TTI, Inc. – Never Short on Solutions

Technical Seminar Series: Capacitor Q&A

AVX

Q: Where is the future of Metallized Film Technology headed?

Chris: The range of Metallized Film Capacitors is extremely wide, including SMT, Leaded and Power Capacitors. Leaded and Power continue to be expanding application areas, but the latest activity being seen is in SMT series as Pb-free systems are emerging. Many SMT parts are b ased on PET or PPS dielectric — at higher Pb-free reflow temperatures, the PET types can overheat, so PPS will emerge as the SMT dielectric of choice. Full details of AVX PPS chip film can be found at: http://www.avxcorp.com/docs/Catalogs/pps.pdf.

Q: Can you make a comparison between the different technologies in terms of reliability?

Chris: At full-rated voltage and temperature, commercial tantalum (standard and polymer) and MLCC have a reliability rating of 1%/1000 hrs, but for standard niobium oxide (NOJ) this is 0.5%/1000 hrs (2x better) and for Low ESR niobium oxide (NOS) it is 0.2% / 1000 Hrs (5x better). These technologies will typically perform an order better than specification. With the exception of tantalum polymer (which has limited self-healing) these technologies are also available as established reliability for mission-critical applications. Apart from reliability, another consideration is failure mode. Unlike other technologies, the niobium pentoxide dielectric, if taken to breakdown, forms a resistive system with the underlying niobium oxide anode. Up to category voltage, the resistance is of order 10k Ohms, so the parts will have an associated leakage current, but will often be good for capacitance, DF and ESR and can continue normal functionality in most applications. By contrast, tantalum (both standard and polymer) and MLCC, break down to hard shorts.

So, overall, the niobium oxide technology offers the best of both worlds – high intrinsic reliability coupled with a graceful failure mode. For a failsafe solution for MLCC, AVX have introduced a new X7R product family – Flexi-Term. The parts have a polymer interlayer within the termination itself that gives compliance between the capacitor and substrate – under extreme conditions the termination can go high resistance or open circuit, but the ceramic capacitor element itself remains completely intact. There are other solutions in the market place – “margin” caps - which have a reduced active overlap area internally. These parts WILL crack under stress, but hopefully the crack will not go through the active region (this is not 100%!). However, such parts can still break or the cracks can absorb flux residues or moisture that can lead to later failure, so Flexi-term is the recommended solution. Full details of this product can be found at: http://www.avxcorp.com/docs/Catalogs/softterm.pdf

Q: High altitude effects on each capacitor. We have seen "outgassing" on tantalums in Colorado.

Chris: At AVX, we manufacture molded tantalum chip to space level, and we do not have outgassing data available confirming compliance with NASA requirements. This includes the latest QPL CWR19/29 and microchip CWR15. Our Oxicap is manufactured using the same materials, and we are also offering and established reliability version for critical applications. Out MLC Ceramic chips are also available to space level — these are monolithic ceramics with no associated outgassing.

Q: If you can, cover the RoHS timeline and assembly issues for the largercomponents and the TC and interaction at assembly and rework.

Chris: This was actually the subject of another webcast last week on behalf of SMT magazine. For details on downloading the information, you can contact John Hughes: Johnh@pennwell.com. AVX currently has the widest range of fully RoHS compliant passives as standard. For full details visit http://www.avxcorp.com/docs/Catalogs/RoHS_Status.pdf.

Q: Why is there a 10V limit on Niobium Caps, and what do you foresee? It seems that if the voltages can be increased, they will replace tantalums.

Chris: This comes from othe basic chemistry of the material — Niobium is a lot less dense than tantalum. To grow the dielectric on an electrolytic capacitor (Ta2O5 for tantalum, Nb2O5 for Niobium or AI2O3 for Aluminum) you apply a voltage and the dielectric (insulation layer) grows so the surface proportional to the voltage applied. Because Niobium is less dense, it grows thicker for every volt applied. There is a maximum thickness to which you can grow the dielectric film based on the powder particle size used — for solid tantalum, this equates to ~50V, for Niobium ~10V. Work is being done on a next generation product to 16V.

Q: How does temperature influence life time in NbO and MnO2?

Chris: The following link (http://www.avxcorp.com/docs/Catalogs/techsum.pdf) will take you to the applications guide for this product. Relability charts are shown starting on page 10; the effect of temperature on failure rate is shown in figure 3 on page 11.

To calculate the effect of temperature on failure rate, you need to multiply the b ase failure rate by the correction factor for temperature. As can be seen, the tantalum and Niobium share the same characteristic to 85°C; above 85°C the Niobium correction factor is a little higher than tantalum.

However, this base failure rate is important — for commercial tantalum this is typically 1%/1,000 hrs, but for standard Niobium Oxide (NOJ) this is 0.5%/1,000 hrs (2x better) and for Low ESR Niobium Oxide (NOS) it is 0.2%/1,000 hrs (5x better), so even at elevated temperature the Niobium Oxide can have a lower failure rate than standard tantalum and much better than tantalum polymer.

Another area of discussion has been the actual failure mode for Niobium Oxide vs. tantalum. The Niobium pentoxide dielectric, if taken to breakdown, forms a resistive system with the underlying Niobium oxide anode. Up to category voltage, the resistance is of order 10k Ohms, so the parts will have an associated leakage current, but will often be good for capacitance, DF and ESR. By contrast, tantalum (both standard and polymer) break down to hard shorts.

So, overall, the Niobium oxide technology offers the best of both worlds — high intrinsic reliability coupled with a graceful failure mode.

Q: What is the failure mechanism of NbO Caps as compared to tantalum caps which short at breakdown?

Chris: The Niobium pentoxide dielectric, if taken to breakdown, forms a resistive system with the underlying Niobium oxide anode. Up to category voltage, the resistance is of order 10K Ohms, so the parts will have an assoicated leakage current, but will often be good for capacitance, DF and ESR. By contrast, tantalum (both standard and polymer) breakdown to hard shorts.

Also of interest is the relative relability — tantalum (standard and polymer) both typically meet 1%/1,000 hrs — standard tantalum will often be better than this, but polymer is more limited due to lack of self-healing. Polymers can also be more susceptible to reflow temperature, althought the AVX series (TCJ) are qualified to 3x 260°C reflow.

The Niobium oxide Oxicap, by contrast, has a higher reliability designation; 0.5%/1,000 hrs for standard (NOJ) series and 0.2%/1,000 hrs for low ESR (NOS series).

Q: What is meant by no wear-out mechanism? Is it only for the dielectric or for the whole packaged device?

Chris: For Niobium pentoxide (and tantalum pentoxide) dielectrics, there is no degeneration during usage; in fact, the reliability actually increases over time as parts are used. An example of capacitor technology with a wear-out mechanism would be an aluminum electrolytic, where dry-out of the wet electrolyte can take place over time.

Q: What causes ESL?

Chris: This is the inductance caused by the termination loop of the component; the larger the case size, the larger the loop and inductance value. These are the same for Oxicap and tantalum chips.

Q: We've had a history with cracking of MLCC dielectrics. This has been much more problematic as the dielectric becomes thinner. How are you approaching the solution to this with High Caps MLCCs?

Chris: This is a good subject that could be covered in a future webcast. As dielectrics become thinner for low voltage applications, and more layers are stacked for higher capacitance, then parts can be more susceptible to flexure failure (note that the conversion to nickel electrode systems has also improved thermal robustness – AVX apply a 260C solder plunge test as part of lot assessment).

AVX have introduced a new X7R product family – Flexi-Term – to address this type of stress, TCE, & thermal cycling issue. The parts have a polymer interlayer within the termination itself that gives compliance between the capacitor and substrate – under extreme conditions the termination can go high resistance or open circuit, but the ceramic capacitor element itself remains completely intact.

There are other solutions in the marked place – “margin” caps which have a reduced active overlap area internally. These parts WILL crack, but hopefully the crack will not go through the active region. However, such parts can still break or the cracks can absorb flux residues or moisture that can lead to later failure, so Flexi-erm is the recommended solution. Full details of this product can be found at: http://www.avxcorp.com/docs/Catalogs/softterm.pdf

Karun (Murata): I believe you are referring to cracking of the chip itself (i.e. mechanical strength). As internal dielectrics become thinner, there are more layers of electrodes crammed into the same thickness. This actually makes the capacitors more robust. As for the overall thickness of the chip itself, usually its mechanical strength drops as it becomes thinner. However, as package sizes grow smaller, cracking rish due to board flexure drops. Nevertheless, the customer should also make sure their production processes are complying with chip specifications (regarding handling, testing, etc.).

Q: I need higher voltage capacitors to handle telecom type supply voltages for switching supplies. 63V as a minimum. We haven't seen much movement in this direction. What is coming up?

Chris: At AVX, we manufacture telecom capacitors specifically for tip & ring applications — typically 48V line with 250V telco specifications and 250% dielectric strength. We also have our high voltage series of MLC for higher stress applications. Details of both series can be seen at:

Karun (Murata): Murata offers a complete line-up of MLCC caps (including HiCaps) in as many small sizes as possible in the 100V rating for SMPS. Our products are available for input, output, snubber and isolation applications in SMPS. Our products are rated from 50V to 2KV. This is a very popular and well acclaimed line-up. Please contact your local Murata sales reprentative or technical contact for further information.

Q: Where is the best place to obtain comparisons of all these (and other) capacitor types in summary form?

Chris: For a quick reference, the following link (http://www.avxcorp.com/docs/techinfo/dielectr.pdf) will give you a comparison of a number of dielectric characteristics. For more specific requirements by application, please contact me at: (843) 444-2868 or creynolds@avxus.com.

Karun (Murata): At Murata, we are not aware of any independent source which summarizes the comparisons among all these cpacitors. However, there are many short publications and open literature discussing these capacitors. Murata does offer a simple comparison in a tabular format. Our product engineers are also available to discuss such comparisons in greater detail. Please contact a Murata sales representative or technical contact for further information.

MURATA

Q: Murata provides a generous sampling services, but there appears to be longer delay in getting to production with these newer parts. Why is this so, and how can this be improved?

Karun: Since the time required to bring a new product from design stage to production can vary by a lot depending upon the application and our customer (a few months to a few years), Murata's first priority is to provide the design engineers with qualification samples of our latest products. Although we try our best to introduce these parts into full scale production at the earliest juncture, there are times when we are unable to meet our customers' expectations (especially for short design cycles.) Nevertheless, we are trying to constantly shorten our design-to-production lead time.

Q: In Murata's presentation, a slide shows that MLCC capacitance versus DC voltage change — in what application would this be a problem?

Karun: This intrinsic property (a drop in capacitance with DC voltage for Hi-K delectrics) does not cause any functional problems for the circuit. In fact, for most applications impedance is the main paramater of concern (and this effect is barely noticeable). However, for applications where bulk capacitance is necessary or precise knowledge of capacitance is required, the design engineer should be aware of this capacitance drop at the operating voltage (since all capacitances are rated at ZERO DC Bias).

Q: How does temperature effect the life of MLCC at 125°C.

Karun: A capacitor within its specification is expected to have a certain MTTF at a specified temperature. Thus, life at 125°C becomes relative. For example, if we limit our discussion to X7R, then the life at 125 (its max temperature) would be much lower than at 85°C. On the other hand, if this capacitor is used at 150°C (outside its specifications and not recommended), life at 150°C would degrade rapidly. Similarly life for a X8R capacitor would be much higher at 125°C as compared to its maximum operations temperature of 150°C.

Q: Does anyone offer 1mm or below height SMT caps?

Karun: Yes, Murata offers a low-profile series of capacitors. These are available in heights shorter than 1mm. Please contact a Murata sales representative or product manager for further information.

Chris (AVX): At this stage, in a molded package, minimum height is 1.2mm; for height less than 1.0mm, see the TAC microchip series.

Q: What is the relative performance of the various technologies for ESL?

Kaurn: I am not sure I understand the question, but if various technologies means various type of capacitors, then low ESL MLCCs are by far the most superior (i.e. have the lowest ESL).

Chris (AVX): For all capacitor technologies, parasitic inductance is largely a function of the packaging geometry.

Tantalum and niobium Oxide technologies share common molded case sizes and share the same ESL characteristics — by case size:

  • A-1.8nH
  • B-1.8nH
  • C-2.2nH
  • D-2.4nH

For MLCC packages, 1206 is ~ 1000pH, 0805 ~ 800pH etc. However, this inductance can be halved if the parts are terminated on their sides rather than on the ends – this is AVX LICC (Low Inductance Chip Capacitor) Reverse Geometry series – an 0204 package will have ~ 150pH inductance. It may be, for ASIC decoupling, you need low inductance AND high bulk cap. For these applications, AVX have inductance-cancelling technology, the IDC (Interdigitated chip) and LICA (Low Inductance Capacitor Array).

The IDC uses inductance canceling to achieve ~ 130pH inductance while retaining high CV (4.7uF / 6v) with an X7R characteristic.

Full details of these products are available at:

Q: What kind of capacitor would you recomment for the smallest absorption (1µF)?

Karun: I am not sure I understand the question, but if you are referring to dielectric absorption, then our low distortion series or C0G/U2J series are highly recommended. In case higher capacitance is required, avoid using Y5V or Z5U.

Chris (AVX): Although this question was directed to Murata, at AVX we do offer plastic film chip, with PPS dielectric. Typically, NP0 ceramic types will have DA ~ 0.6%, while PPS plastic film will be ~ 0.06 — almost an order lower. This is extremely important in low-leakage voltage discriminator applications. Details of AVX PPS film series can be seen at: http://www.avxcorp.com/docs/Catalogs/pps.pdf.

PANASONIC

Q: Why was the OS can capacitor discontinued?

Robert: Oscon capacitors are manufactured by another supplier.  If you are referring to the WA series, the production cost was to expensive for the part to be competitive.

Q: What is the smallest case size targetted for implementation for SP Caps?

Robert: Panasonic manufacturers all of the SP caps in a D case footprint.  Lowest profile is 1.1mm maximum height.  There are no plans at this point to go to smaller a case size.

Q: Please compare/contrast the EEVFK1E471P to your SP caps (esr, derating, life, capactance vs frequency).

Robert: SP caps are only available up to 16V while the FK series you reference here is a 25V rated part. Generally speaking the SP caps are for very low ESR applications single digits to low teens in a small case/low profile SMT case size, up to a few hundred microfarads. Normally for Dc to Dc converters and decoupling. The FK SMT series is also low ESR when compared to conventional aluminum but much higher than the SP caps for the same CV value. The FK series though is available in higher capacitance and voltage values than the SP Caps. FK series has a limited life depending on the operating temperature and the can size due to the electrolyte drying out. SP Caps have no dry out mechanism so life is very long, limitation is more the max operating temperature. Capacitance versus frequency> both are stable until the self resonant point. This will be lower for the SP caps due to lower ESR and ESL.