured in a microfluidic device, where the culture condition of the cells can be modified to mimic the in vivo microenvironment of an organ or tissue. More cues can be micro-engineered within these devices, such as tissue-tissue interaction, mechanical or electrical stimulation, as well as spatiotemporal gradients of chemicals. In response to these cues, the cells have been shown to have more realistic phenotypes, and they can even self-organise to depict organ-like architecture and functions [12], [14]–[17]. This new in vitro model can offer a realistic, species-specific, and high-throughput model for drug screening. Such a model hopes to improve the success rate of clinical trials, therefore lowering the drug development costs. On top of that, it can prevent the loss of promising drug candidates that are falsely eliminated early in the pipeline due to lack of efficacy or safety when tested in unrepresentative animal models.
Since the first organ-on-chip device in 2010 [16], different models of different organs have been developed both for healthy [18]–[20] and diseased [21]–[23] physiologies. Along with the growing library of on-chip models, their applications also broadened beyond drug screening. Similar models are now used for fundamental research as well, delivering new insights on physiology, pathogenesis, and organogenesis [24]–[26]. The number of reports on development towards platforms for clinical trials on-chip [27], [28] and personalised medicine [29] has also been growing.
Accompanying the diversifying applications, the technological library of organ-on-chip is also steadily growing. This includes expanding the cell sources (from cell lines to induced pluripotent stem cells [7], [30]), parallelised and interconnected organ-on-chip devices to increase throughput and emulate organ-to-organ interactions (also named body-on-chip) [30], [31], and the inclusion of microsensors or readout capability [32]–[34].
4
1.2.
Organ-on-chip sensors
Inclusion of microsensors within organ-on-chips would allow evaluation of the cells’ conditions without the need to terminate or interrupt the experiment to obtain samples, providing meaningful information beyond end-point assays. The value of such information came about the realisation of the varied timing of cells’ response to drugs. Some drugs showed an immediate, acute response from the cells, while others triggered chronic, or even delayed response [32], [35]. More often than not, the cells’ response comes down to their metabolism or biochemical pathways, which can often be captured in the expression or expression rates of certain (chemical) metabolites. Considering the ability to capture the cells’ response is imperative for evaluation of drugs efficacy and safety, online in-situ monitoring of the cell’s metabolites is indispensable. A great deal of current organ-on-chip studies also scrutinise the cells’ biochemical pathways, which thus can highly benefit from chemical sensors in ways similar to the drug response studies. In addition, the integrated (micro)sensors can also provide spatiotemporal-specificity to the chemical information, affording added meaningful information; for example, in studies involving gradients or non-homogenous cell layers.
Comparable to the expansion of the organ-chips field, the applications of on-chip chemical sensors have since expanded to include, among others, fundamental research, standardisation of culture parameters, characterisation of stem cells-derived cells. The following part of the introduction will explore the chemical sensing/readout technologies in terms of suitability for organ-on-chips, and discuss the requirements for successful organ-on-chip sensors.
1.2.1. Chemical readout technologies
As pointed out, organ-on-chips can highly benefit from real-time chemical analysis from inside the chips. While there are many sensing techniques capable of providing this information, only a handful are suitable for organ-on-chip integration. Understandably, miniaturisability of the sensors or probes is crucial in integrating them in the micrometre-sized channels or devices. Furthermore, the majority of organ-on-chip devices are perfused with complex biological matrices, which are composed of thousands of components (e.g. proteins, sugars, nucleic acids, ions, radicals, and many other small organic molecules). Therefore real-time sensing would necessitate selective measurement in the presence of biological molecules. As sample-preparation complicates the on-chip operations, label-free and non-destructive measurements are also preferred.
Organ-on-chip sensors
5 Figure 1.1 below shows domains or classes of real-time sensing/readout technology, capable of providing chemical information, mapped based on their miniaturisability and suitability to be operated in complex medium. As can be seen, electrochemical techniques are generally miniaturisable. The overarching miniaturisability limit across electrochemical techniques (i.e. amperometry, potentiometry, and ion- /chemical-selective field-effect transistor) is the need to incorporate a separate/internal filling solution for stable reference electrodes. However, there are well-characterised solid-state materials, that can act as a pseudo- or quasi- reference electrode [37]–[39]. The different electrochemical techniques have different inherent selectivity, however, generally, their suitability to operate in complex matrices can be tuned by modifying the electrode’s surface (in terms of material, morphology, or coating).
Figure 1.1. Different domains of real-time sensing technologies for chemical analysis mapped based on their miniaturisability and suitability for operation in a complex medium. The axes depict increasing miniaturisability or suitability going from – to +.
The optical sensor domain is perhaps the largest domain of on-chip sensing techniques. The domain contains a large number of customisable components that can be combined into new techniques. However, a large part of this domain includes optical labelling which is not practical for continuous on-chip applications. Label-free luminescent indicators would be suited for organ-on-chip applications; however, they are only available for a handful of parameters (e.g. pH, O2, glucose,
6
spectroscopy, surface plasmon resonance, etc.) typically requires light source or an optical setup around the chip that is expensive/difficult to miniaturise.
The physical domain is split into two clusters in Figure 1.1, since the classical techniques relying on physical transducers are not (yet) miniaturisable (e.g. ion mobility spectrometer and nuclear magnetic resonance spectroscopy). However, a sub-class of techniques were developed to make use of microelectromechanical systems (MEMS) physical transducers, making them highly miniaturisable (e.g. piezoelectronics for quartz microbalance and thermistors for thermometric). That being said, these techniques also rely strictly on a biological recognition layer (e.g. DNA, proteins, aptamers) for their (chemical) selectivity. While such layer does lend an excellent selectivity even in complex media, it usually is a trade-off for versatility.
Overall, there are no clear favourites between electrochemical and optical techniques for organ-on-chips. Both domains have extensive adaptability and versatility. On the one hand, optical equipment is already widely available in many in vitro research labs. However, for many, dedicated extra setup (e.g. incubator) needs to be augmented for continuous monitoring using the optical equipment. Electrochemical techniques are generally easier to miniaturise and require cheaper equipment. Therefore, for online measurements, electrochemical methods are more practical to integrate on-chip. It is also worth mentioning that the electrochemical operations are vastly interchangeable, where many different techniques can be operated using the same setup. This gives an extra incentive to integrate an electrochemical setup in a versatile organ-on-chip platform.
1.2.2. Requirements for organ-on-chip sensors
As mentioned, many different applications for organ-on-chip sensors have emerged. The different applications naturally pose different sensing requirements, e.g. in terms of time resolution, spatial resolution, the number of parameters monitored, and length of measurements. Therefore it is unlikely that one sensing solution can be employed for all organ-on-chip needs. Nonetheless, a number of common characteristics among successful organ-on-chip sensors can be identified. The most-cited works on organ-on-chip sensors tend to present robust, regenerable, and versatile/multi-analyte sensors [32], [41]–[44]. All of the mentioned works also relied for the most part on electrochemical sensors, which testifies for the compatibility of electrochemical sensors in organ-on-chip settings.
Research aim
7
1.3.
Research aim
This research aimed to develop sensors that are suitable for organ-on-chips. By expanding the library of organ-on-chips sensors, we hope to increase the information yield from existing organ-on-chip studies, as well as to enable new studies to be conducted. Each sensor in this work was developed and tested for specific applications/objectives. However, generally, we aimed to develop real-time chemical sensors that are practical to use in organ-on-chip research, and that are also robust, reusable/regenerable, and versatile.
1.4.
Framework
This project is a part of the European Research Council (ERC) advanced grant awarded to Prof. Dr. Albert van den Berg titled ‘Vascular Engineering on-chip using
differentiated Stem Cells’ (more commonly known as ‘VESCEL’, grant no. 669768).
The VESCEL project as a whole aimed to create and apply technologies for vasculature-on-chip models. The research for this specific work package was conducted primarily in BIOS Lab-on-a-Chip group at the MESA+ Institute of Nanotechnology and Technical Medical Centre of the University of Twente.
1.5. Thesis outline
The thesis presents the works on the development and applications of organ-on-chip sensors. The sensors are all based on ruthenium oxide (RuOx) nanorods, operated using different electrochemical techniques. The fabrication process of the RuOx electrode is described and discussed in Chapter 2.
Chapter 3 presents the potentiometric pH sensing characterisation of the
developed RuOx electrode, with a focus on its applicability for organ-on-chip studies. In Chapter 4, the amperometric oxygen sensing of the RuOx electrode was
characterised. Together with its pH sensing capabilities, the electrode is presented as a tool for in vitro cell metabolism studies. Its usage is demonstrated in inferring the different types of metabolism of human pluripotent stem cells-derived cardiomyocytes.
The RuOx electrode was also tested as an amperometric nitric oxide (NO) sensor in Chapter 5. This work used a NO chemical donor in testing the electrode. Therefore, the NO generation and the electrochemical oxidation of the generated NO were first confirmed using real-time electrochemical/mass-spectrometry. Finally, we applied the RuOx electrode to detect NO generated by endothelial cell culture.
8
Chapter 6 shows another use of the RuOx electrode in monitoring the
mechanical contractile events of cardiomyocytes using electrochemical impedance spectroscopy. Chapter 7 addresses the demands on the miniaturisation and on-chip
integration of the RuOx electrodes. Results obtained with miniaturised (pH) electrodes are presented, based on which the limits and considerations on miniaturisation are discussed. On-chip integrated (NO) electrodes were also fabricated and tested. The transferability of the RuOx electrode and its fabrication methods are discussed.
Finally, Chapter 8 summarises the main findings of the entire body of work and
discusses the outlook of the project.
1.6.
References
[1] H. H. Kyu, D. Abate, K. H. Abate, S. M. Abay, C. Abbafati, N. Abbasi, et al., “Global, regional, and national disability-adjusted life-years (DALYs) for 359 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017,” Lancet, vol. 392, no.
10159, pp. 1859–1922, Nov. 2018.
[2] J. P. Bunker, “The role of medical care in contributing to health improvements within societies,” Int. J. Epidemiol., vol. 30, no. 6, pp. 1260–1263, 2001.
[3] N. Hassoun, “The Global Health Impact Index: Promoting Global Health,” PLoS One, vol. 10,
no. 12, pp. 1–16, 2015.
[4] J. W. Scannell, A. Blanckley, H. Boldon, and B. Warrington, “Diagnosing the decline in pharmaceutical R&D efficiency,” Nat. Rev. Drug Discov., vol. 11, no. 3, pp. 191–200,
Mar. 2012.
[5] F. Steward and G. Wibberley, “Drug innovation - What’s slowing it down?,” Nature, vol.
284, no. 5752, pp. 118–120, 1980.
[6] M. Weatherall, “An end to the search for new drugs?,” World Health Forum, vol. 3, no. 4, pp.
409–413, 1982.
[7] L. A. Low, C. Mummery, B. R. Berridge, C. P. Austin, and D. A. Tagle, “Organs-on-chips: into the next decade,” Nat. Rev. Drug Discov., Sep. 2020.
[8] S. M. Paul, D. S. Mytelka, C. T. Dunwiddie, C. C. Persinger, B. H. Munos, S. R. Lindborg, et al., “How to improve R&D productivity: the pharmaceutical industry’s grand challenge,”
Nat. Rev. Drug Discov., vol. 9, pp. 203–214, 2010.
[9] J. A. DiMasi, H. G. Grabowski, and R. W. Hansen, “Innovation in the pharmaceutical industry: New estimates of R&D costs,” J. Health Econ., vol. 47, pp. 20–33, 2016.
[10] N. Shanks, R. Greek, and J. Greek, “Are animal models predictive for humans?,” Philos. Ethics, Humanit. Med., vol. 4, no. 1, pp. 1–20, 2009.
References
9 [11] M. J. Prescott and K. Lidster, “Improving quality of science through better animal welfare:
The NC3Rs strategy,” Lab Anim. (NY)., vol. 46, no. 4, pp. 152–156, 2017.
[12] A. D. van der Meer and A. van den Berg, “Organs-on-chips: breaking the in vitro impasse,”
Integr. Biol., vol. 4, no. 5, p. 461, 2012.
[13] M. J. Gómez-Lechón, R. Jover, T. Donato, X. Ponsoda, C. Rodriguez, K. G. Stenzel, et al., “Long-term expression of differentiated functions in hepatocytes cultured in three-dimensional collagen matrix,” J. Cell. Physiol., vol. 177, no. 4, pp. 553–562, 1998.
[14] D. Huh, G. A. Hamilton, and D. E. Ingber, “From 3D cell culture to organs-on-chips,” Trends Cell Biol., vol. 21, no. 12, pp. 745–754, Dec. 2011.
[15] S. N. Bhatia and D. E. Ingber, “Microfluidic organs-on-chips,” Nat. Biotechnol., vol. 32, no. 8,
pp. 760–772, Aug. 2014.
[16] D. Huh, B. D. Matthews, A. Mammoto, M. Montoya-Zavala, H. Y. Hsin, and D. E. Ingber, “Reconstituting Organ-Level Lung Functions on a Chip,” Science (80-. )., vol. 328, no. 5986,
pp. 1662–1668, Jun. 2010.
[17] L. G. Griffith and M. A. Swartz, “Capturing complex 3D tissue physiology in vitro,” Nat. Rev. Mol. Cell Biol., vol. 7, no. 3, pp. 211–224, Mar. 2006.
[18] J. H. Sung, J. Yu, D. Luo, M. L. Shuler, and J. C. March, “Microscale 3-D hydrogel scaffold for biomimetic gastrointestinal (GI) tract model,” Lab Chip, vol. 11, no. 3, pp. 389–392, 2011.
[19] M. J. Wilmer, C. P. Ng, H. L. Lanz, P. Vulto, L. Suter-Dick, and R. Masereeuw, “Kidney-on-a-Chip Technology for Drug-Induced Nephrotoxicity Screening,” Trends Biotechnol., vol. 34,
no. 2, pp. 156–170, 2016.
[20] H. J. Kim, D. Huh, G. Hamilton, and D. E. Ingber, “Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow,” Lab Chip, vol.
12, no. 12, pp. 2165–2174, 2012.
[21] A. Sontheimer-Phelps, B. A. Hassell, and D. E. Ingber, “Modelling cancer in microfluidic human organs-on-chips,” Nat. Rev. Cancer, vol. 19, no. 2, pp. 65–81, 2019.
[22] P. F. Costa, H. J. Albers, J. E. A. Linssen, H. H. T. Middelkamp, L. Van Der Hout, R. Passier, et al., “Mimicking arterial thrombosis in a 3D-printed microfluidic: In vitro vascular model based on computed tomography angiography data,” Lab Chip, vol. 17, no. 16, pp. 2785–
2792, 2017.
[23] K. H. Benam, S. Dauth, B. Hassell, A. Herland, A. Jain, K. J. Jang, et al., “Engineered in vitro disease models,” Annu. Rev. Pathol. Mech. Dis., vol. 10, pp. 195–262, 2015.
[24] G. Wang, M. L. McCain, L. Yang, A. He, F. S. Pasqualini, A. Agarwal, et al., “Modeling the mitochondrial cardiomyopathy of Barth syndrome with induced pluripotent stem cell and heart-on-chip technologies,” Nat. Med., vol. 20, no. 6, pp. 616–623, 2014.
[25] T. Faal, D. T. T. Phan, H. Davtyan, V. M. Scarfone, E. Varady, M. Blurton-Jones, et al., “Induction of Mesoderm and Neural Crest-Derived Pericytes from Human Pluripotent
10
Stem Cells to Study Blood-Brain Barrier Interactions,” Stem Cell Reports, vol. 12, no. 3, pp.
451–460, 2019.
[26] N. M. Cruz, X. Song, S. M. Czerniecki, R. E. Gulieva, A. J. Churchill, Y. K. Kim, et al., “Organoid cystogenesis reveals a critical role of microenvironment in human polycystic kidney disease,” Nat. Mater., vol. 16, no. 11, pp. 1112–1119, 2017.
[27] S. H. Blumenrath, B. Y. Lee, L. Low, R. Prithviraj, and D. Tagle, “Tackling rare diseases: Clinical trials on chips,” Exp. Biol. Med., vol. 245, no. 13, pp. 1155–1162, 2020.
[28] C. Mummery, A. van de Stolpe, B. A. J. Roelen, and H. Clevers, Eds., “Human Stem Cells for Organs-on-Chips,” in Stem Cells, Second Edi., Boston: Elsevier, 2014, pp. 343–361.
[29] A. Van Den Berg, C. L. Mummery, R. Passier, and A. D. Van der Meer, “Personalised organs-on-chips: functional testing for precision medicine,” Lab Chip, vol. 19, no. 2, pp.
198–205, 2019.
[30] C. Luni, E. Serena, and N. Elvassore, “Human-on-chip for therapy development and fundamental science,” Curr. Opin. Biotechnol., vol. 25, pp. 45–50, Feb. 2014.
[31] J. H. Sung, Y. I. Wang, N. Narasimhan Sriram, M. Jackson, C. Long, J. J. Hickman, et al., “Recent Advances in Body-on-a-Chip Systems,” Anal. Chem., vol. 91, no. 1, pp. 330–351,
2019.
[32] Y. S. Zhang, J. Aleman, S. R. Shin, T. Kilic, D. Kim, S. A. Mousavi Shaegh, et al., “Multisensor-integrated organs-on-chips platform for automated and continual in situ monitoring of organoid behaviors,” Proc. Natl. Acad. Sci., vol. 114, no. 12, pp. E2293–E2302,
2017.
[33] M. W. van der Helm, O. Y. F. Henry, A. Bein, T. Hamkins-Indik, M. J. Cronce, W. D. Leineweber, et al., “Non-invasive sensing of transepithelial barrier function and tissue differentiation in organs-on-chips using impedance spectroscopy,” Lab Chip, vol. 19, no. 3,
pp. 452–463, 2019.
[34] J. Kieninger, A. Weltin, H. Flamm, and G. A. Urban, “Microsensor systems for cell metabolism-from 2D culture to organ-on-chip,” Lab Chip, vol. 18, no. 9, pp. 1274–1291,
2018.
[35] D. Bovard, A. Sandoz, K. Luettich, S. Frentzel, A. Iskandar, D. Marescotti, et al., “A lung/liver-on-a-chip platform for acute and chronic toxicity studies,” Lab Chip, vol. 18, no.
24, pp. 3814–3829, 2018.
[36] H. M. U. Farooqi, M. A. U. Khalid, K. H. Kim, S. R. Lee, and K. H. Choi, “Real-time physiological sensor-based liver-on-chip device for monitoring drug toxicity,” J. Micromechanics Microengineering, vol. 30, no. 11, 2020.
[37] P. Bergveld, A. Van Den Berg, P. D. Van Der Wal, M. Skowronska-Ptasinska, E. J. R. Sudhölter, and D. N. Reinhoudt, “How electrical and chemical requirements for REFETs may coincide,” Sensors and Actuators, vol. 18, no. 3–4, pp. 309–327, Jul. 1989.
References
11 [38] A. van den Berg, A. Grisel, H. H. van den Vlekkert, and N. F. de Rooij, “A micro-volume open liquid-junction reference electrode for pH-ISFETs,” Sensors Actuators B. Chem., vol. 1,
no. 1–6, pp. 425–432, 1990.
[39] A. van den Berg, A. Grisel, E. Verney-Norberg, B. H. van der Schoot, M. Koudelka-Hep, and N. F. de Rooij, “On-wafer fabricated free-chlorine sensor with ppb detection limit for drinking-water monitoring,” Sensors Actuators B. Chem., vol. 13, no. 1–3, pp. 396–399, 1993.
[40] P. Gruber, M. P. C. Marques, N. Szita, and T. Mayr, “Integration and application of optical chemical sensors in microbioreactors,” Lab Chip, vol. 17, no. 16, pp. 2693–2712, 2017.
[41] P. M. Misun, J. Rothe, Y. R. F. Schmid, A. Hierlemann, and O. Frey, “Multi-analyte biosensor interface for real-time monitoring of 3D microtissue spheroids in hanging-drop networks,” Microsystems Nanoeng., vol. 2, no. March, 2016.
[42] D. Bavli, S. Prill, E. Ezra, G. Levy, M. Cohen, M. Vinken, et al., “Real-time monitoring of metabolic function in liver-onchip microdevices tracks the dynamics of Mitochondrial dysfunction,” Proc. Natl. Acad. Sci. U. S. A., vol. 113, no. 16, pp. E2231–E2240, 2016.
[43] S. E. Eklund, R. M. Snider, J. Wikswo, F. Baudenbacher, A. Prokop, and D. E. Cliffel, “Multianalyte microphysiometry as a tool in metabolomics and systems biology,” J. Electroanal. Chem., vol. 587, no. 2, pp. 333–339, 2006.
[44] A. Weltin, K. Slotwinski, J. Kieninger, I. Moser, G. Jobst, M. Wego, et al., “Cell culture monitoring for drug screening and cancer research: A transparent, microfluidic, multi-sensor microsystem,” Lab Chip, vol. 14, no. 1, pp. 138–146, 2014.
CHAPTER 2
FABRICATION OF RUTHENIUM OXIDE
NANORODS
Ruthenium oxide (RuOx) is a suitable material for sensing electrode, for its high electrical conductivity, stability, and high catalytic properties. Its biocompatibility also makes it suited for organ-on-chip applications. Nanostructured RuOx can be achieved by a facile heating protocol. This chapter details a similar fabrication method, applied to modify electrodes on glass chips. Its growth process is discussed, as well as the effects of several growth parameters. Finally, electrochemical characterisations show the resulting RuOx nanorods electrodes to be robust and inert, making them suitable for sensing purposes.
2.1. Introduction
In this thesis, the sensing characteristics and applications of ruthenium oxide (RuOx) nanorods electrodes are discussed. RuOx is a material befitting for sensing electrodes for its high electrical conductivity, as well as its high chemical and thermal stability. On top of that, it has also been reported to be highly catalytic for different redox reactions, and it is biocompatible. [1]–[3]
Different ways to synthesise RuOx electrode have been reported [3], several of which include the formation of nanostructures. Nanostructured electrodes are advantageous for sensing applications because of their increased surface area and the ability to adjust the surface’s steric hindrance, which can benefit the sensor’s selectivity [3]–[5]. One of the most practical methods with lowest-cost in achieving nanostructured RuOx electrode was reported by Chen et al [6], where RuOx nanorods were grown from a simple protocol of heating Ru(OH)3 precursors on a substrate. This
chapter describes the process of applying the RuOx nanorods onto on-chip electrodes using a similar protocol.
While the physical process of the RuOx nanorods’ growth is briefly addressed in this chapter, it is not the intention to scrutinise this growth process. The understanding of the process was applied to explore several experimental parameters to end up with a RuOx nanorods electrode that is favourable for our sensing purposes. Finally, the final design/protocol used in the upcoming chapters is presented, as well as the electrochemical characteristics of the developed electrode.
2.2. Fabrication process
2.2.1. Precursor preparation
The precursor was precipitated from RuCl3.xH2O powder (Sigma Aldrich). The
powder was dissolved in ultrapure deionised (DI) water (PURELAB flex, Elga) to make a 5 mM RuCl3 solution. The solution’s pH was measured using a Toledo Mettler-Toledo
SevenMulti pH meter. Due to the hydrolysis of RuCl3.xH2O, the pH of the solution
dropped to around 2.1 (at 20.5°C).
In a glass beaker, 10 mL of the 5 mM RuCl3 solution was continuously stirred
(using a magnetic stirrer) and its pH monitored. A solution of 5 mM NaOH (Sigma Aldrich) was then added into the glass beaker drop by drop, at a rate of 1 drop every ~2 seconds. The formation of precipitate was checked after every 1 mL NaOH addition. Typically, after around 25 mL of 5 mM NaOH was added (pH ~4.0), a noticeable amount of precipitate was formed. At this point, no further NaOH was added, and the solution was let to settle for 30 minutes unstirred.
Fabrication process
A 1 mL aliquot of the suspension was then transferred into a centrifuge tube and was rinsed with DI water three times. First, the aliquot was centrifuged at 2000 rpm for 10 minutes. The supernatant was then pipetted out, and 1 mL DI water was added back into the tube. After the rinsing steps, the separated solid precursor was weighed. The precursor was then resuspended in DI water to make approximately 5 mg/mL concentration. The suspension was then split into aliquots and kept at room temperature until further use.
2.2.2. Chip fabrication
The chips were fabricated using photolithography and thin-film technologies, as illustrated in Figure 2.1. A layer of positive photoresist was spun onto a cleaned glass substrate. Using photolithography, the photoresist was then patterned according to the designed mask (Figure A.1), to expose the electrode design. The substrate was then etched in buffered HF to create recessed facets into which the electrode material was then sputtered (with a thin layer of adhesion metal, typically Ti or Ta). Pt, C, W, Ti, and Ru were considered as electrode material. The excess sputtered material was removed from the chip by lift-off in acetone. When necessary, the steps (from step 2 on) were repeated to pattern a second electrode material. The resulting electrodes should have the same height as the substrate around it, making the chip relatively flat overall.
Figure 2.1. Illustration of the process flow in fabricating flat electrodes on a glass substrate. Illustrated layers thickness and electrode dimensions are not to scale.
2.2.3. Chip modification
Freshly-fabricated chips were modified directly. Tungsten and multi-layered graphene chips were used without prior cleaning. Platinum chips older than one-month-old were first electrochemically cleaned before being used as a substrate. Pt working electrodes (WE) were cleaned by potential cycling in 0.5 M H2SO4 (5 cycles
between -1 and 2 V at 200 mV/s scan rate (SR), ending in 2 V; followed by 20 cycles between -0.2 and 1.2 V at 100 mV/s SR, ending in 1.2 V). The voltammograms were typically very repeatable after the two steps; otherwise, the cleaning cycles (between -0.2 and 1.2 V at 100 mV/s) were repeated. The potential sweeps were performed using
a Bio-Logic SP300 bipotentiostat, measured against a liquid junction Ag/AgCl reference electrode (CH Instruments). The chip was then rinsed with DI water, blown dry with N2, and used as RuOx substrate.
A tube of the prepared precursor suspension was again quickly vortexed before use. Typically, 3 μL of the vortexed suspension was carefully pipetted onto the cleaned and dried electrode (ø 2.4 mm). The chips were then left to dry in room temperature (at least for 30 minutes) and then heated in a preheated oven at 350oC for 4 hours. The
chips were left to cool overnight inside the oven until it reached room temperature.
2.2.4. Characterisation
The modification was imaged using scanning electron microscopy (SEM, using FEI Sirion HR-SEM), under ultra-high-resolution mode. Electrochemical tests were performed using a Teflon chip holder made in-house and operated by a Bio-Logic SP300 bi-potentiostat. All tests were conducted in a closed Faraday cage, at 20±1°C. All potentials were measured against liquid junction Ag/AgCl reference electrode (CH Instruments).
2.3.
RuOx nanorods growth
The growth of the RuOx nanorods from Ru(OH)3 precursor has been reported and
investigated in several different works [5]–[8]. All reported evidence point toward diffusion of the amorphous Ru(OH)3 precursors over the substrate surface. The
precursor then nucleated and grew into crystalline RuO2 nanorods to minimise its
surface energy.
A similar mechanism was observed in our fabrication process. SEM images showed the precursor (Figure 2.2a) formed a thin layer of discontinuous island-like structures when spread on a substrate, as also observed by Chen et al. [6] A higher magnification (Figure 2.2b) showed that the precursor had a grain size of 10-15 nm in diameter.
After heating at ≥300°C for ≥2 hours, rod-like structures were formed. Energy dispersive x-ray analyses (HRSEM-EDX, Oxford Instruments) were performed on the precursor, as precipitated (Figure A.2 in appendix) and after heating at 300°C for 3 hours (Figure A.3 in appendix). The Ru:O ratio changed from 1:3.8 to 1:1.7, going from unheated to heated. This is consistent with the expected Ru(OH)3 to RuO2
transformation taking place during the heating process. Despite similar growth mechanism, several parameters influenced the growth rate and growth behaviour of the nanorods, resulting in a range of nanorods morphology, size, and substrate coverage. These parameters and their influences are discussed separately below.
RuOx nanorods growth
Figure 2.2. (a) SEM image of the Ru(OH)3 precursor on a silicon substrate. (b) A higher
magnification SEM image shows the grain size of the Ru(OH)3 precursor.
2.3.1. Heating temperature
Considering the growth process, the heating temperature was thought to be an important parameter as it brings the system to an unstable thermodynamic state and influences the diffusion rate. The findings of Chen et al. [6] also showed that heating temperature heavily influenced the growth of the nanorods. Growth at temperatures ≤300°C was shown to be limited to more local nucleation processes. As the diffusion rate increased in temperatures higher than 300°C, nuclei or smaller crystals could redissolve and redeposit onto larger crystals (termed Ostwald ripening). This process tends to result in fewer, but larger crystals.
Our experimentally grown nanorods, on the other hand, did not seem to be consistently influenced by the heating temperature. In Figure 2.3a-b, SEM images of two samples of RuOx grown on Pt electrodes can be seen. Both samples were heated at the same temperature (of 350°C) for 3 hours, yet, different RuOx nanorods growth were achieved.
Figure 2.3a shows RuOx nanorods of 134±20 nm long and 24±5 nm wide, which grew mostly on the precursor islands and select (darker) areas on the substrate. Figure 2.3b shows fewer, but bigger RuOx nanorods of 402±73 nm long and 51±12 nm wide, which grew mostly on the substrate. This pointed to the occurance of Ostwald ripening on one substrate and not the other, despite identical heating protocol. The two Pt chips were prepared the same way; however, they were fabricated in different batches at different time. The Pt substrate imaged in Figure 2.3a was around two years old when modified, while the substrate of Figure 2.3b was one month old. None of the newly-fabricated electrodes showed similar darker film, and its origin could not be identified. Since other experiments in studying different heating temperatures also did not deliver consistent results, it was concluded that substrate properties are the
more defining parameter for RuOx nanorods growth when compared to heating temperature.
Figure 2.3. (a) SEM image of grown RuOx nanorods on two-years-old Pt electrode after heated at 350°C for 3 hours. (b) SEM image of another sample of grown RuOx nanorods on one-month-old Pt electrode, also after heated at 350°C for 3 hours.
2.3.2. Substrate
During the growth process, the substrate properties seem to be most influential during the nucleation process. According to general nucleation theories, there are several (bulk) properties of the substrate (e.g. surface tension) that can influence the probability of nucleation [9]. However, as can be seen in Figure A.4a-b (and Figure 2.3), nanorods which grew on areas next to each other can grow to have very different sizes and shapes. This suggested that the local surface properties are more determinant for the nucleation process than the material’s bulk material properties. Therefore, on top of its material choice, other substrate’s properties (among others (atomic scale) roughness, defects, crystallographic orientation) can influence the growth of the nanorods greatly. Overall, this makes theoretical predictions of favoured/suitable substrate material (and its fabrication method) challenging. Consequently, we experimented with several substrate materials and fabrication methods from available techniques and took inspirations from other published works [4], [5].
Out of several substrates tested, the RuOx nanorods grew exceptionally well on carbon substrates (graphite and multi-layer graphene) and sputtered W electrodes (see appendix section A.4 for SEM images). The nanorods covered both the substrates as well as the precursor islands on most of the electrode. Some variations were observed on the edges of the electrodes, with lower coverage as well as different morphologies (Figure A.5c-d). Interestingly, the RuOx nanorods on the precursor islands also seemed to be influenced by the substrate (Figure 2.4, Figures A.4-A.7). RuOx nanorods which grew on precursor islands on Pt substrates tend to have more varied dimensions within the same area, on W substrates longer nanorods were
RuOx nanorods growth
observed, and on carbon substrates, the nanorods tend to grow wider. Since no other works have discussed these findings, and given the high degree of variation we observed, further studies are necessary to arrive at a possible explanation.
After all, the decision on the used substrate for sensor applications was based on practicality. While a number of other substrates showed better nanorods growth than Pt, all subtrates tested exhibit very poor robustness. Elemental W is known to grow a passivating oxide layer only in pH ≤ 2, and despite the very dense coverage of RuOx, the electrodes were still unstable in pH > 2. Carbon electrodes were also rendered non-conducting after a few runs of potential cycling in inert solution. Therefore, eventually, Pt electrodes were still decided to be the best substrate material for sensing purposes. SEM images of the specific electrodes used in each testing/application are shown in the respective chapters. However, typically, the RuOx nanorods grew on Pt electrodes as in Figure 2.4a-b.
Figure 2.4. SEM images of RuOx nanorods grown on Pt substrate. (a) RuOx nanorods covered the precursor island well but did not cover the bare substrate. (b) The grown RuOx nanorods have a high variance in their dimensions.
2.3.3. Heating time
Lastly, a parameter that influenced the resulting RuOx nanorods is the heating time. The figures below show SEM images of the same electrode, first after two hours of heating (Figure 2.5a). Small nanorods were seen, indicating the growth was at its initial stage. Figure 2.5b shows SEM image taken after the same electrode heated for another two hours and left to cool in the oven overnight. As expected, bigger nanorods had grown on the precursor islands, with varied dimensions. Most tested RuOx electrodes were heated for four hours and left to cool down in the oven overnight to obtain electrodes with a rougher surface.
Figure 2.5. SEM images of grown RuOx on Pt substrates, after heating at 350°C for (a) 2 hours and (b) 4 hours.
2.4. Electrochemical characteristics
The electrochemical characterisations of the RuOx nanorods on Pt substrate are shown in Figure 2.6 and in appendix section A.5. The RuOx nanorods showed to be a relatively inert electrode between the hydrogen and oxygen evolution window (Figure 2.6a), giving it a potential window for sensing applications between -0.6 and +1.0 V vs Ag/AgCl at pH ~7. Cyclic voltammograms recorded on RuOx in the presence of redox-active ferri-/ferrocyanide (Figure 2.6b) showed highly reversible curves with low over-potential even in high scan rates (tested up to 250 mV/s). This indicates an electro-active electrode with facile electron transfer.
Figure 2.6. (a) CV (SR = 100 mV/s) recorded on the RuOx electrode in inert solution (100 mM KNO3). The arrows note the scan direction. (b) Cyclic voltammograms recorded on RuOx
electrode in redox-active solution (in the presence of 10 mM ferri-/ferrocyanide and 100 mM KNO3) in varied SRs. The arrows note the scan direction.
Conclusion
We also performed staircase voltammetry and cyclic voltammetry using the RuOx electrode in inert solution (Figure A.9) to determine the surface roughness. The disparity between the two voltammograms can be attributed to the capacitive current. Using the capacitive current and RuOx’s specific double-layer capacitance (following similar calculation shown in section D.2), the RuOx nanorods was calculated to typically increase the electrode’s electrochemical surface area by 20-30 times its geometric surface area. Further modifications brought by the RuOx nanorods for each sensing application are presented and discussed in the respective chapters.
2.5. Conclusion
Crystalline RuOx nanorods could be grown by heating Ru(OH)3 precursor on a
substrate. The growth was found to be highly influenced by the surface characteristics of the substrate. Among the tested substrates, Pt electrode was found to be the most practical substrate for sensing purposes. Pt electrodes could be sufficiently modified with RuOx nanorods by heating at 350°C for > 4 hours. Electrochemical characterisations showed conducting and highly active RuOx-modified Pt electrodes, that are suitable for sensing applications.
2.6. Acknowledgements
We would like to express our gratitude to André ten Elshof for his scientific input regarding the nanorods growth process. Special thanks are also extended to Wesley van den Beld for the discussions and the preparation of multi-layer graphene chips. We would also like to thank Johan Bomer and Johnny Sanderink for their advice and assistance in the cleanroom processes. Mark Smithers is acknowledged for the EDX analyses.
2.7. References
[1] D. Galizzioli, F. Tantardini, and S. Trasatti, “Ruthenium dioxide: a new electrode material. I. Behaviour in acid solutions of inert electrolytes,” J. Appl. Electrochem., vol. 4,
no. 1, pp. 57–67, 1974.
[2] M. Brischwein, H. Grothe, J. Wiest, M. Zottmann, J. Ressler, and B. Wolf, “Planar ruthenium oxide sensors for cell-on-a-chip metabolic studies,” Brischwein, M., Grothe, H., Wiest, J., Zottmann, M., Ressler, J., Wolf, B. 2009. Planar ruthenium oxide sensors cell-on-a-chip Metab. Stud. Chem. Anal. Vol. 54, No. 6 1193-1201., vol. 1193, no. Vol. 54, No. 6, pp. 1193–
1201, 2009.
[3] H. Over, “Surface chemistry of ruthenium dioxide in heterogeneous catalysis and electrocatalysis: From fundamental to applied research,” Chem. Rev., vol. 112, no. 6, pp.
3356–3426, 2012.
grown on a single carbon fiber for the real-time direct nitric oxide sensing,” Sensors Actuators, B Chem., vol. 191, pp. 298–304, 2014.
[5] S. J. Kim, Y. K. Cho, J. Seok, N. S. Lee, B. Son, J. W. Lee, et al., “Highly Branched RuO2 Nanoneedles on Electrospun TiO2 Nanofibers as an Efficient Electrocatalytic Platform,”
ACS Appl. Mater. Interfaces, vol. 7, pp. 15321–15330, 2015.
[6] Z. G. Chen, F. Pei, Y. T. Pei, and J. T. M. De Hosson, “A Versatile Route for the Synthesis of Single Crystalline Oxide Nanorods: Growth Behavior and Field Emission Characteristics,” Cryst. Growth Des., vol. 10, no. 6, pp. 2585–2590, 2010.
[7] J. Park, J. W. Lee, B. U. Ye, S. H. Chun, S. H. Joo, H. Park, et al., “Structural Evolution of Chemically- Driven RuO 2 Nanowires and 3-Dimensional Design for Photo- Catalytic Applications,” Nat. Publ. Gr., no. November 2014, pp. 1–10, 2015.
[8] M. Kang, Y. Lee, H. Jung, J. H. Shim, N. S. Lee, J. M. Baik, et al., “Single carbon fiber decorated with RuO 2 nanorods as a highly electrocatalytic sensing element,” Anal. Chem., vol. 84, no. 21, pp. 9485–9491, Oct. 2012.
[9] S. Karthika, T. K. Radhakrishnan, and P. Kalaichelvi, “A Review of Classical and Nonclassical Nucleation Theories,” Cryst. Growth Des., vol. 16, no. 11, pp. 6663–6681, 2016.
CHAPTER 3
RUTHENIUM OXIDE AS pH SENSOR FOR
ORGAN-ON-CHIP
A ruthenium oxide (RuOx) sensor for potentiometric pH sensing is being developed for organs-on-chip purposes. The sensor was fabricated from Ru(OH)3 precursor,
resulting in RuOx nanorods after heating. Open-circuit potential of the RuOx electrode showed near-Nernstian response of -58.05 mV/pH, with good selectivity against potentially interfering ions (lithium, sulfate, chloride, and calcium ions). The preconditioned electrode (stored in liquid) had long-term drift of -0.8 mV/hour and response time was less than 2 seconds. Sensitivity to oxygens was observed, at an order of magnitude lower than other reported metal-oxide pH sensors, causing limited interference (in the order of 0.050 pH unit) at physiological conditions. Together with miniaturizability, the RuOx pH sensor proves to be a suitable pH sensor for organs-on-chip studies.
This chapter is adapted from:
E. Tanumihardja, W. Olthuis, and A. van den Berg, “Ruthenium Oxide Nanorods as Potentiometric pH Sensor for Organ-On-Chip Purposes,” Sensors, vol. 18, no. 9, p. 2901, Sep. 2018.
3.1.
Introduction
Organs-on-chips are new in vitro models of human tissue, where cells are cultured and perfused in microfluidic devices. The devices can be configured to mimic cell’s microenvironment as realistic as possible (in terms of physical/chemical cues, tissue-tissue interfaces, and perfusion). This way, the cells can replicate human physiology and pathology at organ-level, thus providing an improved model to the too-simplistic 2D (or even 3D) cell culture. Application of organ-on-chips can offer an alternative to the animal models that is more ethical, physiologically representative/predictive, and can be personalized. [1]–[3]
Over the years, many different organs-on-chips have been developed [3]. Typically, the cells are cultured in micrometre size chambers or channels, leaving low accessibility to probe or monitor the cells from outside the chip. The small dimensions also mean that very minute amount of medium is involved, making offline sampling also limited. Consequently, microsensors can be of great value in providing readouts for organs-on-chip studies. Miniaturized sensors can be placed close to the cells, enabling accurate non-invasive online monitoring of minute analytes.
This work explores on-chip pH sensors, which are useful for long-term cell-culture/differentiation and tissue studies. Metal oxide is a well-known class of on-chip pH sensors, for its robust, inert, and miniaturizable properties. The metal oxides (commonly investigated are SbO2, TiO2, CuO, IrO2, and RuO2) show Nernstian response
to proton concentration [4]. However, metal oxide electrodes often suffer from sensitivity to other oxidizing/reducing agents [4], [5]. An especially critical and unavoidable oxidizing agent for organs-on-chip studies is dissolved O2, where it is
often an operational parameter. Studies on IrO2, for instance, have reported potential
response as much as 80 mV going from deoxygenated to oxygenated buffer [6], [7]. Ruthenium oxide (RuOx), however, might be an exception to this shortcoming [4], [8] and is therefore the focus of this work.
RuOx is proton sensitive due to reversible redox equilibrium between two different solid phases/oxidation states of the metal oxides, in which proton is involved [8]–[10]:
RuO2 · 2 H2O + H+ + e- ↔ H2O + Ru(OH)3 (eq. 3.1)
An adsorbed proton leads to the attraction of an electron through this conducting oxide, resulting in the reduction of the Ru ion. This leads to a net potential change, generating measurable change in open-circuit potential.
Recent years have seen many works on RuOx pH sensor for different applications. Table 3.1 summarizes recent works on RuOx pH sensor, their results, and applications. While pH sensitivity, range, drift, and response time are commonly
Introduction
characterized, very few touched on oxygen sensitivity. Lonsdale et al. [11] described elimination of oxygen sensitivity through addition of Ta2O5 layer on RuO2. Xu et al.
[12] showed high resemblance between calibration curves of RuO2-MWCNTs
measuring in oxygen-saturated and nitrogen-saturated buffers. No work so far has characterized oxygen sensitivity of RuO2 in organs-on-chip configuration.
Table 3.1. Summary of recent works on RuO2 pH sensors, their results, and applications.
Material pH range Sensitivity (mV/pH) Response time (s) Drift (mV/hr) Application Ref RuO2-Ta2O5 - -55.3 5-136 7.2 pH of common beverages [11] RuO2 1-10 -77.74 <20 93.3 Helicobacter pylori detection [13] RuO2 2-11 -56 60-120 - - [14] Pt-doped RuO2 2-13 -58 1-2 0.002 Water quality monitoring [15] RuO2-MWCNTs 2-12 -55 <40 - - [12] RuO2 nanorods 2-10 -58 <2 -0.8 Organs-on-chip This work
- : not (quantitatively) characterized/specified
In this work, the RuOx electrode was fabricated from Ru(OH)3 precursor, which
was precipitated from Ru3+ and OH- ions. The precursor was heated to form RuO 2 with
nanorods morphology. The same electrode is intended to be developed into a dual sensor: potentiometric pH sensor and amperometric nitric oxide sensor, where the nanorods morphology is desired for the latter. Therefore, the same RuOx nanorods are studied as pH sensor, although the morphology does not theoretically improve the potentiometric signal.
This contribution evaluates the suitability of RuO2 nanorods as an on-chip pH
sensor in novel organs-on-chip settings. Its drift behavior, selectivity, response time, and oxygen sensitivity are presented and discussed. First organs-on-chip application for the RuOx pH sensor envisions hypoxia study of cardiomyocytes. In hypoxic condition, cardiomyocytes undergo anaerobic glycolysis which results in pH drop as much as 1-2 pH unit. During the study, oxygen level can be expected to be varied from fully oxygenated to deoxygenated. Applicability of the sensor is currently assessed with this application in mind.
3.2. Materials and Methods
3.2.1.
Electrode fabrication
RuOx electrode was fabricated according to the protocol of Chen et al. [16] RuCl3.xH2O (Aldrich, 99.98%) was dissolved in de-ionized water to make 10 mL of 5
mM RuCl3 solution. 5 mM NaOH (Aldrich, 98%) solution was then added drop by drop
to the solution, until solid Ru(OH)3 precursor precipitated. Precipitation typically
occurred around pH 4 (from initial pH of around 2). The precursor was then isolated by centrifugation and suspended in DI water. The resuspension was spread on clean substrate (sputtered circular Pt electrode, 2.4 mm in diameter, on glass chip), left to dry in room temperature, and then heated to 350oC in a preheated oven (atmospheric
environment) for 3 hours. RuO2 nanorods were formed after the heat treatment, as
confirmed by scanning electron microscopy (SEM) imaging.
Pt electrodes on glass chips were first electrochemically cleaned before used as RuOx substrate by applying cyclic potential sweeps (20 times with scan rate 100 mV/s in 0.5 M H2SO4 between -0.6 and 1 V (vs. Ag/AgCl), ending in 1 V).
3.2.2. Experimental setup
The modified chip with grown RuOx nanorods was used for most experiments, unless stated otherwise. The glass chip was used with a Teflon chip holder, made in-house. The chip holder exposes the electrodes active area to the electrolyte chamber. All potentials were measured against liquid-junction Ag/AgCl (satd. KCl) reference electrode (CH Instruments), typically placed ~5 mm from the RuOx electrode. The setup was placed inside a Faraday cage during all measurements. All measurements were carried out using Bio-Logic SP300 bipotentiostat (input impedance of > 100 GΩ) and were performed in around 22oC room temperature.
3.2.3. Measurement protocol
Open-circuit potential measurements were performed in buffer solutions. pH 4 used potassium hydrogen phthalate buffer, pH 5 acetic acid/sodium acetate buffer, pH 7 disodium hydrogen phosphate/potassium hydrogen phosphate buffer, pH 8 phosphate buffered saline, and pH 10 boric acid/sodium hydroxide buffer. The pH of the solutions was confirmed with a Mettler-Toledo SevenMulti pH meter, calibrated with standard buffer solutions of pH 4, 7, and 10.
For pH response experiment, open-circuit potential was recorded for 150 s at 1 s interval. The last point was taken as the potential response to determine the pH sensitivity. Interference experiments were done in air-saturated pH 7.4 PBS buffer. Ions of different concentrations were dissolved in the buffer and measured.
Results
Longer measurements in air saturated pH buffer were done for drift experiment. Drift was calculated using the linear-fit slope of the potential over the measurement period.
Similar recording was also performed for the response time experiment. The open-circuit potential was recorded in, initially, 5 mL stirred solution of 20 mM KH2PO4, 137 mM NaCl, and 5 mM KCl. The solution was stirred with a magnetic stirrer
hovering above the electrode. Over time, different amounts of 0.5 M K2HPO4 were
added (by pipetting) to change proton concentration by 0.6 pH units (except for the first addition, which resulted in change of 1.2 pH units due to the unbuffered nature of the initial solution). Response time was deduced by linear fitting the slope and calculating the intersecting time points between the stable potential plateau and the slope.
Oxygen sensitivity measurement was done in pH 7 buffer which had been purged with N2 gas for 1 hour. Open-circuit measurement began with the buffer still purged
with N2 gas. After ~4 hours of measurement, the N2 gas was stopped and the buffer
solution was purged with instrument air (compressed filtered air free of contaminates) instead. The measurement was left to record for another ~4 hours.
3.3. Results
3.3.1. Fabrication and characterization
Every heat-annealed RuOx was imaged by SEM to confirm the creation of nanorods. An example of the produced nanorods sample is given in Figure 3.1b. As a comparison, Figure 3.1a shows the freshly precipitated precursor on silicon. As can be seen, the heat treatment did not convert all amorphous precursor into nanorods. The conversion seemed to depend on few different parameters, most notably are the substrate material and cleanness, as well as presence of organics during the heat treatment. On (electrochemically) cleaned Pt substrate and in absence of organics during heat treatment, nanorods typically grew (rather sparingly) on the precursor, as shown in Figure 3.1b. Most of nanorods are around 50 nm wide and 200-350 nm long, forming pointy rod shape as its width tapers along the length.
RuO2 stoichiometry was confirmed using energy-dispersive X-ray spectroscopy
(EDX). EDX measurement of annealed RuOx with visible nanorods morphology sample (shown in appendix, Figure A.2) revealed Ru to O ratio of approximately 1:2.
Figure 3.1. SEM images of the RuOx electrode: (a) Precipitated Ru(OH)3 precursor, isolated and
spread on silicon surface; (b) Heat-treated precursor (at 350oC, for 3 hours) on platinum electrode
surface formed RuO2 nanorods with width of around 50 nm, and length of 200-350 nm.
3.3.2. pH sensitivity and selectivity
Open-circuit potential of the RuOx electrode as a function of pH is shown in Figure 3.2. Figure 3.2a shows the typical potential response in air-saturated pH buffers from pH 4-10. Mean of three different measurements (using three different electrodes) is plotted, with the error bar showing one standard deviation among the data. Slope of linear fitting through the measurement points indicates the electrode pH sensitivity of -58.1 ± 1.2 mV/pH, with extrapolated E0 of 736 ± 18 mV.
Figure 3.2. Experimental results of RuOx electrode: (a) Open-circuit potential of RuOx electrode in different pH buffer solutions. Error bar shows one standard deviation out of three different sets of measurement. Largest cross-sensitivity, towards lithium ions, is plotted as comparison. Dashed lines are linear fitting through the points; (b) Open-circuit potential of RuOx electrode with presence of different ions (lithium, sulfate, chloride ions) in pH 7.4 buffer. Dashed lines are linear fitting through the points.
The RuOx electrode was also tested for its selectivity against possibly interfering ions, namely lithium, sulfate, chloride, and calcium ions. The potential response
Results
towards the different ions is given in Figure 3.2b. Highest sensitivity, taken from the slope of linear fitting through the points, was found towards lithium ions. This response of 1.06 mV/decade [Li+] is also plotted in Figure 3.2a as comparison to the pH
response.
The RuOx electrode was also tested for its pH response in a physiologically relevant milieu for future organs-on-chip applications. A solution comprising of 137 mM NaCl, 5 mM KCl, and 2 mM KH2PO4, which make up the inorganic constituent of
the human body fluids, was used. Different amount of 0.5 M KHPO4 was added to the
solution to vary the pH between 6 and 8, while making sure the solution was relatively isotonic to the human body fluids. An open-circuit potential measurement of a RuOx electrode is plotted in Figure 3.3 as a function of the solution pH. Sensitivity of -58.23 mV/pH was estimated from linear fitting.
Figure 3.3. Open-circuit potential of (preconditioned) RuOx electrode in isotonic PBS buffer with pH range of 6-8. Dashed line is linear fitting through the points.
3.3.3. Drift and aging
Drift behavior of the RuOx electrode was also studied. Freshly prepared or dry-stored electrodes were observed to undergo large short-term drift, which decays into linear long-time drift after 8-10 hours of measurement in a buffer. Preconditioning of the electrode, simply by storing it in liquid, allows it to immediately attain the lesser long-term drift, as shown below in Figure 3.4a.
Open-circuit potential of RuOx electrode was measured over a long period of time in air-saturated pH 7 buffer, firstly after it had been stored in air (dry-stored RuOx graph). The measurement was then repeated with the same electrode after it had been stored in pH 7 buffer overnight (wet-stored RuOx graph). Dry-stored RuOx electrode showed high drift of around -7.5 mV/hour. Wet-stored RuOx electrode showed significantly lower drift of -0.8 mV/hour. The behavior was highly reversible, the wet-stored electrodes again undergo the large drift after drying (dry-blowing with N2).
Furthermore, the wet-stored RuOx electrode performed comparably to the freshly prepared/air-stored RuOx electrode in terms of pH sensitivity, as shown in Figure 3.4b.
Figure 3.4. (a) Open-circuit potential of RuOx recorded over time in pH 7 buffer, fitted linearly (dotted lines). RuOx electrode showed high drift when stored in air (dry-stored) and significantly lower drift when stored in liquid (wet-stored); (b) Comparison of open-circuit potential of wet-stored vs. dry-wet-stored RuOx in different pH buffers, fitted linearly.
At the time of writing, a number of RuOx electrodes have been used for as much as 10 months since their fabrication. The electrodes are stored in DI water for most of the time. So far, no detrimental effects on pH sensitivity, drift, or response time have been noticed from using, rinsing, and pH cycling (between pH 2-10) of the electrodes. Exposure of the electrodes to biological cell culture medium or bovine serum albumin also did not affect electrode’s pH sensing performance. pH response calibration after extensive exposure to culture medium and bovine serum albumin showed the reproducible -58 mV/pH sensitivity.
3.3.4. Response time
Filtered recording of RuOx open-circuit potential is shown in Figure 3.5, during which the solution pH was changed (the first change by 1.2 pH unit, the following by 0.6 pH unit). Due to the introduction of a magnetic stirrer in the Faraday cage during measurement, a noisy signal was acquired. For ease of analysis, a low-pass (cut off at ~40 Hz) and band-stop (cut off around the frequency of magnetic stirrer, between 3 and 3.5 Hz) filters were applied, resulting in the graph shown in Figure 3.5. The raw signal can be seen in appendix, Figure B.3. Analysis of the slopes results in calculated response time between 2 to 3 seconds.
Results
Figure 3.5. Open-circuit potential recording of RuOx electrode in changing pH, low-pass filtered with cut-off frequency of ~40 Hz.
3.3.5. Oxygen sensitivity
Open-circuit potential recording of the RuOx electrode in pH buffer, with changing oxygen level is shown in Figure 3.6. The negatively sloping potential recording in the first 4 hours (linearly fitted with blue dashed line) occurred in deaerated buffer solution, and is of the same order of magnitude as the earlier observed wet-stored drift. A positive slope was observed in the following 4 hours as reoxygenation took place in the buffer. Over the entire 4 hours, the open-circuit potential drifted 3 mV as the buffer went from deoxygenated to oxygenated. Large spikes in the recording were motion artefacts.
Figure 3.6. Open-circuit potential recording of wet-stored RuOx electrode, for the first 4 hours in deaerated pH 7 buffer, followed by 4 hours of reoxygenated pH 7 buffer. Fitted linearly by the dashed lines.
