Final Report Summary - SNOW CONTROL (Integrated Real-Time Measurement Platforms for Nanoparticles and Nanoparticle Thin Films)
Executive Summary:
One of the major barrier towards the successful introduction of nanoparticles (NP) into many applications is the lack of a tight control on their properties, i.e. size, shape, crystallinity, core-shell, functionalization, etc.) This can only be resolved by the simultaneous use of several metrology methods and to extract the relevant information in real-time. In this Snow Control project, we take the first steps towards this goal by developing real-time characterization metrology tools to measure the properties of NP, functionalised NP and NP thin films. The Snow Control project realized two novel real-time integrated characterization systems.
Firstly, an integrated real-time measurement platform linked to an automated NP generation system was developed. To enable a real-time and a complete characterization of NP properties, we reviewed and drastically extended the capabilities of Dynamic Light Scattering (DLS), Zeta Potential (ZP) and Small Angle X-ray Scattering (SAXS). Since there exits no perfect universal single technique for complete NP characterization, these enhanced techniques were combined into a novel integrated real-time NPs measuring platform. Therefore, a new DLS probe (from Cordouan) was designed, which was remotely connected to the laser source and to the laser detectors. The SAXS instrument (from Bruker AXS) was improved and gained a significantly higher X-ray flux, using a liquid Ga metal jet x-ray source. The ZP measurement capability was extended on the basis of electrophoresis, and led to the development of a unique Wallis system (by Cordouan), with increased resolution and accuracy by a factor 3 to 10 over existing solutions. Subsequently, real-time combined measurements were realized, and a setup was successfully developed which allows simultaneous DLS and SAXS measurements under flowing NP conditions. Since it was technically not feasible to perform ZP measurements on these flowing particles, the ZP tool was installed at the end of the synthesis process, to allow for offline measurements. To realize real-time measurements, the measurement time should match the synthesis timescale. A measurement time of 10 seconds was within reach. A successful demonstration of the combined setup was accomplished during the project. Combined SAXS and DLS techniques can follow the size evolution of NPs in real-time.
Secondly, an integrated thin film deposition system for NPs was developed. A novel vacuum system for the deposition of a NP thin film (from DCA) was designed and fabricated. To enable the deposition of NPs in vacuum, the liquid injection/vaporization unit (from Kemstream) was redesigned. The vacuum system was then built with this integrated NP injector, a RHEED (Reflection high-energy electron diffraction) analyzer and a newly designed laser based fluorescence NP flux monitor. NP thin films were successfully grown, as shown by RHEED and GI-SAXS experiments. These results were further confirmed by AFM data.
Systematic characterization of the different NP was carried out using the methods available in the project. But it became clear that some methods are missing, such as UV-VIS absorption and TGA/FTIR. In particular, KUL demonstrated that TGA/FTIR is the only method that can provide a quantitative measure of how much functional (organic) groups are effectively attached to NP.
Finally, the methods developed in this project were evaluated and assessed for further use in industrial manufacturing applications. To do so, the specifications and testing procedure of the different metrology tools, the cost of ownership, the NP standards and the simultaneous measurement protocol were documented and evaluated by the end-user. The main conclusion is that the tools performed as expected and that a combined SAXS/DLS analysis on NPs, in particular during real-time experiments, is certainly feasible. Also, for the first time, NP thin films were grown and monitored from a NP solution.
Project Context and Objectives:
A. Project Context
Nanomaterials are currently already used in a variety of applications in the fields of energy, health, information and communication technology, as well as materials engineering. Among these nanomaterials, the nanoparticles (NP) are a major category and exist in many shapes, sizes and structures. The different shapes are ranging from the spherical single particle, rod like, or the multi-shell nanorice particle to highly branched nanostructure. For most industrial and/or medical applications, a very narrow size & shape distribution is highly desired, to establish toxicity standards for instance. Also small differences in size, shape and crystal structure can have a large effect on the performance, in the field of catalysis for example. Unfortunately, today none of the known methods allows for an accurate determination of the physical and chemical properties of many of the mentioned NPs.
Ensuring that the EU can realize the commercial potential of nanotechnology, industry and society will require reliable and quantitative means of characterization, as well as measurement techniques that will underpin the competitiveness and reliability of future products and services. Metrology and standards need to be developed to facilitate rapid development of the technology as well as to provide users with the necessary confidence in their process and product performance.
This project is the first step towards full feedback control of NP synthesis by developing and integrating dedicated measuring techniques. For instance, if the size of the NPs can be measured in real-time then the multi-shell structure can be resolved during the synthesis. However, it is not the goal of this project to close the feedback loop (i.e. not using the obtained data to manipulate the process), but to focus on building integrated measurements systems that will be tested in realistic industrial conditions. The project provides two unique integrated systems, that will allow for innovative developments in measurement techniques needed to cope with the demands of nanotechnology. The systems allow to develop standard procedures for nanomaterial synthesis, a set of reference materials for property measurement standardization, standard methods for physical and chemical properties evaluation, and finally to develop standards for material evaluation in interdisciplinary applications such as biomaterials, photonics and photovoltaics, nanomedicine, etc.
B. Project Objectives
Our research activities aimed to fulfill five scientific and technology objectives. The first objective aimed to improve the size, shape and temporal resolution of measurement techniques, which already exist today. The goal of the second objective was to develop novel NP characterization methods that do not exist today. The third objective aimed to create a unique combination of the best elements of the first two objectives into 2 different real-time measurement platforms, for NPs and NP thin films, respectively. The fourth objective focused on reproducibility and standardization using the combined measuring methods, while the final objective demonstrated the integration into a two production lines.
More specifically, the project had the following objectives:
Objective 1: Demonstrate DLS, ZP and SAXS with improved size, shape and temporal resolution.
This objective includes an array of multi-optical DLS and ZP heads along the line of a microfluidic reaction path as well as the use of a new liquid Ga metal jet X-ray source for SAXS.
Objective 2: Develop real-time concepts, VUV techniques and probes for novel properties.
This objective includes developing a real-time concept to follow the synthesis of NP at different locations in a flow chemistry line as well as a VUV based NP flux sensor (in a vacuum system).
Objective 3: Integrate different techniques into two real-time characterization platforms.
This objective includes the integration of ZP, DLS and SAXS measurements together along the flow chemistry line to follow in real-time the synthesis of NP. This includes also the real-time measurement of the NP flux in a vacuum system.
Objective 4: Towards real-time multiple measurement standards and standardized NP.
This objective aims to devise standard operating procedures for the preparation of standardized NP and functionalized NP. The objective also includes the establishment of standard operating procedures for the simultaneous measurement and analysis of NP properties obtained via the real-time measurement platform.
Objective 5: Demonstrate integration into a (high throughput) production line.
This objective includes the demonstration of the integrated NP measurement platform on a flow chemistry process as well as the demonstration of the NP thin film measurements on a thin film deposition system.
Project Results:
The Snow Control project realized two novel real-time integrated characterization systems for the measurements of the NP size and shape. One demonstrator system has an integrated real-time measurement platform linked to an automated NP generation system. The other demonstrator is a thin film deposition system for NPs.
The two real-time characterization methods developed work in two automated production cycles. The first cycle uses Dynamic Light Scattering (DLS), Zeta Potential (ZP) and Small angle X-ray scattering (SAXS) to automatically characterize NPs of the selected size, while the second one uses in-line UV flux sensing and RHEED to create NP thin films of selected structure and thickness. Both platforms were demonstrated and evaluated from the perspective of an end-user and potential customer, namely IBM.
Below we will give a concise overview of the work done throughout the whole project. The development of the integrated real-time measurement platform, linked to an automated NP generation system, will be discussed first, followed by the description of the work related to the development of the thin film deposition system for NPs. The work on the 2 systems was performed in parallel throughout the whole project.
1. Development of an integrated real-time measurement platform, linked to an automated NP generation system
We first evaluated and extended current metrology methods, namely DLS, SAXS and ZP. We performed real-time measurements with these extended methods, which were then combined and integrated into one measurement platform. A set of standard procedures were designed and optimized, followed by a demonstration of the reliability and automation of this NP production line. The tools were finally evaluated from an end-user point of view. The outcome of each step will be discussed in the following paragraphs.
Evaluation of current metrology methods
This task addressed the evaluation of the current metrology methods and is materialized by the release of a report, that reviews the most relevant metrology techniques currently in use and identifies their limits with respect to sensitivity, accuracy and time resolution. The central goal of this review was to understand the physical principles that limit these techniques and determine ways for further performance optimization.
In the review paper, it was concluded that developing an efficient cost effective system for in-line & real time NP characterization (size, charge, shape/structure) on concentrated suspensions is a real technical challenge. The technology review presented in this report has allowed to highlight and explain the principle, the capability and the limitation of existing technology in regards to the SNOW project targets and requirements. Light has been specifically shed on DLS, Electrophoresis and SAXS, which are the selected/preferred technologies to be further developed in order to improve size, zeta and shape/structure measurements. Based on the current state of the art, some ways of improvement have been clearly identified in order to reach the technical objectives listed below:
- Measurement without stop flow: for in situ and real time measurement straight from the reactor/tank
- Measurement on concentrated and/or opaque suspensions: when no sample preparation, no dilution, no filtration, no sampling via external vial (no risk of contamination) is possible
- Measurement in harsh synthesis conditions: for measurements adaptable to any type of chemical production industry set-up (independently from pH, solvent types and/or temperature behavior up to 100 °C).
- Remote measurement and control: for work under stringent factory conditions
- Measurement automation: for time resolved particle size kinetics and process feedback control loops.
This review was complemented with an experimental evaluation review. For a comparative evaluation of the current metrology methods, our strategy consisted on testing all the available methods including TEM, SAXS, DLS, ZP, TGA (FTIR), UV on a set of NP model systems, including NP NIST “standards” (metallic, metal oxide, and organic). However, except for Latex NPs, no commercial source for NIST Reference Material could be found at that moment in the case of GNP or SNP.
After extensive experimental testing, we can conclude that there exists no unique perfect technique that can be selected as the “best” method for the characterization of NP properties (including size, shape, temporal evolution, functionalization etc.), but rather, a method is chosen to balance the restriction of the type of sample, the information required, time constraints and the cost of the analysis. Indeed, it is necessary to apply all of these techniques on the same set of NP since each technique provides specific information about only one or a few aspects of their properties. During the course of the project, we continued to apply these different methods to the same set of particles in order to continuously improve our understanding as well as the measurement techniques themselves.
The initial configuration of the Vasco DLS of Cordouan seemed to be not versatile and fast enough to allow for simultaneous DLS and SAXS measurements. Moreover, we experienced difficulties to couple DLS with SAXS in the initial configuration. Therefore, an extension of DLS was necessary to meet the requirements of the Snow Control project for in-line and real-time measurements. Based on the work described above, it was decided to develop a new concept of a fiber remote DLS probe head, which would allow simultaneous DLS and SAXS measurements. After a few design iterations, one fully functional prototype was designed, constructed and further used in the project.
Regarding enhanced SAXS, Bruker-AXS has fabricated, delivered and tested a completely new SAXS system. This new system uses a high flux liquid Ga metal jet x-ray source and modified Montel mirrors in replacement of the standard rotating anode source.
Initial measurements indicate that the x-ray flux is significantly higher (3 – 4 times) than with other sources. A further increase of X-Ray flux is realized by using “scatterless” pinholes. These kind of pinholes are designed to drastically reduce the parasitic edge scattering of the pinhole. Bruker-AXS also demonstrated with this system that a measurement time of a few seconds on a NP solution is enough to extract the NP size and distribution. Next step was to combine DLS and SAXS measurement simultaneously.
To extend the ZP measurement capability for the requirements of the Snow Control project, Cordouan has explored in parallel two paths: one “classical” opto-electronic method called electrophoresis and a second more exploratory one based on acoustophoresis.
The electrophoresis method is based on optical heterodyning measurement. It has lead, after several design iterations, to the development of a unique prototype named Wallis and its associated software called ZetaQ. It uses a dip cell design combined with the use of amorphous carbon electrodes and a high time resolution acquisition card to improve ZP measurement resolution and accuracy by a factor 3 to 10 over existing solutions.
Acoustophoresis is an original method that relies on the use of an acoustic sound wave to monitor the charges of the particles. Despites its promise, this path has been stopped at the board demonstration level, since the development lead time would not be compatible with the project timeline. Nevertheless, this work has opened up new technology opportunities, which will be further extend beyond the demonstrator.
To summarize, we successfully developed and validated an innovative fiber remote DLS probe. This new Vasco Flex was successfully integrated in the SAXS sample chamber. For SAXS, we successfully integrated the Metaljet and Nanostar, while the Montel optics was adapted to the Ga radiation and the beam collimation was upgraded with scatterless pinholes. Regarding the ZP metrology tool, artifact free measurements were demonstrated, with improved LDE technology and enhanced high resolution, which are also cost-effective.
Real-time measurements
The goal was to speed up the measurement time of the techniques (ZP, SAXS and DLS) and to show that these fast measurements can be performed continuously, in principle during the whole time needed for a reaction. To ensure extraction of relevant data over a long period, the goals are to optimize the computer architecture, to ensure that the data analysis follows the correlations and trends over a long period and to run the techniques DLS, ZP, SAXS under similar hardware and software platforms.
During the course of the project, Cordouan and Bruker AXS successfully achieved all these goals and demonstrated combined SAXS and DLS measurements during a long term measurement. The ZP device was connected at the end of the NP production and characterization line, when the nucleation and growth of NPs has been completed. All measurements are running on the same computer.
Furthermore, an exposure time of 10 s for an individual SAXS measurement was made possible. For DLS measurements, a user-friendly and straightforward interface was developed, allowing for automatable measurements combined with real-time measurements.
Integrated measurement platform for NP
We first created combined measurement cells, integrated the different techniques onto one platform and finally performed simultaneous measurements with the integrated techniques.
The goal of the first task was to design and test a measurement cell around which all the characterization techniques, SAXS, DLS and ZP, can be integrated to simultaneously measure the NPs. The original design provided for up to 10 DLS/ZP heads to determine the synthesis status at several locations along the flow. However, in the course of the project, this concept turned out to be not feasible due to economic reasons, practicability in the further industrial use and space constrains related to the integration of the DLS/ZP heads into the SAXS sample chamber. Hence, a new design of the cell was proposed.
This new design was based on the consideration that, to gain a full picture of the synthesis process, three synthesis steps comprising the nucleation, growth and functionalization has to be monitored. Consequently, 3 probe heads assigned to one of these synthesis steps, respectively, would be sufficient. Furthermore, as the ZP method is a mean to measure the stability of the NP suspension, ZP is applied at the end of the flow, because only at the end of the production the stability of the “end product” is of interest. These considerations led to a 3-capillary cell concept. Between the capillaries flow-reactors can be interconnected to supply the NP suspension with additional agents to functionalize the NP, and to accelerate or stop the growth. The measurement cell was designed, including the option to incorporate 3 capillaries. However, in the course of the project, a single-flow capillary was implemented and connected to a flow-reactor. It consists of a Quartz glass capillary with a wall thickness of 100 – 200 m, which is hold in place by a plastic frame. The cell can be aligned perpendicular to the laser beam, in order to accurately measure the flowing NPs.
Secondly, we integrated SAXS, DLS and ZP around the measurements cell. This implies the integration of the DLS device into the SAXS vacuum sample chamber (see Figure 1). In order to achieve this goal, a novel DLS head was designed especially for the purpose of performing simultaneous measurements together with SAXS, as described above. Thus, in other words DLS is brought to the sample position. This concept of the novel DLS device differs significantly from the standard functionality of state-of-the-art devices, where the sample is placed inside the device. This in turn makes the state-of-the-art DLS devices not practicable and feasible for in-situ and real-time measurements.
The outcome of the development is a novel DLS device, which fulfils all requirements and is named “VASCO Flex”. The main feature of the VASCO Flex device is an optical fiber remote head. This makes it versatile applicable, even beyond the Snow Control project. The device was mounted on the XY-table of the chamber together with the capillary holder as illustrated in Figure 1. As mentioned above the ZP device is applied only at the end of the production process. Therefore, ZP is not integrated into the SAXS chamber, but the ZP measurements are performed separately from SAXS and DLS after the production process.
The successful integration of the VASCO Flex DLS head allowed performing simultaneous real-time DLS and SAXS measurements. The complete system is shown in Figure 2. During these measurements the raw data were transferred to a common computer and analyzed automatically. SAXS and DLS were providing complementary information on mean radius, size distribution and hydrodynamic radius. To optimize the measurement results, the optimal sample geometry and sample volume were found. The above mentioned single flow-through capillary is sufficient for both SAXS and DLS. For SAXS, the capillary walls were sufficiently thin so that scattering and absorption effects are minimal. To optimize DLS measurements, refraction and reflection of the laser beam was minimized by proper alignment of the capillary.
For SAXS data processing, self-written software was used in the PYTHON programming language. For DLS and ZP measurements and data processing, proprietary software named ‘NanoQ’ and ‘ZetaQ’ was provided by COR. All software runs under WinXP on a common computer.
Simultaneous measurements with SAXS and DLS were successfully demonstrated at the same capillary location. SAXS measurements were unaffected by varying the flow rate of the NP, DLS measurements however varied depending on the different flow rates. Therefore, the DLS setup was modified so that the measured NP displacement is perpendicular to the flow direction. Under those conditions, simultaneous measurements were possible under flow conditions. Additional TEM measurements were performed to validate the SAXS and DLS results. This work resulted in a publication of A. Schwamberger et al. in the journal ‘Nuclear Instruments and Methods in Physics Research B’ in January 2015.
Development of materials and standards
We first optimized the NP models synthesis and subsequently defined the standard operation procedure related to the combined measurement methods.
For the first task, different types of metallic and metal oxide NPs were produced using several methods and were tested for the integrated NPs measurement platform. The NPs include colloids of different sizes as well as core shell heterostructure particles. The produced NP systems were chemically modified at their surface using either polyethylene glycol chain (PEG), a biocompatible polymer, or Bovine Serum Albumin (BSA) as a large biomolecule.
KU Leuven has prepared several model synthesis routes, including synthesis in batch reactors and micro-flow-reactors. The synthesis in micro-flow-reactors is focused on silica NPs and on gold NPs. To briefly summarize, we have described the methods used to generate standard NPs of BSTO, gold and silica. We have also shown the capability of functionalization of the NPs and to create an extra outer shell around NPs. To obtain these results, we have done synthesis experiments on different NPs and analyzed them with TGA/FTIR, TEM, SAXS and DLS.
Secondly, we developed standard measurement protocols for characterization of the NPs prepared. This procedure will include the three integrated techniques DLS, SAXS and ZP that operate simultaneously as described above. Furthermore, the coupled TGA/FTIR system is presented. The goal of this task is to develop specific procedures that exclude systematic errors and lead to optimal resolution and accuracy. The performance of a SAXS system integrated with DLS and ZP must produce as good and reliable data as when the same system is a standalone system. Since the sample conditions are different form the standalone situations, new procedures must be derived to ensure that the same level of precision and accuracy can be maintained. Furthermore, care must be taken to ascertain that the different techniques do not interfere, or interact with the sample in an irreversible manner.
A comparison was made between the measurements of the standalone systems, together with measurement systems integrated into single system. The measurements were performed on a set of standards based on silica NPs. These silica NPs were extensively characterized.
Small angle X-ray scattering (SAXS)
The major difference between the two systems is the sample preparation itself, and related background correction. With standalone SAXS measurements it is standard to take a background measurement of the buffer where the NPs are dispersed in. The NPs are also extensively washed before characterization and as such, the background is pure and free from most of the by-products. This ensure a good background subtraction and hence a better understanding of the properties of the NPs. However, for combined measurements extra measurements need to be performed before accurate background corrections can be done. The reactants need to be mixed together in order to have the background without the NPs, but with the same by-products. Since this is impossible in a continuous flow setup, a more complicated process is needed to correct for this. Two protocols are described to ensure good SAXS measurements in a flow/synthesis setup: 1) a simplified background correction, with the assumption that the background is more or less a weighted average between the different solvents and reactants, and 2) an exact background correction, applicable when the NP synthesis process is temperature driven.
Dynamic light scattering (DLS)
In a standalone DLS measurement, a measurement is started only after the sample is loaded in the center of a cell, ensuring a good alignment, which is followed by optimization of the laser intensity and determining the sample concentration. Five runs of 60 seconds are used, to allow for a statistical correlation of the data and fitting with the cumulants and Pade-Laplace algorithms. The real-time measurements with the VASCO-flex are completely different from the standard standalone system. Therefore, new protocols were designed in order to deal with the technical hurdles. The major challenge was the unknown (and changeable) position of the flow capillary compared to the VASCO-flex head. Since the flow heavily influences the DLS spectra (see below), the system was designed in such a way that the flow of the NPs was perpendicular to the laser beam and detector plane, which minimizes the influence of the flow on the spectra. After this alignment, the measurements can be done in the same way as with the standalone system.
Zeta-Potential (ZP)
As described above, it was decided to install the ZP system at the end of the synthesis process, when the particle size is stabilized to the targeted value, as a final control of the process, and before storage or deposition process with injectors. Therefore, the sample is collected at the end of the line and ZP measurements are done via the standalone system. After a thorough cleaning of the measurement cell and measuring the exact pH value, ZP measurements can be performed. A minimum of 7 runs per sample is used to establish measurement repeatability.
Thermogravimetric analysis coupled to Fourier transform infrared spectroscopy (TGA/FTIR)
In this project, we found that TGA/FTIR is about the only method to estimate the amount of functionalized groups around a NP. Simultaneous TGA/FTIR analysis is done via the Netzsch 449 A – Bruker Vertex 80V FTIR thermal analyzer (see Figure 4), which yields a thorough understanding of the decomposition mechanism of the ligands/molecules attached to the NPs. This is essential to optimize the synthesis conditions to obtain phase pure products and to estimate the amount of functional groups.
TGA is a method of thermal analysis in which changes in physical and chemical properties of materials are measured as a function of increasing temperature (with constant heating rate), or as a function of time (with constant temperature and/or constant mass loss). The Netzsch system, with its 25 ng resolution, allows us to accurately calculate the weight losses in different stages during the measurement, with just a few milligrams (1-3 mg) of the NPs required for a meaningful result. Since sample preparation is crucial in the outcome of the measurements, it is ensured that the NPs are either centrifuged or dried (if in a liquid state) to minimize the effect of solvent on the TGA profile.
TGA-FTIR coupled measurement is an interesting way to combine quantitative and qualitative analysis of the various functional groups present in the sample. In a typical measurement, the gases evolved from the TGA heating are fed through a transfer line to a gas cell and are analyzed by an FTIR detector. Thus, the different functional groups evolving at different temperature regions can be traced in real time. By comparing the observed FTIR spectra extracted at different temperatures with the FTIR spectra database, we confirmed the presence of the functional groups (typically organic) attached to the NPs.
Assembly of the NP proto-type production line and demonstration of its reliability and automation
A NP production line was connected by a NP production device to the measurement site inside the SAXS chamber. The NP synthesis is accomplished by a micro-flow-reactor. The synthesized NP are pumped into the capillary to be measured and characterized by SAXS and DLS. After passing the capillary, the NP are measured with the ZP device. The system is schematically illustrated in Figure 3.
Subsequently, the NP production line was demonstrated together with the different real-time characterization tools. For this demonstration SiO2 NPs were synthesized. The real-time capabilities of the system were tested on the synthesis of SiO2 NPs. In this context, the accuracy and the precision of the determination of the size and size distribution were investigated. To demonstrate the long term reliability of the measurement platforms, long term experiments were performed. During this period, the extracted parameters were monitored and displayed on a screen in real-time.
To achieve automation, a dedicated software was implemented to monitor the NP size and size distribution in real-time. The software features an automated background correction algorithm and a graphical user interface, which displays the NP size and the standard deviation of the size distribution as function of time.
Owing to the ability to track the NP size for different synthesis conditions and as function of time, it is possible to investigate the influence of different synthesis parameters on the size of the NPs. The graphical user interface (GUI) facilitates the investigation and makes the optimization of the synthesis conditions more efficient as it displays the NP size in real-time. Different synthesis parameters were investigated. By this means, one can find optimal synthesis parameters for a certain size.
For a proper assignment of a certain size to a synthesis parameter set, high accuracy and precision are required. The accuracy represents the degree of conformity of the measured NP size to the true NP size. To estimate the accuracy of a SAXS measurement the obtained SAXS result is compared to a TEM result. It was successfully demonstrated that for fully grown NPs, the SAXS technique provides size values in good agreement with the TEM results. Therefore, one can conclude that SAXS has a high accuracy (+/- 0.5nm).
The precision is the degree to which consecutive measurements give the same or similar results. For SAXS measurements the precision depends on the signal-to-noise ratio. In turn, the signal-to-noise ratio depends on the X-ray flux at sample position and the exposure time. The noise in a SAXS curve leaves/allows freedom in the fitting procedure and extraction of the NP size. Therefore, a high noise level leads to less reproducibility and thus to low precision. To investigate the influence of the noise on the reproducibility of the results, an experiment with different exposure times was performed. Based on the outcome, it was concluded that the reproducibility of SAXS is ultimately within +/- 0.5nm. Therefore, the high accuracy and precision of SAXS allow investigation of the influence of different synthesis parameters. It can be assumed that the size values displayed in the GUI correspond to the true size of the NPs inside the capillary.
Moreover, long-term measurements over at least 16 hours were performed to demonstrate the reliability of the real-time monitoring procedure. The experiment was divided into two parts: the combined SAXS, DLS and ZP measurements were performed during the day with variable synthesis parameters, followed by overnight measurements, during which these parameters were kept constant.
With this task, the feasibility of real-time monitoring of NP production was demonstrated. Owing to the high accuracy and precision of the determination of the size and size distribution it is possible to track the temporal development of the NP size inside the capillary. This results in the possibility to investigate the influence of different synthesis parameters. The optimization is facilitated by a dedicated software including an automated data processing and a graphical user interface which displays the size and size distribution in real-time on a screen. By this means, the optimization becomes efficient and fast. Consequently, synthesis parameters can be assigned to a certain NP size. In turn, this results in the ability to synthesize NPs of a particular size on request. Furthermore, reliability of the monitoring process and the SAXS device was demonstrated over at least 16 hours. During all experiments the devices were running without problems.
End-user evaluation
The performance of the individual metrology methods was evaluated from the perspective of an end-user and potential customer. This evaluation was done on the prototype tools developed in the project.
The task had two parts. First the specifications and testing/measurement procedure of the different metrology tools developed within the Snow project, namely the DLS, ZP and SAXS, were gathered. The main technical specifications for the primary characterization of the NP (accuracy, resolution and repeatability in the extraction of the diameter) was compared to the measured specifications obtained on the toolset actually installed at KU Leuven. In this task mainly data obtained on gold NPss from Cytodiagnostics Analytics are used for comparison. Second, the user-oriented parameters (ease of maintenance and maintenance cost, long-term reliability, total cost of operation and economic viability of the system) are estimated. This aimed at doing an evaluation of the cost of ownership by looking at the maintenance and repair costs during the project. Similarly, this evaluation is based on the experience of the industrial partners together with KU Leuven employees with the tool.
The interest of IBM in the SAXS tool as an end-user is two-fold. The purpose of the visit for evaluation was first to estimate the brilliance of the Excillium source, and then to evaluate whether the tool installed at KU Leuven could potentially be used for first GISAXS tests. The comments are below:
•
Excillium source: A simple protocol has been defined to estimate the flux of photons emitted by the Excillium source, and to compare it with the rotating anode generator from Bruker AXS installed at IBM premises. Since the direct beam could damage the GADDS detector, the intensity of the direct beam has been measured at KUL using the detector and a sandwich of Al and Zn foils. For this measurement, the source had been operated at 200W and with a beam size after pinholes of about 0.5x0.2mm. The exact same absorber has been mounted on the D8 Discover at IBM Zurich, and the direct beam measured as well to compare for the respective intensities. Assuming a similar detection efficiency for the GADDS and scintillator detector, and assuming a 50% larger absorption for a CuKx radiation compared to GaKa, the flux estimated on the Excillium source is about 1-2.1010 photons/s. This is about a factor 5 lower than the RAG – albeit with line focus instead of the point focus used here – making it an interesting alternative for our application, in particular for microdiffraction on test pads.
•
GISAXS tests: As a preliminary test, a thin film sample has been prepared, made of a thin, epitaxial layer of BaTiO3 directly grown on silicon. The composition of the oxide matches the one that is developed in the NP thin film tool. The sample has been mounted in the SAXS chamber, although the holding stage was by far not optimized for GISAXS geometry. Despite this not suited geometry, a clear GISAXS signal could be observed, and fringes observed on the GADDS detector, related to the finite size of the BaTiO3 sample along the growth direction.
A second part of the task was to evaluate the standard operating protocols used during the simultaneous measurements. Indeed, the experimental condition under which a sample - such as a NP solution - is measured with the individual SAXS, DLS, and ZP techniques is quite different from the experimental conditions needed to enable a simultaneous measurement.
There is also a substantial risk to make systematic errors due to not optimal measurement cells. Clearly, the obtained results from a combined measurement that includes SAXS must be as good as those obtained from an individual SAXS measurements.
The result of this evaluation shows that simultaneous DLS and SAXS measurements at the same capillary position and the same time and real-time monitoring of NP synthesis are possible. While SAXS measurements under flow conditions are still reliable, using DLS under flow conditions however is more challenging task and need a precise alignment of the measurement geometry perpendicular to the flow direction.
To summarize, the different tools have been evaluated by IBM and they performed as expected, in particular in a research environment. As far as we could conclude, the main weakness of the installed tool set is essentially related now to the additional requirements for a combined analysis using DLS and SAXS. It can be foreseen that a “routine” analysis using the combined method will require a redesign of the sample holding stage and of the alignment unit. Furthermore, a multi-sample stage should be designed for such a combined DLS/SAXS tool, so that a semi-automatized analysis routine can be envisioned. In this case, two independent alignment stages for SAXS, resp. DLS will have to be designed. For the moment indeed, the design of the stage allows for multiple samples to be loaded for SAXS, but the DLS alignment stage can only be focused on one single position. An independent alignment stage for the DLS head would enable multiple samples to be loaded for both SAXS and DLS.
2. Development of a thin film deposition system for NPs.
For the development of the second integrated measurement platform, a number of novel methods needed to be developed first. A novel laser based sensor was developed, to measure the flux of NP. A novel high vacuum system was designed and fabricated, followed by the development of a novel NP vaporizer. These tools were combined and integrated into a complete NP thin film deposition system. We performed simultaneous measurements, followed by a demonstration of the reliability and automation of this NP thin film system. The work was finalized by a public demonstration and end-user evaluation. The results of each step will discussed in the following paragraphs.
Development of novel VUV systems
The main goals were to measure the flux of NP in the vapor phase, as well as to make a diffraction pattern from NP thin films. However, it became clear during the studies of the first year, that for the latter undertaking a coherent source of VUV light is necessary. This was outside of the capabilities (technical, scientific and financial) of the project. Hence, most of the efforts concentrated on a non-coherent VUV source, as well as on an UV diode laser (coherent) source. The advantages and disadvantages of both approaches were studied and, taking into account the technical and financial capabilities of the project, it was decided to use a diode laser as an excitation source.
To test the idea of the flux measurements, DCA designed and manufactured a system to measure the atomic Ga flux with a diode laser as an excitation source. The Ga source temperature was varied between 1000 °C and 1200 °C and the measurement system was capable to detect an atomic Ga fluorescent flux signal proportional to the temperature. After extensive designing and testing, DCA developed a compact, single chamber flange flux sensor (see Figure 5) with associated flux sensor software, which allows to control the fluorescence sensor and to perform data acquisition.
Newly designed vacuum measurement system and NP vaporizer
DCA designed and built a new vacuum measurement system (see Figure 5). Designing the system was a challenging and time consuming task for DCA. Only load lock and main parts of the sample manipulator were standard DCA components. The process chamber was designed from scratch in close partnership with Kemstream. The sample manipulator was modified to be able to handle high pressure regimes, with a presence of various gases and liquids that might be used during the process. This, together with the requirement to reach temperatures up to 1000 °C, took a lot of testing and assembling hours. Measurement systems which are integrated to the process chamber, are a RHEED analyzer and the laser based NP fluorescence sensor, which DCA developed as described above. Operation of the RHEED analyzer was demonstrated during initial silica NP depositions. Operation of the NP sensor was demonstrated later when the NP thin films were grown, as described below.
Kemstream built a modified version of Vapbox 1500 NP injector and all its associated accessories (hardware and software, see Figure 5). All these elements were delivered to DCA. Kemstream provided technical support to DCA for the integration of NP injector on the newly designed DCA vacuum system.
Integration of the newly designed tool into one measurement platform
When DCA was assembling the integrated measurement platform, it also integrated the laser based fluorescence sensor and the RHEED analyzer. After performing first functional tests of the NP injector installed on DCA vacuum system, DCA and Kemstream provided training to KU Leuven in the use of the system. Following the training, first silica NPs injection and deposition on silicon wafers were done by DCA (Markku Rajala), KU Leuven (Bert de Roo) and Kemstream (Hervé Guillon). Finally, DCA delivered the complete system to KU Leuven, to perform simultaneous measurements, which was the main goal of the last task within this WP.
These procedures took a lot of developing, testing and assembling hours. Before the system could be installed, the necessary safety measures needed to be in place, as prescribed by the KULeuven team of Health, Environment and Safety (HES). The details of the installation as well as the different modes of operation and maintenance were discussed and tested. In general, the following precautions were taken to avoid NPs to become airborne: 1) examination of the possibility to deposit a transparent TEOS-based Si-oxide film on top of the NP film; 2) installation of an additional exhaust in the lab with a movable exhaust hood and a filter cabinet, where filters can be replaced without having to expose the filter to the lab air. Additionally, special safety measures were taken to protect the user when loading a sample into the system. The necessary safety procedures were derived from IBM’s procedures and were adapted to the specific Snow Control toolset, where a special plastic glove box was installed in the system.
Assembly of the integrated measurement platform for NP thin films together with a thin film production line
The assembly of the thin film system was mainly performed by Kemstream and DCA. With this production system, standard thin films can be grown by DLI-CVD (Direct Liquid Injection Chemical Vapour Deposition). The thin film production system used for this demonstration was a low pressure DLI-CVD system put at the disposition of the project by DCA for the purposes of these demonstration activities. The system consists of several metal organic precursor lines, a DLI vaporizer, pressure and flow controllers, temperature regulation, pumps, precursor inlet valves, purge and oxygen gas lines, etc. These components were all attached to the newly designed vacuum system. SiO2 thin films were chosen as standard DLI-CVD thin films. Two different silica precursors were chosen in this work: TEOS (tetraethyl orthosilicate) and DADBS (diacetoxyditertiobutoxysilane).
In a first set of experiment we varied the different deposition parameters like time, temperature and pressure, and compared these the corresponding changes in thickness and roughness parameters of the thin films, which were measured by X-ray reflectivity (XRR) measurements (via X’Pert Power X-ray platform at KU Leuven) and Scanning Ellipsometry (SE) measurements (at IBM Zurich).
Specific fringes in the XRR spectrum indicate a clear temperature-thickness and time-thickness relation. This indicates that a higher substrate temperature results in a higher thickness as well as longer deposition times in a higher thickness. The CVD growth is in the kinetic regime where the growth rate is limited by the temperature. At a given deposition temperature, the film thickness linearly increases with the deposition duration.
When we want to use this precursor and its inherent high decomposition temperature together with NPs some problems may arise with the thermal stability of the NPs. Thus we want a “better” precursor, i.e. one that decomposes at a much lower temperature. This is the reason we changed our precursor to DADBS.
Similar to the first experiments with TEOS, we varied the different synthesis parameters and correlated these to difference s in thin film thickness and roughness. Again, specific fringes in the XRR spectrum indicate a clear temperature-thickness relation. It was concluded that a higher substrate temperature results in a higher thickness. The growth is in the kinetic regime where the growth rate is limited by the temperature. Similarly, it was shown that a longer time result in higher thicknesses. At a given deposition temperature, the film thickness linearly increases with the deposition duration.
To summarize, we showed that we successfully integrated the measurement platform on a working thin film production line. With this system we could grow silica thin films from different precursors, namely DADBS and TEOS. With TEOS, we could only grow at an elevated temperature of around 900°C. When we used the precursor to DADBS, we could grow at 600°C, which implies very significant diminish of the thermal load.
In order to further lower the deposition temperature, it was decided to add an ozone line. The ozone reacts with the precursors and reduced the decomposition temperature. For this purpose, DCA developed a system consisting of dedicated gas lines with an ozone generator, mass flow controller, pressure gauge and pneumatic valves. With this system oxygen gas enriched with about 10% of ozone can be introduced into the system.
Simultaneous measurements
Simultaneous measurements with the flux sensor and RHEED turned out to be difficult since the pressure requirements are not compatible. The RHEED filament would burn out too fast. However, sequential operation can be done whereby first the NP’s are deposited and then the chamber pressure is reduced so that the RHEED measurements can be performed.
Once the system was operational at KUL, CdSe/ZnS core-shell type quantum dots (QDs)were successfully used, in combination with a silica thin films. The deposition was monitored in real-time using the newly designed NP fluorescence flux sensor. RHEED was used to analyze the structure of the NP thin films. On the fluorescent RHEED screen, different features and spots could be seen. These spots correspond to the different crystal structures available on the surface of the substrate and from the NPs themselves. These data also confirm the crystalline nature of the QDs NP’s.
The GI-SAXS experiments have also taken place subsequently after the films have been deposited. From the GI-SAXS measurements we can obtain information about the size of the NPs and the ordering of these NPs onto the surface of the substrate. The measured size of NPs corresponded with the expected values. The results were also confirmed by AFM measurements.
To summarize, the instrumentation developed for the integrated measurement of NP deposition in thin films such as the laser based flux sensor, the RHEED and the GI-SASX all work well and deliver results as expected. The complete system is illustrated in Figure 7.
Demonstration of the reliability and automation of the NP thin film proto-type production line
We successfully demonstrated the NP thin film growth inside the DLI-CVD thin film deposition system. For this demonstration, two different thin film growth procedures were used. Firstly, gold NP thin films were deposited. Secondly, BTO NPs were used. To achieve automation, a dedicated software was implemented to monitor the gas and liquid flow in real-time. The software featured a GUI, which displays the gas (setpoint, actual flow, actual tON) and liquid parameters (setpoint, actual flow, actual tON) as function of time. If the deposition parameters were unchanged, the different flow rates remained constant. If deposition parameters were changed, a jump was observed in the flow versus time graphs. Additionally, the temperature and the resistance was followed on the screen. Additionaly, the Kemstream Control Software Vapview was able to program a predefined recipe. The deposition was started manually by the experimenter.
Owing to the ability to track the flow for different deposition conditions and as function of time, it is possible to investigate the influence of different deposition parameters on the thickness of the films. The GUI facilitates this investigation and makes the optimization of the deposition conditions more efficient as it displays the flows in real-time. Different deposition parameters were investigated. Therefore, one can find the optimal deposition parameters for a certain film thickness and structure.
Long term measurements were performed over at least 24hrs to assess the reliability and stability of the deposition system. The temperature of the substrate holder, pressure of the chamber and the NP flow were measured as a function of time.
Secondly, gold NP were deposited on a silica substrate. Different deposition parameters, like duration, vaporizer temperature, wafer holder temperature and solution flow rates, were varied during different experiments to account for difference in deposition of gold NPs. The gold NPs were clearly visible after deposition on the silica substrate.
The NP thin film was also characterized by XRR and AFM measurements, confirming the presence of gold NP in the thin film. However, from these measurements it became clear that the gold NPs were not dispersed in a homogenous way, but formed clusters of different sizes. In addition, a difference in height was observed, which varied from few nm to a 10-fold increase in nm.
To resolve these issues, new NP thin films were grown in which the vaporizer temperature and flow rate were varied. The NP thin films were again measured with XRR and AFM, which revealed a more homogenous distribution of NP as compared to the previously grown films. Furthermore, the height of the NPs clusters was smaller and the dispersion was higher. These experiments suggest that depositing gold NPs during a longer time at a smaller flow rate could be beneficial. Finally, commercially available BaTiO3 (BTO) NPs were deposited on a silica substrate. After optimization of the total amount of solution that was injected by varying different deposition parameters like solution flow rate and time, a homogenous distribution of NP was obtained. The results were confirmed by XRR, AFM and SAXS measurements.
Of note, these results differ from the data obtained with gold NPs, which implies that the suitable conditions for one type of NPs does not have to be the same for a different kind. We found that, in general, deposition of BaTiO3 seems to be more uniform
Demonstration of the complete system
The demonstration was given on Thursday the 4th of June 2015 during the final Snow Control meeting in Leuven for all the different partners. The assembly of the NP production line was connected with the NP thin film deposition system. CdSe/ZnS core-shell type quantum dots were chosen as NP to visualize simultaneously the flow measurements, as described above. This was done because no NP made with continuous flow system has been fluorescent so far, and therefore impossible to detect with the NP flux meter. This technical hurdle implies also that we are unable to show the complete system during the synthesis of NP. However, since the production process parameters are mimicked in this setup, the results will not change when we change to a NP production setup.
The main difficulties to prepare the demonstration arose with the connection between the NP synthesis flow reactor and the NP injection system. For this connection new equipment needed to be assembled and one of these tools is a new top flange of the fluid canisters. Indeed, to allow continuous injection in the Vapbox/DCA machine of NPs suspension synthesized and delivered by the Syrris microreactor flow system, a new top flange needed to be developed. Before, the system consisted of two lines: a 3-position valve at the solvent port and a 2-position valve at the pressurization port. To be able to inject solution from the Syrris reactor, an additional port needed to be installed on top of one of the NPs suspension tanks. Therefore, a new top flange was developed including this extra port (which is also a 2-position valve). This connection is suited for the 1/16” diameter of the tubings.
The assembly of the NP production line is shown in Figure 6. For the public demonstration, the Syrris pumps were connected to the NP thin film system. Due to the impossibility to move the SAXS to the NP thin film platform and vice versa, SAXS measurement systems were not included in this demonstration. The complete system demonstration is done with CdSe/ZnS core-shell type quantum dots, with a diameter around 5 nm and stabilized with octadecylamine ligands as used before. They have a fluorescence signal λem @ 600 nm (solid) and were deposited inside the deposition chamber. These QDs are chosen because of their clear emission peak at 600 nm falling in the range of our NP flux sensor (580-627 nm).
The different parts of the NP thin film system were connected with the microfluidic reactor. The connecting liquid line is indicated with a red arrow. When the flow was running, the injection of the NPs into the system was started. The flow of these NPs was also monitored with the NP flux meter. The deposition is monitored in real-time using the NP flux sensor. The fluorescence signal is detected by the photomultiplier and the signal is converted to voltage by the high-accuracy voltage meter.
End-user evaluation
The NP production system has been evaluated with respect to basic parameters relevant in a production environment. Within the limited data set acquired with the system fully integrated, the long term stability and the reproducibility of the experimental conditions have been quantified. Although more time is required to establish the characteristics of the integrated system as a production unit, the basic requirements are fulfilled, in particular for the long term stability of the different components. Finally the cost of ownership for such an integrated system was estimated, on the basis of maintenance costs experienced during the project. The cost of parts necessary amounts to about 13K and required a down time of about 8 days over the last 18 months.
Potential Impact:
The Snow Control project was a large project wherein a lot of metrology tools were developed. These tools were then assembled together into two novel real-time integrated characterization systems for the properties of NP and deposited NP thin films. To illustrate the impact of the Snow Control project on European research and industry, we will describe how the project led to new collaborations and future research projects, but also how the project enhanced the current scientific knowledge on the different metrology tools. Furthermore, we will discuss how Snow Control led to new products and associated increases in business and sales of the different SME involved.
One demonstrator system has an integrated real-time measurement platform linked to an automated NP generation system. The real-time characterization methods developed work in an automated production cycle, which uses combined DLS, ZP and SAXS to automatically characterize NPs of the selected size. This project opened new markets for interdisciplinary sectors, such as nano-medicine, where custom designed reactors can be adapted for the synthesis of NP in the search for new molecules and drugs. Based on the Snow Control Results, we applied with partially the same partners of the Snow Control project, namely KUL, Cordouan and Bruker AXS, for an Horizon2020 NMP-11 project relating to ‘Nanomedicine for cancer therapy’. The project successfully passed the first evaluation phase and is currently in preparation for the second phase submission.
Furthermore, Cordouan participated in a FUI project ‘SAXSize’ project and Horizon2020 project ‘Improcore’, both of which are under evaluation and are based on the work done within Snow Control. The newly developed DLS probe ‘Vascoflex’ was very interesting for the maturation of their current technology, since this tool enables to perform DLS measurements on ‘in bottle’ NP. For this new market segment it is very important that measurements can be be done without exposing the NP to the environment, and therefore it is applicable in bio-applications.
For future research possibilities, Bruker AXS will further investigate the flowthrough capillary, in order to improve the transmission characteristics for X-rays. The combination with DLS will be explored for a better alignment in height, automation and the simultaneous display of SAXS and DLS results. Therefore, the unique combination of DLS and SAXS opened the horizon for new potential customer interests, and is waiting for some additional investments to bring the combined tools to the market. For the NP synthesis, correlations between SAXS and DLS will be investigated, including focusing on monodisperse NPs, their functionalization and also on core-Shell NPs.
The other demonstrator of the project is a thin film deposition system for NPs, opening a wide range of applications. An obvious application is the deposition on a substrate of catalyst particles for CNT or nanowire growth. We also expect this system to present a breakthrough for the production of photonic or photovoltaic devices with tailored band-gap on cheap substrates.
Both research paths paved the way for further research opportunity and currently additional projects are in preparation. Furthermore, the systems developed are the first ones available worldwide with the above mentioned capabilities, therefore assuring an immediate high interest from potential end users as well as the possibility to develop new materials with tailored properties and characteristics across different industrial sectors. This is already obvious from the high number of orders received for the tools developed within Snow Control by the industrial partners, as discussed below.
Snow Control was an accelerator lever for R&D of new products. For the SAXS systems of Bruker AXS, Snow Control significantly enhanced the knowledge on the MetalJet and Nanostar technology, leading to increased costumer interests and sales. This is translated into increased business by the sale of 4 new SAXS systems (0,5 M€ per system) to new clients, implying a significant increase in sales revenues of about 2 M€. These 4 systems were installed in Singapore, Bratislava, Denmark and Germany. Also, one Metaljet system for diffraction was sold.
Cordouan was able to expand their portfolio by two new products developed within Snow Control, namely the VASCO flex (remote in situ DLS) and WALLIS system (ZP analyzer). So far, Cordouan already sold 11 Wallis and 5 VASCO flex units. This implied a revenue in their sales which is worth about 0,5M€ turnover. Furthermore, they expanded their capacity in staff by creating 2 engineer positions created by the means of Snow Control. Finally, Cordouan is involved in several on going scientific publications with reference labs using their VASCO flex, contributing to the broad range dissemination potential of Snow Control.
For DCA, they invested a lot in the development of novel methods, namely in the newly designed NP flux sensor, which is proven to work very well both for atomic fluxes as well as for fluorescent NP fluxes. DCA will further develop it for use in molecular beam epitaxy, and thereby enhancing its application potential. A high number of interested customers have already been identified. The sensor will also be expanded for accurate flux measurement of group V elements in quaternary compounds like InGaAsP and AlGaAsSb, taken into account the limitations of currently available laser diodes.
For KEMSTREAM, the new VAPBOX design was a very important investment. It expanded their current direct liquid injector (DLI) technology to include NP solutions into the injection process. This opens new avenues for their DLI – CVD technology. The injector was then assembled into a thin film deposition process together with DCA and resulted in the first NP thin films grown with this method. This first-of-a-kind tool opens new markets in many of the thin film areas where the innovations that become possible with NP can now be implemented into practical solution.
List of Websites:
https://fys.kuleuven.be/vsm/
Prof. Jean-Pierre Locquet
Department of Physics and Astronomy
Celestijnenlaan 200D, B-3001 Leuven, Belgium
Phone: +32 (0)16 327290
Fax: +32 (0)16 327983
One of the major barrier towards the successful introduction of nanoparticles (NP) into many applications is the lack of a tight control on their properties, i.e. size, shape, crystallinity, core-shell, functionalization, etc.) This can only be resolved by the simultaneous use of several metrology methods and to extract the relevant information in real-time. In this Snow Control project, we take the first steps towards this goal by developing real-time characterization metrology tools to measure the properties of NP, functionalised NP and NP thin films. The Snow Control project realized two novel real-time integrated characterization systems.
Firstly, an integrated real-time measurement platform linked to an automated NP generation system was developed. To enable a real-time and a complete characterization of NP properties, we reviewed and drastically extended the capabilities of Dynamic Light Scattering (DLS), Zeta Potential (ZP) and Small Angle X-ray Scattering (SAXS). Since there exits no perfect universal single technique for complete NP characterization, these enhanced techniques were combined into a novel integrated real-time NPs measuring platform. Therefore, a new DLS probe (from Cordouan) was designed, which was remotely connected to the laser source and to the laser detectors. The SAXS instrument (from Bruker AXS) was improved and gained a significantly higher X-ray flux, using a liquid Ga metal jet x-ray source. The ZP measurement capability was extended on the basis of electrophoresis, and led to the development of a unique Wallis system (by Cordouan), with increased resolution and accuracy by a factor 3 to 10 over existing solutions. Subsequently, real-time combined measurements were realized, and a setup was successfully developed which allows simultaneous DLS and SAXS measurements under flowing NP conditions. Since it was technically not feasible to perform ZP measurements on these flowing particles, the ZP tool was installed at the end of the synthesis process, to allow for offline measurements. To realize real-time measurements, the measurement time should match the synthesis timescale. A measurement time of 10 seconds was within reach. A successful demonstration of the combined setup was accomplished during the project. Combined SAXS and DLS techniques can follow the size evolution of NPs in real-time.
Secondly, an integrated thin film deposition system for NPs was developed. A novel vacuum system for the deposition of a NP thin film (from DCA) was designed and fabricated. To enable the deposition of NPs in vacuum, the liquid injection/vaporization unit (from Kemstream) was redesigned. The vacuum system was then built with this integrated NP injector, a RHEED (Reflection high-energy electron diffraction) analyzer and a newly designed laser based fluorescence NP flux monitor. NP thin films were successfully grown, as shown by RHEED and GI-SAXS experiments. These results were further confirmed by AFM data.
Systematic characterization of the different NP was carried out using the methods available in the project. But it became clear that some methods are missing, such as UV-VIS absorption and TGA/FTIR. In particular, KUL demonstrated that TGA/FTIR is the only method that can provide a quantitative measure of how much functional (organic) groups are effectively attached to NP.
Finally, the methods developed in this project were evaluated and assessed for further use in industrial manufacturing applications. To do so, the specifications and testing procedure of the different metrology tools, the cost of ownership, the NP standards and the simultaneous measurement protocol were documented and evaluated by the end-user. The main conclusion is that the tools performed as expected and that a combined SAXS/DLS analysis on NPs, in particular during real-time experiments, is certainly feasible. Also, for the first time, NP thin films were grown and monitored from a NP solution.
Project Context and Objectives:
A. Project Context
Nanomaterials are currently already used in a variety of applications in the fields of energy, health, information and communication technology, as well as materials engineering. Among these nanomaterials, the nanoparticles (NP) are a major category and exist in many shapes, sizes and structures. The different shapes are ranging from the spherical single particle, rod like, or the multi-shell nanorice particle to highly branched nanostructure. For most industrial and/or medical applications, a very narrow size & shape distribution is highly desired, to establish toxicity standards for instance. Also small differences in size, shape and crystal structure can have a large effect on the performance, in the field of catalysis for example. Unfortunately, today none of the known methods allows for an accurate determination of the physical and chemical properties of many of the mentioned NPs.
Ensuring that the EU can realize the commercial potential of nanotechnology, industry and society will require reliable and quantitative means of characterization, as well as measurement techniques that will underpin the competitiveness and reliability of future products and services. Metrology and standards need to be developed to facilitate rapid development of the technology as well as to provide users with the necessary confidence in their process and product performance.
This project is the first step towards full feedback control of NP synthesis by developing and integrating dedicated measuring techniques. For instance, if the size of the NPs can be measured in real-time then the multi-shell structure can be resolved during the synthesis. However, it is not the goal of this project to close the feedback loop (i.e. not using the obtained data to manipulate the process), but to focus on building integrated measurements systems that will be tested in realistic industrial conditions. The project provides two unique integrated systems, that will allow for innovative developments in measurement techniques needed to cope with the demands of nanotechnology. The systems allow to develop standard procedures for nanomaterial synthesis, a set of reference materials for property measurement standardization, standard methods for physical and chemical properties evaluation, and finally to develop standards for material evaluation in interdisciplinary applications such as biomaterials, photonics and photovoltaics, nanomedicine, etc.
B. Project Objectives
Our research activities aimed to fulfill five scientific and technology objectives. The first objective aimed to improve the size, shape and temporal resolution of measurement techniques, which already exist today. The goal of the second objective was to develop novel NP characterization methods that do not exist today. The third objective aimed to create a unique combination of the best elements of the first two objectives into 2 different real-time measurement platforms, for NPs and NP thin films, respectively. The fourth objective focused on reproducibility and standardization using the combined measuring methods, while the final objective demonstrated the integration into a two production lines.
More specifically, the project had the following objectives:
Objective 1: Demonstrate DLS, ZP and SAXS with improved size, shape and temporal resolution.
This objective includes an array of multi-optical DLS and ZP heads along the line of a microfluidic reaction path as well as the use of a new liquid Ga metal jet X-ray source for SAXS.
Objective 2: Develop real-time concepts, VUV techniques and probes for novel properties.
This objective includes developing a real-time concept to follow the synthesis of NP at different locations in a flow chemistry line as well as a VUV based NP flux sensor (in a vacuum system).
Objective 3: Integrate different techniques into two real-time characterization platforms.
This objective includes the integration of ZP, DLS and SAXS measurements together along the flow chemistry line to follow in real-time the synthesis of NP. This includes also the real-time measurement of the NP flux in a vacuum system.
Objective 4: Towards real-time multiple measurement standards and standardized NP.
This objective aims to devise standard operating procedures for the preparation of standardized NP and functionalized NP. The objective also includes the establishment of standard operating procedures for the simultaneous measurement and analysis of NP properties obtained via the real-time measurement platform.
Objective 5: Demonstrate integration into a (high throughput) production line.
This objective includes the demonstration of the integrated NP measurement platform on a flow chemistry process as well as the demonstration of the NP thin film measurements on a thin film deposition system.
Project Results:
The Snow Control project realized two novel real-time integrated characterization systems for the measurements of the NP size and shape. One demonstrator system has an integrated real-time measurement platform linked to an automated NP generation system. The other demonstrator is a thin film deposition system for NPs.
The two real-time characterization methods developed work in two automated production cycles. The first cycle uses Dynamic Light Scattering (DLS), Zeta Potential (ZP) and Small angle X-ray scattering (SAXS) to automatically characterize NPs of the selected size, while the second one uses in-line UV flux sensing and RHEED to create NP thin films of selected structure and thickness. Both platforms were demonstrated and evaluated from the perspective of an end-user and potential customer, namely IBM.
Below we will give a concise overview of the work done throughout the whole project. The development of the integrated real-time measurement platform, linked to an automated NP generation system, will be discussed first, followed by the description of the work related to the development of the thin film deposition system for NPs. The work on the 2 systems was performed in parallel throughout the whole project.
1. Development of an integrated real-time measurement platform, linked to an automated NP generation system
We first evaluated and extended current metrology methods, namely DLS, SAXS and ZP. We performed real-time measurements with these extended methods, which were then combined and integrated into one measurement platform. A set of standard procedures were designed and optimized, followed by a demonstration of the reliability and automation of this NP production line. The tools were finally evaluated from an end-user point of view. The outcome of each step will be discussed in the following paragraphs.
Evaluation of current metrology methods
This task addressed the evaluation of the current metrology methods and is materialized by the release of a report, that reviews the most relevant metrology techniques currently in use and identifies their limits with respect to sensitivity, accuracy and time resolution. The central goal of this review was to understand the physical principles that limit these techniques and determine ways for further performance optimization.
In the review paper, it was concluded that developing an efficient cost effective system for in-line & real time NP characterization (size, charge, shape/structure) on concentrated suspensions is a real technical challenge. The technology review presented in this report has allowed to highlight and explain the principle, the capability and the limitation of existing technology in regards to the SNOW project targets and requirements. Light has been specifically shed on DLS, Electrophoresis and SAXS, which are the selected/preferred technologies to be further developed in order to improve size, zeta and shape/structure measurements. Based on the current state of the art, some ways of improvement have been clearly identified in order to reach the technical objectives listed below:
- Measurement without stop flow: for in situ and real time measurement straight from the reactor/tank
- Measurement on concentrated and/or opaque suspensions: when no sample preparation, no dilution, no filtration, no sampling via external vial (no risk of contamination) is possible
- Measurement in harsh synthesis conditions: for measurements adaptable to any type of chemical production industry set-up (independently from pH, solvent types and/or temperature behavior up to 100 °C).
- Remote measurement and control: for work under stringent factory conditions
- Measurement automation: for time resolved particle size kinetics and process feedback control loops.
This review was complemented with an experimental evaluation review. For a comparative evaluation of the current metrology methods, our strategy consisted on testing all the available methods including TEM, SAXS, DLS, ZP, TGA (FTIR), UV on a set of NP model systems, including NP NIST “standards” (metallic, metal oxide, and organic). However, except for Latex NPs, no commercial source for NIST Reference Material could be found at that moment in the case of GNP or SNP.
After extensive experimental testing, we can conclude that there exists no unique perfect technique that can be selected as the “best” method for the characterization of NP properties (including size, shape, temporal evolution, functionalization etc.), but rather, a method is chosen to balance the restriction of the type of sample, the information required, time constraints and the cost of the analysis. Indeed, it is necessary to apply all of these techniques on the same set of NP since each technique provides specific information about only one or a few aspects of their properties. During the course of the project, we continued to apply these different methods to the same set of particles in order to continuously improve our understanding as well as the measurement techniques themselves.
The initial configuration of the Vasco DLS of Cordouan seemed to be not versatile and fast enough to allow for simultaneous DLS and SAXS measurements. Moreover, we experienced difficulties to couple DLS with SAXS in the initial configuration. Therefore, an extension of DLS was necessary to meet the requirements of the Snow Control project for in-line and real-time measurements. Based on the work described above, it was decided to develop a new concept of a fiber remote DLS probe head, which would allow simultaneous DLS and SAXS measurements. After a few design iterations, one fully functional prototype was designed, constructed and further used in the project.
Regarding enhanced SAXS, Bruker-AXS has fabricated, delivered and tested a completely new SAXS system. This new system uses a high flux liquid Ga metal jet x-ray source and modified Montel mirrors in replacement of the standard rotating anode source.
Initial measurements indicate that the x-ray flux is significantly higher (3 – 4 times) than with other sources. A further increase of X-Ray flux is realized by using “scatterless” pinholes. These kind of pinholes are designed to drastically reduce the parasitic edge scattering of the pinhole. Bruker-AXS also demonstrated with this system that a measurement time of a few seconds on a NP solution is enough to extract the NP size and distribution. Next step was to combine DLS and SAXS measurement simultaneously.
To extend the ZP measurement capability for the requirements of the Snow Control project, Cordouan has explored in parallel two paths: one “classical” opto-electronic method called electrophoresis and a second more exploratory one based on acoustophoresis.
The electrophoresis method is based on optical heterodyning measurement. It has lead, after several design iterations, to the development of a unique prototype named Wallis and its associated software called ZetaQ. It uses a dip cell design combined with the use of amorphous carbon electrodes and a high time resolution acquisition card to improve ZP measurement resolution and accuracy by a factor 3 to 10 over existing solutions.
Acoustophoresis is an original method that relies on the use of an acoustic sound wave to monitor the charges of the particles. Despites its promise, this path has been stopped at the board demonstration level, since the development lead time would not be compatible with the project timeline. Nevertheless, this work has opened up new technology opportunities, which will be further extend beyond the demonstrator.
To summarize, we successfully developed and validated an innovative fiber remote DLS probe. This new Vasco Flex was successfully integrated in the SAXS sample chamber. For SAXS, we successfully integrated the Metaljet and Nanostar, while the Montel optics was adapted to the Ga radiation and the beam collimation was upgraded with scatterless pinholes. Regarding the ZP metrology tool, artifact free measurements were demonstrated, with improved LDE technology and enhanced high resolution, which are also cost-effective.
Real-time measurements
The goal was to speed up the measurement time of the techniques (ZP, SAXS and DLS) and to show that these fast measurements can be performed continuously, in principle during the whole time needed for a reaction. To ensure extraction of relevant data over a long period, the goals are to optimize the computer architecture, to ensure that the data analysis follows the correlations and trends over a long period and to run the techniques DLS, ZP, SAXS under similar hardware and software platforms.
During the course of the project, Cordouan and Bruker AXS successfully achieved all these goals and demonstrated combined SAXS and DLS measurements during a long term measurement. The ZP device was connected at the end of the NP production and characterization line, when the nucleation and growth of NPs has been completed. All measurements are running on the same computer.
Furthermore, an exposure time of 10 s for an individual SAXS measurement was made possible. For DLS measurements, a user-friendly and straightforward interface was developed, allowing for automatable measurements combined with real-time measurements.
Integrated measurement platform for NP
We first created combined measurement cells, integrated the different techniques onto one platform and finally performed simultaneous measurements with the integrated techniques.
The goal of the first task was to design and test a measurement cell around which all the characterization techniques, SAXS, DLS and ZP, can be integrated to simultaneously measure the NPs. The original design provided for up to 10 DLS/ZP heads to determine the synthesis status at several locations along the flow. However, in the course of the project, this concept turned out to be not feasible due to economic reasons, practicability in the further industrial use and space constrains related to the integration of the DLS/ZP heads into the SAXS sample chamber. Hence, a new design of the cell was proposed.
This new design was based on the consideration that, to gain a full picture of the synthesis process, three synthesis steps comprising the nucleation, growth and functionalization has to be monitored. Consequently, 3 probe heads assigned to one of these synthesis steps, respectively, would be sufficient. Furthermore, as the ZP method is a mean to measure the stability of the NP suspension, ZP is applied at the end of the flow, because only at the end of the production the stability of the “end product” is of interest. These considerations led to a 3-capillary cell concept. Between the capillaries flow-reactors can be interconnected to supply the NP suspension with additional agents to functionalize the NP, and to accelerate or stop the growth. The measurement cell was designed, including the option to incorporate 3 capillaries. However, in the course of the project, a single-flow capillary was implemented and connected to a flow-reactor. It consists of a Quartz glass capillary with a wall thickness of 100 – 200 m, which is hold in place by a plastic frame. The cell can be aligned perpendicular to the laser beam, in order to accurately measure the flowing NPs.
Secondly, we integrated SAXS, DLS and ZP around the measurements cell. This implies the integration of the DLS device into the SAXS vacuum sample chamber (see Figure 1). In order to achieve this goal, a novel DLS head was designed especially for the purpose of performing simultaneous measurements together with SAXS, as described above. Thus, in other words DLS is brought to the sample position. This concept of the novel DLS device differs significantly from the standard functionality of state-of-the-art devices, where the sample is placed inside the device. This in turn makes the state-of-the-art DLS devices not practicable and feasible for in-situ and real-time measurements.
The outcome of the development is a novel DLS device, which fulfils all requirements and is named “VASCO Flex”. The main feature of the VASCO Flex device is an optical fiber remote head. This makes it versatile applicable, even beyond the Snow Control project. The device was mounted on the XY-table of the chamber together with the capillary holder as illustrated in Figure 1. As mentioned above the ZP device is applied only at the end of the production process. Therefore, ZP is not integrated into the SAXS chamber, but the ZP measurements are performed separately from SAXS and DLS after the production process.
The successful integration of the VASCO Flex DLS head allowed performing simultaneous real-time DLS and SAXS measurements. The complete system is shown in Figure 2. During these measurements the raw data were transferred to a common computer and analyzed automatically. SAXS and DLS were providing complementary information on mean radius, size distribution and hydrodynamic radius. To optimize the measurement results, the optimal sample geometry and sample volume were found. The above mentioned single flow-through capillary is sufficient for both SAXS and DLS. For SAXS, the capillary walls were sufficiently thin so that scattering and absorption effects are minimal. To optimize DLS measurements, refraction and reflection of the laser beam was minimized by proper alignment of the capillary.
For SAXS data processing, self-written software was used in the PYTHON programming language. For DLS and ZP measurements and data processing, proprietary software named ‘NanoQ’ and ‘ZetaQ’ was provided by COR. All software runs under WinXP on a common computer.
Simultaneous measurements with SAXS and DLS were successfully demonstrated at the same capillary location. SAXS measurements were unaffected by varying the flow rate of the NP, DLS measurements however varied depending on the different flow rates. Therefore, the DLS setup was modified so that the measured NP displacement is perpendicular to the flow direction. Under those conditions, simultaneous measurements were possible under flow conditions. Additional TEM measurements were performed to validate the SAXS and DLS results. This work resulted in a publication of A. Schwamberger et al. in the journal ‘Nuclear Instruments and Methods in Physics Research B’ in January 2015.
Development of materials and standards
We first optimized the NP models synthesis and subsequently defined the standard operation procedure related to the combined measurement methods.
For the first task, different types of metallic and metal oxide NPs were produced using several methods and were tested for the integrated NPs measurement platform. The NPs include colloids of different sizes as well as core shell heterostructure particles. The produced NP systems were chemically modified at their surface using either polyethylene glycol chain (PEG), a biocompatible polymer, or Bovine Serum Albumin (BSA) as a large biomolecule.
KU Leuven has prepared several model synthesis routes, including synthesis in batch reactors and micro-flow-reactors. The synthesis in micro-flow-reactors is focused on silica NPs and on gold NPs. To briefly summarize, we have described the methods used to generate standard NPs of BSTO, gold and silica. We have also shown the capability of functionalization of the NPs and to create an extra outer shell around NPs. To obtain these results, we have done synthesis experiments on different NPs and analyzed them with TGA/FTIR, TEM, SAXS and DLS.
Secondly, we developed standard measurement protocols for characterization of the NPs prepared. This procedure will include the three integrated techniques DLS, SAXS and ZP that operate simultaneously as described above. Furthermore, the coupled TGA/FTIR system is presented. The goal of this task is to develop specific procedures that exclude systematic errors and lead to optimal resolution and accuracy. The performance of a SAXS system integrated with DLS and ZP must produce as good and reliable data as when the same system is a standalone system. Since the sample conditions are different form the standalone situations, new procedures must be derived to ensure that the same level of precision and accuracy can be maintained. Furthermore, care must be taken to ascertain that the different techniques do not interfere, or interact with the sample in an irreversible manner.
A comparison was made between the measurements of the standalone systems, together with measurement systems integrated into single system. The measurements were performed on a set of standards based on silica NPs. These silica NPs were extensively characterized.
Small angle X-ray scattering (SAXS)
The major difference between the two systems is the sample preparation itself, and related background correction. With standalone SAXS measurements it is standard to take a background measurement of the buffer where the NPs are dispersed in. The NPs are also extensively washed before characterization and as such, the background is pure and free from most of the by-products. This ensure a good background subtraction and hence a better understanding of the properties of the NPs. However, for combined measurements extra measurements need to be performed before accurate background corrections can be done. The reactants need to be mixed together in order to have the background without the NPs, but with the same by-products. Since this is impossible in a continuous flow setup, a more complicated process is needed to correct for this. Two protocols are described to ensure good SAXS measurements in a flow/synthesis setup: 1) a simplified background correction, with the assumption that the background is more or less a weighted average between the different solvents and reactants, and 2) an exact background correction, applicable when the NP synthesis process is temperature driven.
Dynamic light scattering (DLS)
In a standalone DLS measurement, a measurement is started only after the sample is loaded in the center of a cell, ensuring a good alignment, which is followed by optimization of the laser intensity and determining the sample concentration. Five runs of 60 seconds are used, to allow for a statistical correlation of the data and fitting with the cumulants and Pade-Laplace algorithms. The real-time measurements with the VASCO-flex are completely different from the standard standalone system. Therefore, new protocols were designed in order to deal with the technical hurdles. The major challenge was the unknown (and changeable) position of the flow capillary compared to the VASCO-flex head. Since the flow heavily influences the DLS spectra (see below), the system was designed in such a way that the flow of the NPs was perpendicular to the laser beam and detector plane, which minimizes the influence of the flow on the spectra. After this alignment, the measurements can be done in the same way as with the standalone system.
Zeta-Potential (ZP)
As described above, it was decided to install the ZP system at the end of the synthesis process, when the particle size is stabilized to the targeted value, as a final control of the process, and before storage or deposition process with injectors. Therefore, the sample is collected at the end of the line and ZP measurements are done via the standalone system. After a thorough cleaning of the measurement cell and measuring the exact pH value, ZP measurements can be performed. A minimum of 7 runs per sample is used to establish measurement repeatability.
Thermogravimetric analysis coupled to Fourier transform infrared spectroscopy (TGA/FTIR)
In this project, we found that TGA/FTIR is about the only method to estimate the amount of functionalized groups around a NP. Simultaneous TGA/FTIR analysis is done via the Netzsch 449 A – Bruker Vertex 80V FTIR thermal analyzer (see Figure 4), which yields a thorough understanding of the decomposition mechanism of the ligands/molecules attached to the NPs. This is essential to optimize the synthesis conditions to obtain phase pure products and to estimate the amount of functional groups.
TGA is a method of thermal analysis in which changes in physical and chemical properties of materials are measured as a function of increasing temperature (with constant heating rate), or as a function of time (with constant temperature and/or constant mass loss). The Netzsch system, with its 25 ng resolution, allows us to accurately calculate the weight losses in different stages during the measurement, with just a few milligrams (1-3 mg) of the NPs required for a meaningful result. Since sample preparation is crucial in the outcome of the measurements, it is ensured that the NPs are either centrifuged or dried (if in a liquid state) to minimize the effect of solvent on the TGA profile.
TGA-FTIR coupled measurement is an interesting way to combine quantitative and qualitative analysis of the various functional groups present in the sample. In a typical measurement, the gases evolved from the TGA heating are fed through a transfer line to a gas cell and are analyzed by an FTIR detector. Thus, the different functional groups evolving at different temperature regions can be traced in real time. By comparing the observed FTIR spectra extracted at different temperatures with the FTIR spectra database, we confirmed the presence of the functional groups (typically organic) attached to the NPs.
Assembly of the NP proto-type production line and demonstration of its reliability and automation
A NP production line was connected by a NP production device to the measurement site inside the SAXS chamber. The NP synthesis is accomplished by a micro-flow-reactor. The synthesized NP are pumped into the capillary to be measured and characterized by SAXS and DLS. After passing the capillary, the NP are measured with the ZP device. The system is schematically illustrated in Figure 3.
Subsequently, the NP production line was demonstrated together with the different real-time characterization tools. For this demonstration SiO2 NPs were synthesized. The real-time capabilities of the system were tested on the synthesis of SiO2 NPs. In this context, the accuracy and the precision of the determination of the size and size distribution were investigated. To demonstrate the long term reliability of the measurement platforms, long term experiments were performed. During this period, the extracted parameters were monitored and displayed on a screen in real-time.
To achieve automation, a dedicated software was implemented to monitor the NP size and size distribution in real-time. The software features an automated background correction algorithm and a graphical user interface, which displays the NP size and the standard deviation of the size distribution as function of time.
Owing to the ability to track the NP size for different synthesis conditions and as function of time, it is possible to investigate the influence of different synthesis parameters on the size of the NPs. The graphical user interface (GUI) facilitates the investigation and makes the optimization of the synthesis conditions more efficient as it displays the NP size in real-time. Different synthesis parameters were investigated. By this means, one can find optimal synthesis parameters for a certain size.
For a proper assignment of a certain size to a synthesis parameter set, high accuracy and precision are required. The accuracy represents the degree of conformity of the measured NP size to the true NP size. To estimate the accuracy of a SAXS measurement the obtained SAXS result is compared to a TEM result. It was successfully demonstrated that for fully grown NPs, the SAXS technique provides size values in good agreement with the TEM results. Therefore, one can conclude that SAXS has a high accuracy (+/- 0.5nm).
The precision is the degree to which consecutive measurements give the same or similar results. For SAXS measurements the precision depends on the signal-to-noise ratio. In turn, the signal-to-noise ratio depends on the X-ray flux at sample position and the exposure time. The noise in a SAXS curve leaves/allows freedom in the fitting procedure and extraction of the NP size. Therefore, a high noise level leads to less reproducibility and thus to low precision. To investigate the influence of the noise on the reproducibility of the results, an experiment with different exposure times was performed. Based on the outcome, it was concluded that the reproducibility of SAXS is ultimately within +/- 0.5nm. Therefore, the high accuracy and precision of SAXS allow investigation of the influence of different synthesis parameters. It can be assumed that the size values displayed in the GUI correspond to the true size of the NPs inside the capillary.
Moreover, long-term measurements over at least 16 hours were performed to demonstrate the reliability of the real-time monitoring procedure. The experiment was divided into two parts: the combined SAXS, DLS and ZP measurements were performed during the day with variable synthesis parameters, followed by overnight measurements, during which these parameters were kept constant.
With this task, the feasibility of real-time monitoring of NP production was demonstrated. Owing to the high accuracy and precision of the determination of the size and size distribution it is possible to track the temporal development of the NP size inside the capillary. This results in the possibility to investigate the influence of different synthesis parameters. The optimization is facilitated by a dedicated software including an automated data processing and a graphical user interface which displays the size and size distribution in real-time on a screen. By this means, the optimization becomes efficient and fast. Consequently, synthesis parameters can be assigned to a certain NP size. In turn, this results in the ability to synthesize NPs of a particular size on request. Furthermore, reliability of the monitoring process and the SAXS device was demonstrated over at least 16 hours. During all experiments the devices were running without problems.
End-user evaluation
The performance of the individual metrology methods was evaluated from the perspective of an end-user and potential customer. This evaluation was done on the prototype tools developed in the project.
The task had two parts. First the specifications and testing/measurement procedure of the different metrology tools developed within the Snow project, namely the DLS, ZP and SAXS, were gathered. The main technical specifications for the primary characterization of the NP (accuracy, resolution and repeatability in the extraction of the diameter) was compared to the measured specifications obtained on the toolset actually installed at KU Leuven. In this task mainly data obtained on gold NPss from Cytodiagnostics Analytics are used for comparison. Second, the user-oriented parameters (ease of maintenance and maintenance cost, long-term reliability, total cost of operation and economic viability of the system) are estimated. This aimed at doing an evaluation of the cost of ownership by looking at the maintenance and repair costs during the project. Similarly, this evaluation is based on the experience of the industrial partners together with KU Leuven employees with the tool.
The interest of IBM in the SAXS tool as an end-user is two-fold. The purpose of the visit for evaluation was first to estimate the brilliance of the Excillium source, and then to evaluate whether the tool installed at KU Leuven could potentially be used for first GISAXS tests. The comments are below:
•
Excillium source: A simple protocol has been defined to estimate the flux of photons emitted by the Excillium source, and to compare it with the rotating anode generator from Bruker AXS installed at IBM premises. Since the direct beam could damage the GADDS detector, the intensity of the direct beam has been measured at KUL using the detector and a sandwich of Al and Zn foils. For this measurement, the source had been operated at 200W and with a beam size after pinholes of about 0.5x0.2mm. The exact same absorber has been mounted on the D8 Discover at IBM Zurich, and the direct beam measured as well to compare for the respective intensities. Assuming a similar detection efficiency for the GADDS and scintillator detector, and assuming a 50% larger absorption for a CuKx radiation compared to GaKa, the flux estimated on the Excillium source is about 1-2.1010 photons/s. This is about a factor 5 lower than the RAG – albeit with line focus instead of the point focus used here – making it an interesting alternative for our application, in particular for microdiffraction on test pads.
•
GISAXS tests: As a preliminary test, a thin film sample has been prepared, made of a thin, epitaxial layer of BaTiO3 directly grown on silicon. The composition of the oxide matches the one that is developed in the NP thin film tool. The sample has been mounted in the SAXS chamber, although the holding stage was by far not optimized for GISAXS geometry. Despite this not suited geometry, a clear GISAXS signal could be observed, and fringes observed on the GADDS detector, related to the finite size of the BaTiO3 sample along the growth direction.
A second part of the task was to evaluate the standard operating protocols used during the simultaneous measurements. Indeed, the experimental condition under which a sample - such as a NP solution - is measured with the individual SAXS, DLS, and ZP techniques is quite different from the experimental conditions needed to enable a simultaneous measurement.
There is also a substantial risk to make systematic errors due to not optimal measurement cells. Clearly, the obtained results from a combined measurement that includes SAXS must be as good as those obtained from an individual SAXS measurements.
The result of this evaluation shows that simultaneous DLS and SAXS measurements at the same capillary position and the same time and real-time monitoring of NP synthesis are possible. While SAXS measurements under flow conditions are still reliable, using DLS under flow conditions however is more challenging task and need a precise alignment of the measurement geometry perpendicular to the flow direction.
To summarize, the different tools have been evaluated by IBM and they performed as expected, in particular in a research environment. As far as we could conclude, the main weakness of the installed tool set is essentially related now to the additional requirements for a combined analysis using DLS and SAXS. It can be foreseen that a “routine” analysis using the combined method will require a redesign of the sample holding stage and of the alignment unit. Furthermore, a multi-sample stage should be designed for such a combined DLS/SAXS tool, so that a semi-automatized analysis routine can be envisioned. In this case, two independent alignment stages for SAXS, resp. DLS will have to be designed. For the moment indeed, the design of the stage allows for multiple samples to be loaded for SAXS, but the DLS alignment stage can only be focused on one single position. An independent alignment stage for the DLS head would enable multiple samples to be loaded for both SAXS and DLS.
2. Development of a thin film deposition system for NPs.
For the development of the second integrated measurement platform, a number of novel methods needed to be developed first. A novel laser based sensor was developed, to measure the flux of NP. A novel high vacuum system was designed and fabricated, followed by the development of a novel NP vaporizer. These tools were combined and integrated into a complete NP thin film deposition system. We performed simultaneous measurements, followed by a demonstration of the reliability and automation of this NP thin film system. The work was finalized by a public demonstration and end-user evaluation. The results of each step will discussed in the following paragraphs.
Development of novel VUV systems
The main goals were to measure the flux of NP in the vapor phase, as well as to make a diffraction pattern from NP thin films. However, it became clear during the studies of the first year, that for the latter undertaking a coherent source of VUV light is necessary. This was outside of the capabilities (technical, scientific and financial) of the project. Hence, most of the efforts concentrated on a non-coherent VUV source, as well as on an UV diode laser (coherent) source. The advantages and disadvantages of both approaches were studied and, taking into account the technical and financial capabilities of the project, it was decided to use a diode laser as an excitation source.
To test the idea of the flux measurements, DCA designed and manufactured a system to measure the atomic Ga flux with a diode laser as an excitation source. The Ga source temperature was varied between 1000 °C and 1200 °C and the measurement system was capable to detect an atomic Ga fluorescent flux signal proportional to the temperature. After extensive designing and testing, DCA developed a compact, single chamber flange flux sensor (see Figure 5) with associated flux sensor software, which allows to control the fluorescence sensor and to perform data acquisition.
Newly designed vacuum measurement system and NP vaporizer
DCA designed and built a new vacuum measurement system (see Figure 5). Designing the system was a challenging and time consuming task for DCA. Only load lock and main parts of the sample manipulator were standard DCA components. The process chamber was designed from scratch in close partnership with Kemstream. The sample manipulator was modified to be able to handle high pressure regimes, with a presence of various gases and liquids that might be used during the process. This, together with the requirement to reach temperatures up to 1000 °C, took a lot of testing and assembling hours. Measurement systems which are integrated to the process chamber, are a RHEED analyzer and the laser based NP fluorescence sensor, which DCA developed as described above. Operation of the RHEED analyzer was demonstrated during initial silica NP depositions. Operation of the NP sensor was demonstrated later when the NP thin films were grown, as described below.
Kemstream built a modified version of Vapbox 1500 NP injector and all its associated accessories (hardware and software, see Figure 5). All these elements were delivered to DCA. Kemstream provided technical support to DCA for the integration of NP injector on the newly designed DCA vacuum system.
Integration of the newly designed tool into one measurement platform
When DCA was assembling the integrated measurement platform, it also integrated the laser based fluorescence sensor and the RHEED analyzer. After performing first functional tests of the NP injector installed on DCA vacuum system, DCA and Kemstream provided training to KU Leuven in the use of the system. Following the training, first silica NPs injection and deposition on silicon wafers were done by DCA (Markku Rajala), KU Leuven (Bert de Roo) and Kemstream (Hervé Guillon). Finally, DCA delivered the complete system to KU Leuven, to perform simultaneous measurements, which was the main goal of the last task within this WP.
These procedures took a lot of developing, testing and assembling hours. Before the system could be installed, the necessary safety measures needed to be in place, as prescribed by the KULeuven team of Health, Environment and Safety (HES). The details of the installation as well as the different modes of operation and maintenance were discussed and tested. In general, the following precautions were taken to avoid NPs to become airborne: 1) examination of the possibility to deposit a transparent TEOS-based Si-oxide film on top of the NP film; 2) installation of an additional exhaust in the lab with a movable exhaust hood and a filter cabinet, where filters can be replaced without having to expose the filter to the lab air. Additionally, special safety measures were taken to protect the user when loading a sample into the system. The necessary safety procedures were derived from IBM’s procedures and were adapted to the specific Snow Control toolset, where a special plastic glove box was installed in the system.
Assembly of the integrated measurement platform for NP thin films together with a thin film production line
The assembly of the thin film system was mainly performed by Kemstream and DCA. With this production system, standard thin films can be grown by DLI-CVD (Direct Liquid Injection Chemical Vapour Deposition). The thin film production system used for this demonstration was a low pressure DLI-CVD system put at the disposition of the project by DCA for the purposes of these demonstration activities. The system consists of several metal organic precursor lines, a DLI vaporizer, pressure and flow controllers, temperature regulation, pumps, precursor inlet valves, purge and oxygen gas lines, etc. These components were all attached to the newly designed vacuum system. SiO2 thin films were chosen as standard DLI-CVD thin films. Two different silica precursors were chosen in this work: TEOS (tetraethyl orthosilicate) and DADBS (diacetoxyditertiobutoxysilane).
In a first set of experiment we varied the different deposition parameters like time, temperature and pressure, and compared these the corresponding changes in thickness and roughness parameters of the thin films, which were measured by X-ray reflectivity (XRR) measurements (via X’Pert Power X-ray platform at KU Leuven) and Scanning Ellipsometry (SE) measurements (at IBM Zurich).
Specific fringes in the XRR spectrum indicate a clear temperature-thickness and time-thickness relation. This indicates that a higher substrate temperature results in a higher thickness as well as longer deposition times in a higher thickness. The CVD growth is in the kinetic regime where the growth rate is limited by the temperature. At a given deposition temperature, the film thickness linearly increases with the deposition duration.
When we want to use this precursor and its inherent high decomposition temperature together with NPs some problems may arise with the thermal stability of the NPs. Thus we want a “better” precursor, i.e. one that decomposes at a much lower temperature. This is the reason we changed our precursor to DADBS.
Similar to the first experiments with TEOS, we varied the different synthesis parameters and correlated these to difference s in thin film thickness and roughness. Again, specific fringes in the XRR spectrum indicate a clear temperature-thickness relation. It was concluded that a higher substrate temperature results in a higher thickness. The growth is in the kinetic regime where the growth rate is limited by the temperature. Similarly, it was shown that a longer time result in higher thicknesses. At a given deposition temperature, the film thickness linearly increases with the deposition duration.
To summarize, we showed that we successfully integrated the measurement platform on a working thin film production line. With this system we could grow silica thin films from different precursors, namely DADBS and TEOS. With TEOS, we could only grow at an elevated temperature of around 900°C. When we used the precursor to DADBS, we could grow at 600°C, which implies very significant diminish of the thermal load.
In order to further lower the deposition temperature, it was decided to add an ozone line. The ozone reacts with the precursors and reduced the decomposition temperature. For this purpose, DCA developed a system consisting of dedicated gas lines with an ozone generator, mass flow controller, pressure gauge and pneumatic valves. With this system oxygen gas enriched with about 10% of ozone can be introduced into the system.
Simultaneous measurements
Simultaneous measurements with the flux sensor and RHEED turned out to be difficult since the pressure requirements are not compatible. The RHEED filament would burn out too fast. However, sequential operation can be done whereby first the NP’s are deposited and then the chamber pressure is reduced so that the RHEED measurements can be performed.
Once the system was operational at KUL, CdSe/ZnS core-shell type quantum dots (QDs)were successfully used, in combination with a silica thin films. The deposition was monitored in real-time using the newly designed NP fluorescence flux sensor. RHEED was used to analyze the structure of the NP thin films. On the fluorescent RHEED screen, different features and spots could be seen. These spots correspond to the different crystal structures available on the surface of the substrate and from the NPs themselves. These data also confirm the crystalline nature of the QDs NP’s.
The GI-SAXS experiments have also taken place subsequently after the films have been deposited. From the GI-SAXS measurements we can obtain information about the size of the NPs and the ordering of these NPs onto the surface of the substrate. The measured size of NPs corresponded with the expected values. The results were also confirmed by AFM measurements.
To summarize, the instrumentation developed for the integrated measurement of NP deposition in thin films such as the laser based flux sensor, the RHEED and the GI-SASX all work well and deliver results as expected. The complete system is illustrated in Figure 7.
Demonstration of the reliability and automation of the NP thin film proto-type production line
We successfully demonstrated the NP thin film growth inside the DLI-CVD thin film deposition system. For this demonstration, two different thin film growth procedures were used. Firstly, gold NP thin films were deposited. Secondly, BTO NPs were used. To achieve automation, a dedicated software was implemented to monitor the gas and liquid flow in real-time. The software featured a GUI, which displays the gas (setpoint, actual flow, actual tON) and liquid parameters (setpoint, actual flow, actual tON) as function of time. If the deposition parameters were unchanged, the different flow rates remained constant. If deposition parameters were changed, a jump was observed in the flow versus time graphs. Additionally, the temperature and the resistance was followed on the screen. Additionaly, the Kemstream Control Software Vapview was able to program a predefined recipe. The deposition was started manually by the experimenter.
Owing to the ability to track the flow for different deposition conditions and as function of time, it is possible to investigate the influence of different deposition parameters on the thickness of the films. The GUI facilitates this investigation and makes the optimization of the deposition conditions more efficient as it displays the flows in real-time. Different deposition parameters were investigated. Therefore, one can find the optimal deposition parameters for a certain film thickness and structure.
Long term measurements were performed over at least 24hrs to assess the reliability and stability of the deposition system. The temperature of the substrate holder, pressure of the chamber and the NP flow were measured as a function of time.
Secondly, gold NP were deposited on a silica substrate. Different deposition parameters, like duration, vaporizer temperature, wafer holder temperature and solution flow rates, were varied during different experiments to account for difference in deposition of gold NPs. The gold NPs were clearly visible after deposition on the silica substrate.
The NP thin film was also characterized by XRR and AFM measurements, confirming the presence of gold NP in the thin film. However, from these measurements it became clear that the gold NPs were not dispersed in a homogenous way, but formed clusters of different sizes. In addition, a difference in height was observed, which varied from few nm to a 10-fold increase in nm.
To resolve these issues, new NP thin films were grown in which the vaporizer temperature and flow rate were varied. The NP thin films were again measured with XRR and AFM, which revealed a more homogenous distribution of NP as compared to the previously grown films. Furthermore, the height of the NPs clusters was smaller and the dispersion was higher. These experiments suggest that depositing gold NPs during a longer time at a smaller flow rate could be beneficial. Finally, commercially available BaTiO3 (BTO) NPs were deposited on a silica substrate. After optimization of the total amount of solution that was injected by varying different deposition parameters like solution flow rate and time, a homogenous distribution of NP was obtained. The results were confirmed by XRR, AFM and SAXS measurements.
Of note, these results differ from the data obtained with gold NPs, which implies that the suitable conditions for one type of NPs does not have to be the same for a different kind. We found that, in general, deposition of BaTiO3 seems to be more uniform
Demonstration of the complete system
The demonstration was given on Thursday the 4th of June 2015 during the final Snow Control meeting in Leuven for all the different partners. The assembly of the NP production line was connected with the NP thin film deposition system. CdSe/ZnS core-shell type quantum dots were chosen as NP to visualize simultaneously the flow measurements, as described above. This was done because no NP made with continuous flow system has been fluorescent so far, and therefore impossible to detect with the NP flux meter. This technical hurdle implies also that we are unable to show the complete system during the synthesis of NP. However, since the production process parameters are mimicked in this setup, the results will not change when we change to a NP production setup.
The main difficulties to prepare the demonstration arose with the connection between the NP synthesis flow reactor and the NP injection system. For this connection new equipment needed to be assembled and one of these tools is a new top flange of the fluid canisters. Indeed, to allow continuous injection in the Vapbox/DCA machine of NPs suspension synthesized and delivered by the Syrris microreactor flow system, a new top flange needed to be developed. Before, the system consisted of two lines: a 3-position valve at the solvent port and a 2-position valve at the pressurization port. To be able to inject solution from the Syrris reactor, an additional port needed to be installed on top of one of the NPs suspension tanks. Therefore, a new top flange was developed including this extra port (which is also a 2-position valve). This connection is suited for the 1/16” diameter of the tubings.
The assembly of the NP production line is shown in Figure 6. For the public demonstration, the Syrris pumps were connected to the NP thin film system. Due to the impossibility to move the SAXS to the NP thin film platform and vice versa, SAXS measurement systems were not included in this demonstration. The complete system demonstration is done with CdSe/ZnS core-shell type quantum dots, with a diameter around 5 nm and stabilized with octadecylamine ligands as used before. They have a fluorescence signal λem @ 600 nm (solid) and were deposited inside the deposition chamber. These QDs are chosen because of their clear emission peak at 600 nm falling in the range of our NP flux sensor (580-627 nm).
The different parts of the NP thin film system were connected with the microfluidic reactor. The connecting liquid line is indicated with a red arrow. When the flow was running, the injection of the NPs into the system was started. The flow of these NPs was also monitored with the NP flux meter. The deposition is monitored in real-time using the NP flux sensor. The fluorescence signal is detected by the photomultiplier and the signal is converted to voltage by the high-accuracy voltage meter.
End-user evaluation
The NP production system has been evaluated with respect to basic parameters relevant in a production environment. Within the limited data set acquired with the system fully integrated, the long term stability and the reproducibility of the experimental conditions have been quantified. Although more time is required to establish the characteristics of the integrated system as a production unit, the basic requirements are fulfilled, in particular for the long term stability of the different components. Finally the cost of ownership for such an integrated system was estimated, on the basis of maintenance costs experienced during the project. The cost of parts necessary amounts to about 13K and required a down time of about 8 days over the last 18 months.
Potential Impact:
The Snow Control project was a large project wherein a lot of metrology tools were developed. These tools were then assembled together into two novel real-time integrated characterization systems for the properties of NP and deposited NP thin films. To illustrate the impact of the Snow Control project on European research and industry, we will describe how the project led to new collaborations and future research projects, but also how the project enhanced the current scientific knowledge on the different metrology tools. Furthermore, we will discuss how Snow Control led to new products and associated increases in business and sales of the different SME involved.
One demonstrator system has an integrated real-time measurement platform linked to an automated NP generation system. The real-time characterization methods developed work in an automated production cycle, which uses combined DLS, ZP and SAXS to automatically characterize NPs of the selected size. This project opened new markets for interdisciplinary sectors, such as nano-medicine, where custom designed reactors can be adapted for the synthesis of NP in the search for new molecules and drugs. Based on the Snow Control Results, we applied with partially the same partners of the Snow Control project, namely KUL, Cordouan and Bruker AXS, for an Horizon2020 NMP-11 project relating to ‘Nanomedicine for cancer therapy’. The project successfully passed the first evaluation phase and is currently in preparation for the second phase submission.
Furthermore, Cordouan participated in a FUI project ‘SAXSize’ project and Horizon2020 project ‘Improcore’, both of which are under evaluation and are based on the work done within Snow Control. The newly developed DLS probe ‘Vascoflex’ was very interesting for the maturation of their current technology, since this tool enables to perform DLS measurements on ‘in bottle’ NP. For this new market segment it is very important that measurements can be be done without exposing the NP to the environment, and therefore it is applicable in bio-applications.
For future research possibilities, Bruker AXS will further investigate the flowthrough capillary, in order to improve the transmission characteristics for X-rays. The combination with DLS will be explored for a better alignment in height, automation and the simultaneous display of SAXS and DLS results. Therefore, the unique combination of DLS and SAXS opened the horizon for new potential customer interests, and is waiting for some additional investments to bring the combined tools to the market. For the NP synthesis, correlations between SAXS and DLS will be investigated, including focusing on monodisperse NPs, their functionalization and also on core-Shell NPs.
The other demonstrator of the project is a thin film deposition system for NPs, opening a wide range of applications. An obvious application is the deposition on a substrate of catalyst particles for CNT or nanowire growth. We also expect this system to present a breakthrough for the production of photonic or photovoltaic devices with tailored band-gap on cheap substrates.
Both research paths paved the way for further research opportunity and currently additional projects are in preparation. Furthermore, the systems developed are the first ones available worldwide with the above mentioned capabilities, therefore assuring an immediate high interest from potential end users as well as the possibility to develop new materials with tailored properties and characteristics across different industrial sectors. This is already obvious from the high number of orders received for the tools developed within Snow Control by the industrial partners, as discussed below.
Snow Control was an accelerator lever for R&D of new products. For the SAXS systems of Bruker AXS, Snow Control significantly enhanced the knowledge on the MetalJet and Nanostar technology, leading to increased costumer interests and sales. This is translated into increased business by the sale of 4 new SAXS systems (0,5 M€ per system) to new clients, implying a significant increase in sales revenues of about 2 M€. These 4 systems were installed in Singapore, Bratislava, Denmark and Germany. Also, one Metaljet system for diffraction was sold.
Cordouan was able to expand their portfolio by two new products developed within Snow Control, namely the VASCO flex (remote in situ DLS) and WALLIS system (ZP analyzer). So far, Cordouan already sold 11 Wallis and 5 VASCO flex units. This implied a revenue in their sales which is worth about 0,5M€ turnover. Furthermore, they expanded their capacity in staff by creating 2 engineer positions created by the means of Snow Control. Finally, Cordouan is involved in several on going scientific publications with reference labs using their VASCO flex, contributing to the broad range dissemination potential of Snow Control.
For DCA, they invested a lot in the development of novel methods, namely in the newly designed NP flux sensor, which is proven to work very well both for atomic fluxes as well as for fluorescent NP fluxes. DCA will further develop it for use in molecular beam epitaxy, and thereby enhancing its application potential. A high number of interested customers have already been identified. The sensor will also be expanded for accurate flux measurement of group V elements in quaternary compounds like InGaAsP and AlGaAsSb, taken into account the limitations of currently available laser diodes.
For KEMSTREAM, the new VAPBOX design was a very important investment. It expanded their current direct liquid injector (DLI) technology to include NP solutions into the injection process. This opens new avenues for their DLI – CVD technology. The injector was then assembled into a thin film deposition process together with DCA and resulted in the first NP thin films grown with this method. This first-of-a-kind tool opens new markets in many of the thin film areas where the innovations that become possible with NP can now be implemented into practical solution.
List of Websites:
https://fys.kuleuven.be/vsm/
Prof. Jean-Pierre Locquet
Department of Physics and Astronomy
Celestijnenlaan 200D, B-3001 Leuven, Belgium
Phone: +32 (0)16 327290
Fax: +32 (0)16 327983