Given suitable materials, clean substrates, and a machine with substrate-holder geometry to provide the required distribution accuracy, the main remaining problem is controlling the deposition of layers to achieve the characteristics required by the coating or filter design. Among the various properties of interest, refractive index and optical thickness are the most critical.
At present, there is no satisfactory method for dynamically measuring the refractive index of the portion of a film being deposited. Such measurements can only be made after deposition. For real-time control, practitioners typically focus on maintaining consistent deposition parameters that affect refractive index. This ensures that the resulting index for any given material or mixture remains stable across different deposition runs. While not ideal, this approach is effective in practice. The energetic processes—notably sputtering—are particularly advantageous due to their excellent stability in refractive index and other properties.
In contrast, film thickness can be measured and controlled more readily. The simplest monitoring systems display a signal for the machine operator to interpret, enabling manual termination of deposition at the appropriate time. More advanced systems are entirely automated, requiring little to no operator intervention. In the thin-film community, the term “monitoring” encompasses both measurement and control.
Methods of Monitoring Thickness
The thickness of a film can be monitored by observing a parameter that varies predictably with thickness. Common parameters include:
- Mass: Monitored using a quartz-crystal microbalance.
- Electrical Resistance: Variations correlate with thickness in certain materials.
- Optical Density: Measured via changes in reflectance or transmittance.
- Time of Deposition: Useful in highly stable processes.
Among these methods, optical monitoring (reflectance or transmittance) and mass monitoring with a quartz-crystal microbalance are the most widely used.
Monitoring Accuracy and Method Selection
The choice of monitoring method depends on the required accuracy of layer deposition. Surprisingly, determining the exact tolerances necessary for a thin-film coating is challenging. The monitoring method and allowable tolerances are interdependent. Accurate monitoring directly affects the feasibility of tighter tolerances, and vice versa.
For simplicity, this discussion focuses on common monitoring arrangements, briefly addressing their accuracy. Detailed considerations of tolerances and advanced monitoring techniques are addressed in later sections.
1. Optical Monitoring Techniques
Optical monitoring systems are designed to regulate the deposition of thin films by analyzing the reflected or transmitted light from a test substrate. These systems consist of a light source illuminating a test substrate and a detector to interpret the optical signal. The monitoring system enables precise layer deposition termination by using a shutter mechanism to block the evaporation source instantly. This ensures accurate thickness control compared to merely switching off the evaporation power, which results in delayed material emission cessation. A shutter mechanism is shown in Figure 11.3.
Historical Overview: Visual Monitoring
The earliest optical monitoring techniques relied on the human eye to assess film thickness. Operators visually inspected the color of the film under white light to determine its thickness. For example, a blooming layer of correct thickness typically exhibits a magenta tint due to reduced green reflectance. This visual method sufficed for simple single-layer coatings like blooming, where minor errors had negligible impact. A comprehensive description of this method is provided by Mary Banning, with associated data summarized in Table 13.1.
However, as multilayer coatings gained prominence, the limitations of visual methods became apparent. These methods lacked the precision required for complex designs, prompting the adoption of photoelectric monitoring techniques. Polster describes one such early system, which was based on single-point monitoring.
Single-Point Monitoring Fundamentals
For non-absorbing films, reflectance and transmittance at specific wavelengths oscillate cyclically as a function of thickness. This behavior is depicted in Figure 13.6. The oscillations correspond to optical thicknesses that are integer multiples of quarter-wavelengths. High-precision monitoring systems leverage this behavior to track layer thickness during deposition.
A typical single-point monitoring setup is illustrated in Figure 13.7. A monochromator or narrow-band filter isolates the wavelength of interest, allowing the operator to track reflectance or transmittance changes. For multilayer high-reflectance stacks, reflectance oscillations are monitored layer by layer, as shown in Figure 13.8. Reflectance increases with each high-index layer and decreases with each low-index layer, creating a predictable oscillatory pattern.
While reflectance monitoring is effective for approximately four layers, transmission monitoring is viable for up to 21 layers due to higher signal-to-noise ratios. Beyond this, noise levels often become prohibitive.
Enhancing Accuracy in Optical Monitoring
Skilled operators use chart recorders to track detector output, enabling them to identify turning points with high accuracy. Properly trained operators can achieve monitoring precision within 2% of the target layer thickness. However, this does not necessarily translate to identical precision across the entire batch, as other sources of error may influence final results.
To mitigate noise and improve signal clarity, light sources are typically modulated using a chopper mechanism. The chopper, placed before the light enters the coating machine, reduces stray light and protects the detector from damage caused by excessive exposure. When narrow-band filters are used, careful attention must be paid to suppressing sidebands that might introduce stray signals.
Error Analysis in Single-Point Monitoring
Errors in single-point monitoring arise primarily from inaccuracies in identifying reflectance turning points. For a given reflectance error \( \gamma = \Delta R / R \), the corresponding phase thickness error \( \phi \) can be expressed as:
\[
\phi = \frac{\pi}{2} – \varphi
\]
The reflectance \( R \) is related to the surface admittance \( Y \) via:
\[
R = \left| \frac{Y – Y_{\text{sub}}}{Y + Y_{\text{sub}}} \right|^2
\]
Expanding these relationships yields the error propagation equation:
\[
\gamma = \sigma \phi^2, \quad \text{where } \sigma \text{ is a sensitivity coefficient.}
\]
In practical cases, such as monitoring a quarter-wave layer of zinc sulfide on glass, the fractional thickness error can exceed 8%—highlighting the inherent limitations of this approach.
Advanced Optical Monitoring Techniques
To achieve higher accuracy, alternative methods terminate deposition at points of rapid signal change rather than at turning values. These methods rely on predicting the reflectance or transmittance of the monitoring substrate at the desired thickness. However, challenges such as calibration drift and substrate differences complicate their implementation.
A common solution involves using separate monitoring substrates for each layer. Specialized monitor changers, capable of handling stacks of substrates, facilitate this process. However, using multiple monitors introduces the risk of discrepancies between monitor layers and batch layers.
Giacomo and Jacquinot’s Maximètre
The maximètre, developed by Giacomo and Jacquinot, measures the derivative of reflectance versus wavelength. At reflectance turning points, the derivative transitions sharply from positive to negative (or vice versa). This method achieves sub-nanometer accuracy, typically within 0.2–0.3 nm.
Dual-Wavelength Monitoring
A method devised by Ring and Lissberger involves measuring reflectance at two wavelengths and detecting their difference. By placing the wavelengths at points of maximum slope on the reflectance curve, this technique provides high contrast and precise termination signals.
Broad-Spectrum Monitoring and Ellipsometry
Modern monitoring systems extend these principles to broad spectral regions, utilizing advanced detectors and data analysis. Ellipsometric monitoring, which tracks changes in polarization, offers additional insights into refractive index stability but requires complex high-angle incidence setups.
Single-Point Monitoring in Modern Systems
Despite advancements, single-point monitoring remains widely used in systems like the SYRUSpro (Leybold Optics GmbH). These systems monitor all layers on a single substrate within the batch, ensuring representativeness and control accuracy. Such setups, depicted in Figure 11.17a, continue to deliver reliable performance.
2. The Quartz-Crystal Monitor
Quartz-crystal microbalance (QCM) monitoring is a widely used technique for measuring film thickness during deposition. The piezoelectric properties of quartz allow the crystal to vibrate at a precise resonant frequency, which shifts as material accumulates on its surface, making it an ideal tool for mass-based monitoring of thin films.
Principle of Operation
The resonant frequency of the quartz crystal changes in response to mass deposition. This change is due to the additional mass altering the mechanical vibrational properties of the crystal. The frequency shift is directly proportional to both the square of the initial resonant frequency and the mass of the deposited layer. Two primary methods are used to measure this shift:
1. Frequency Comparison: Compares the frequency of the coated crystal to an uncoated reference crystal.
2. Digital Counting: Measures the number of oscillations over a fixed time interval.
The measured frequency change is often converted into film thickness using film constants provided by the operator. This makes QCM monitoring particularly effective for automated systems.
Crystal Properties
The mechanical vibration modes of quartz are highly complex but can be controlled through precise cutting, dimensioning, and support of the crystal. The AT-cut quartz crystal is the most common for thin-film applications due to its stability over a broad temperature range (-40°C to +90°C). This cut, oriented 35°15′ relative to the z-axis of the crystal, supports a high-frequency shear mode with minimal temperature dependence (±10\(^{-6}\) °C\(^{-1}\)).
Typical QCM frequencies range from 5 MHz to 6 MHz, although frequencies up to 50 MHz are used in specialized applications. Higher frequencies provide better sensitivity but may require more careful thermal management, as temperatures exceeding 120°C significantly degrade performance.
Accuracy and Limitations
The accuracy of QCM monitoring depends on various factors, including crystal stability, environmental control, and proper calibration. Some key considerations include:
– Sensitivity: Sensitivity decreases as the deposited mass increases, limiting the amount of material that can accumulate before the crystal requires cleaning or replacement.
– Calibration: QCMs measure mass, not optical thickness. Calibration must account for the acoustic impedance mismatch between the crystal and the deposited material, which can affect frequency shifts. Commercial systems often allow operators to input material-specific constants for this purpose.
For multilayer coatings, the impedance mismatch at each interface complicates calculations. Using separate crystals for each material can mitigate this issue but adds complexity to the process.
Advancements in QCM Technology
Recent developments have addressed some traditional limitations of QCM monitoring:
1. Multiple-Mode Vibrations: By utilizing multiple vibration modes, modern QCMs can estimate acoustic impedance mismatches more accurately, eliminating the need for user-supplied values.
2. Automatic Crystal Switching: Systems with multiple crystal heads can automatically replace crystals as they reach their mass limits.
3. Distributed Sensors: Placing multiple sensors around the deposition chamber allows for real-time adjustments based on material plume variations.
Comparison to Optical Monitoring
While QCMs excel in stability and automation, they measure mass rather than optical properties, making them less ideal for applications where precise optical performance is critical. Optical monitoring remains the preferred choice for narrowband filters and other designs requiring tight optical tolerances. In such cases, QCMs are often used in conjunction with optical methods, primarily for rate control and source stability.
Applications and Guidelines
QCM monitoring is widely used across various deposition processes. However, to maximize its effectiveness:
– Regular cleaning or replacement of crystals is essential to maintain sensitivity.
– Calibration must be performed for each material and adjusted for temperature variations.
– For processes requiring high precision, combining QCM monitoring with optical techniques is recommended.
Conclusion
Quartz-crystal monitoring is a powerful and versatile tool for thin-film deposition, offering high sensitivity and compatibility with automated systems. Recent advancements in multiple-mode vibration and distributed sensing have further enhanced its capabilities. However, for applications requiring precise optical performance, QCM monitoring is best used in combination with optical methods.
3. Monitoring by Deposition Time
The inherent stability of sputtering deposition processes enables highly consistent control over the thickness of deposited materials. This consistency arises from the incremental deposition methods described in previous tutorials, allowing layer thickness to be controlled effectively by time or by the number of substrate drum rotations.
Time-Controlled Deposition
Time-controlled deposition relies on the stable addition of material over specific time intervals. Successful applications include:
1. Single-Cavity Narrowband Filters: Spencer describes the use of time control in a small ophthalmic coater with sputtering sources of ZrO\(_2\) and SiO\(_2\), demonstrating effective production of optical coatings.
2. Narrowband Notch Filters: Pervac et al. successfully produced filters using pure time control with the Helios machine, shown in Figure 11.12.
3. Edge Filters: Gibson et al. achieved run-to-run reproducibility of ±0.3% for TiO\(_2\) and SiO\(_2\) edge filters in the visible spectrum using a close-field magnetron sputtering system (Figure 11.10).
Figure 13.10 highlights the stability of time monitoring in producing successive batches of antireflection coatings using SiO\(_2\) and Si\(_3\)N\(_4\) with the RAS process (Figure 11.18). The 10-layer design demonstrates the robustness of this approach.
In-Line Sputtering Systems
In-line sputtering systems often use a variant of time control, where substrates move through the system at a constant rate. Key features include:
– Short-Term Stability: Essential for maintaining consistent deposition during each batch.
– Long-Term Drift Compensation: Over extended periods, the deposition thickness may drift. This is mitigated by measuring the thickness of deposited layers and adjusting sputtering power incrementally to maintain process stability.
Advantages and Considerations
Time-controlled monitoring offers several advantages:
– Simplicity: It eliminates the need for real-time optical or mass-based monitoring during deposition.
– Reproducibility: Results are consistent across batches, provided the deposition process remains stable.
– Long-Term Process Stability: Gradual adjustments to deposition parameters ensure reliability over time.
However, time-controlled deposition assumes a high degree of process stability. Any significant variation in deposition rate due to environmental or equipment changes can affect layer thickness accuracy. Regular calibration and process monitoring are essential to maintain desired results.
Conclusion
Time-based monitoring is a practical and reliable method for controlling layer thickness in stable sputtering deposition processes. While it lacks the real-time feedback of optical or quartz-crystal monitoring systems, its simplicity and effectiveness make it a valuable tool for producing consistent optical coatings, particularly in systems with well-maintained stability.
