Advancements in base metal electrode capacitor materials & processing technologies extend performance benefits to space circuits
Manufactured for more than twenty years, ceramic capacitors utilizing base metal electrode (BME) technology are routinely designed into applications spanning several industries. Widely employed throughout the automotive industry, BME capacitors have long been qualified to the AEC-Q200 specification, which, by default, became the primary standard used to generalize and compare the performance characteristics of BME capacitors with those of PME (precious metal electrode) capacitors, an essential technology with more than fifty years of proven performance.
The primary advantage of BME capacitors over PME capacitors is their ability to provide high capacitance and a wide voltage range (including lower voltage ratings for digital applications) in packages up to four times smaller than equivalent PME capacitors. Consequently, BME capacitors have gradually replaced PME capacitors in commercial and automotive applications over the last two decades. Space applications have remained the exception, as PME capacitors have been mandated for all mission critical applications. However, as a result of major technological advancements in ceramic materials and processing methodologies in recent years, BME capacitors now exhibit much higher reliability performance than previous generations while retaining higher volumetric efficiency, making them an attractive alternative to PME capacitors in non-mission critical and lower voltage space applications.
One technology driver, particle size reduction, has significantly increased dielectric strength (V/μ) in BME materials systems. Another, improvement in manufacturing capabilities with regard to casting, printing, and stacking extremely thin ceramic layers, has enabled the production of lightweight components with extended capacitance capabilities. Finally, and most importantly, the successful optimization of the X7R dielectric material for the specific process requirements of BME systems has resulted in improved dielectric constant (K) and reliability.
BME capacitors are primarily composed of non-reducing perovskite dielectric materials, typically barium titanate infused with a range of intermediate ionic sized rare earth ions, which are largely responsible for improving the reliability performance of the dielectric material. Unlike PME capacitors, which utilize palladium/silver (Pd/Ag) electrodes and Pd/Ag or Ag terminations, BME capacitors utilize nickel (Ni) electrodes and copper (Cu) terminations. These materials, when co-fired in a reducing atmosphere with the optimized dielectric materials that comprise the latest generation of BME capacitors, form extremely reliable multilayer ceramic capacitors (MLCCs) capable of exhibiting performance characteristics comparable to their PME (Pd/Ag) equivalents, but in smaller and significantly lighter packages, which is critical for space applications since payload launch cost is typically $10,000 per kg.
European Space Agency BME Capacitor Evaluation Project
Space applications require small, lightweight, high capacitance, and high reliability components; and, although PME capacitors satisfy all of these criteria, BME capacitors provide far greater volumetric efficiency (i.e. the capacity to provide higher capacitance in a smaller, lighter package), making them an ideal solution for replacing PME capacitors in non-life-critical space applications.
In 2008, the European Space Agency (ESA) partnered with AVX, a leading global manufacturer of passive components, to evaluate the performance of surface mount ceramic BME capacitors using highly accelerated stress conditions that would enable component evaluation for space applications. The ESA proposed that six BME MLCC capacitors spanning standard EIA case sizes 0603 to 1812 be subjected to a defined reliability testing program designed to overstress the components via temperature and voltage acceleration. The voltage range selected for evaluation was from 25V to 100V, which are common in high reliability applications and well within the currently available BME range of 4V to 3kV.
Figure 1 shows the current BME range, including the subset available for ESA applications, which are shaded blue, and the actual values utilized in the ESA evaluation, which are circled on the chart. The yellow shaded areas represent the current automotive range and the gray shaded areas are voltage ranges currently under development.
Additionally, it is worth noting that, although 0402 BME capacitors were not selected for this initial evaluation, their actual usage in aerospace applications in recent years has increased dramatically.
BME MLCC product design
MLCC product design is based upon four key elements: dielectric layer thickness, end and side margin dimensions, and capacitor cover layer thickness.
Ceramic layer dielectric thickness
BME capacitors currently utilize fired dielectric thicknesses ranging from less than 2µm for low voltage (4V) X5R devices to 80µm for the higher voltage (2kV) X7R devices. Based on these parameters, the 25V to 100V X7R ratings have a dielectric thickness range from 5µm to 18µm, depending on actual voltage rating. The ESA evaluation adopted a conservative design approach by implementing a minimum dielectric layer thicknesses specification above that of the present automotive designs. For example, the present 50V automotive grade would use a minimum dielectric layer of around 4.5µm to 5µm, while the equivalent space part evaluated by the ESA program featured a minimum ceramic layer thickness of 11µm to provide an additional safety factor.
Capacitor end and side margins
Capacitor margins protect the inner electrode structure from the outside environment (side margins) and the end terminations opposing polarity (end margins). Typical margins are designed as small as possible for manufacturing purposes in order to maximize the active area of the electrode plate within the capacitor body, i.e. to maximize volumetric efficiency.
Typically, the minimum designed side and end green margins for a commercial part are around 75µm; for an automotive part, margins would be 100µm. The side and end margins of the BMA capacitors evaluated by the ESA for space level designs were set at 170µm for 25V rated parts.
Dielectric cover layers top and bottom
The cover layers on the top and bottom of the internal electrode stack are typically designed with a minimum thickness of 75µm for a commercial part and 100µm for an automotive part. For space level designs, the 25V parts have a minimum cover layer thickness of 112µm, while the 50V and 100V parts have minimum cover layer thicknesses of 160µm.
ESA test evaluation program
The ESA evaluation program consisted of the four testing groups outlined below.
Group 1: Initial electrical, visual, and dimensional analysis on 25pcs from the six part numbers by selecting random samples from production lots.
· Sub Group 2a: Thermal Shock test on 25pcs from each lot.
· Sub Group 2b: Voltage and temperature stress tests whereby the samples are deliberately over stressed in an increasing step sequence until 50% of the samples have failed.
1. Initial test at 125°C for 168 hours, then increasing multiples of rated voltage (e.g. 4x rated voltage first step for 168hrs and 125°C, 5x rated voltage second step, 6x rated third step, continued until 50% failure rate is achieved).
2. Initial test at rated voltage for 168 hours, then increasing temperatures in 25° increments from 100°C up to 225°C.
Group 3: Elevated voltage and temperature combinations for up to a maximum 2000hrs until 50% of the parts have failed. Actual voltage and temperature conditions based on the output from Group 2 overstress testing (selected by ESA).
Group 4: ESD test on a sample from each of the six part numbers. Test conditions are accelerated well above normal operation conditions and are designed to take the product to failure.
Testing and acceleration factors
The ESA test matrix of six BME MLCC values was selected for testing to maximum of 2000 hours at accelerated temperatures and voltages above standard reliability test conditions. The samples were surface mount assembled to test cards and were subjected to the voltage and temperature conditions in Table 1.
Capacitance (C), dissipation factor (DF), and insulation resistance (IR) measurements were recorded at specified intervals in the 2000 hour maximum test cycle. The test group had three subsets (T1, T2, and T3) that each used different temperature and voltage combinations (defined by ESA) to overstress the components for a maximum of 2000 hours or until 50% of the samples were driven to failure, the primary indicator of which is IR. The maximum temperature used in this group was 150°C (T1) and the maximum voltage was 8x the rated voltage (T3).
The temperature and voltage overstress conditions applied during the steady-state accelerated life test were used to calculate the acceleration factor from the model devised by Prokopowicz and Vaskas. These have been calculated for the six BME MLCC ratings at each of the set test conditions, which are listed in Table 2. These values compare the relative acceleration of tests T1, T2, and T3 to the normal acceleration used for standard life testing (defined as the T0 series 2x rated voltage and 125°C in Table 2).
Primary test results “group three testing”
Table 3 lists the results from test T1, in which the temperature was 125°C and the voltage was 3x rated for the 100V parts, 4x for the 50V parts, and 6x for the 25V parts. Three of the ratings had zero failures despite the severe overstressing.
The results for the three ratings are as follows:
· Within each voltage rating, the parts subjected to the highest V/µm stress during testing (25V rated parts) experienced earlier failures and/or higher numbers of failures than the parts subjected to lower stresses. Similar results were recorded for each test subset T1, T2, and T3.
· The higher temperatures subjected to group T2 (150°C) resulted in an earlier onset of failures for all groups due to the additional thermal energy, which allowed the conductive mobile species to move more easily within the barium titanate structure.
· The accelerated voltages subjected to group T3 also resulted in earlier onset failures in all groups. Voltage stress in excess of the dielectric voltage capability of the fired ceramic (which is influenced by a number of factors, dielectric thickness, the number of grains within each ceramic layer, and grain core/core-shell characteristics) drove the failure mechanism.
Highly accelerated life test data for 1210 50V 1.0µF part
For reference, the 1210 50V 1.0uF rating was selected to correlate accelerated testing to lifetime mean time to failure (MTTF) values. This rating displayed zero failures at 4x rated voltage and 125°C for 2000 hours of test time, as shown in Table 3. This is equivalent to an MTTF value of 16,000 hours at 2x rated voltage and 125°C or 26,000,000 hours at 0.5x rated voltage at 85°C.
The ESA evaluation of a leading manufacturer’s BME capacitors has proven that the long-term reliability performance of these devices under highly accelerated temperature and voltage conditions satisfies the stringent requirements of space applications. The BME capacitor ratings were deliberately overstressed with voltage and temperature to generate data for long-term life performance evaluation. Consequently, the actual accelerated life capability for the ratings tested to 2000 hours shows MTTF values ≥ 26,000,000 hours (or around 3,000 years) for the sample sizes used.
Additionally, reliability performance is only one element in the definition of a space level product. Other elements include wider electrical performance limits (e.g. combining voltage coefficient with temperature coefficient limits) and lot acceptance tests (Group A, Group B, etc.). For future accommodation of these parameters, the design model took a conservative approach compared to the present BME range of capacitors available for the automotive and commercial markets. As such, these designs have incorporated greater dielectric layer thickness and larger margins/cover layers surrounding the electrode stack to provide the devices with additional protection.
Furthermore, the values tested by the ESA were the maximum values for their respective size, voltage, and capacitance categories. Therefore, lower values and ratings are expected to meet (if not exceed) the long-term accelerated life data displayed in this evaluation.
The BME product range described in this article is now available on the ESA website as the European Preferred Parts Listing 2 (EPPL2), which may be accessed via the following link: https://escies.org/epplcomponent/show?id=42326. For more information about BME MLCCs for space applications or the specific AVX components utilized in the ESA testing, please contact Pat Hollenbeck, Field Applications Engineer at AVX at Pat.Hollenbeck@avx.com.