Understanding Common Specifications for MEMS Silicon Dies

To evaluate the quality and performance of a MEMS silicon die, customers must rely on specifications, at least until they can test parts for themselves. This article will discuss the most common specifications related to these pressure-sensor dies.

The primary thing to understand about MEMS dies is that when they are exposed to either pressure or temperature, they will produce a corresponding output, which will be in millivolts, provided that an input voltage, or excitation voltage, has been supplied. The millivolt output from the MEMS die is essentially the pressure value. Therefore, the general characteristic to look for in any MEMS die is a stable and repeatable output when the die is tested under various conditions.

This article discusses common specifications used to characterize a pressure-sensor die’s performance under different operating conditions.

The first groups of specifications we will discuss are commonly used to characterize how the MEMS die will perform at room temperature (25 °C).

Bridge resistance (or impedance): This indicates the resistance (from Ohm’s Law the voltage divided by the current) measured across the bridge. Due to our Wheatstone bridge design along with our Sentium® and MeritUltra™ processes, the input resistance (+E to -E) and the output resistance (+O to -O) on all of our dies are the same.

Offset (or zero-pressure output voltage): This indicates the difference, at zero pressure, between zero output and the actual output of the MEMS die. With absolutely no offset, at zero pressure the output would be 0 mV/V. However, with an offset of ±10 mV/V, the difference with 5 volts of excitation could be ±50 mV. Refer to the image of the transfer function below.

Sensitivity (or span): Sensitivity and span are, in general, synonymous. The two terms are used to indicate the electrical output, or the response, of the MEMS die to an applied pressure and supply voltage. It is typically represented by the slope of a line on a graph with output on one axis and pressure (for a given supply voltage) on the other axis. Refer to the image of the transfer function below. Sensitivity is generally stated in terms of microvolts per volt per psi (µV/V/psi).

Transfer Function Graph for S Series 15 psi MEMS Die

Non-linearity (or linearity): This shows how linear/non-linear the output is. The ideal output is perfectly linear. For example, at a constant 5-volt supply, for every pound per square inch that the pressure were to increase, the output in millivolts would increase linearly, as shown in the image of the transfer function above. The pressure non-linearity is calculated by measuring—at the mid-point of the pressure range—either of two differences: One is between the actual output and the best-fit straight line (BFSL) or the other is between the actual output and the invisible line that connects the two endpoints of the actual output. This line is called the end-point line or terminal base. Refer to the image below. The actual output shown in this image has been exaggerated for illustration. Whether the pressure non-linearity is based on the BFSL or end-point line, it is expressed as a percentage of the full-scale output (FSO).

MEMS Die Pressure Non-Linearity Example

Pressure hysteresis: This shows the delta, or difference, of the output at zero pressure and then up to full-scale pressure and back to zero pressure. It would be ideal to have no pressure hysteresis, meaning the output would be the exact same every time the pressure returned to zero. This specification will give you one indication of the die’s repeatability. Pressure hysteresis is expressed as a percentage of full-scale output (FSO).

The next three specifications indicate how a part will behave over a specified temperature range. At Merit Sensor all MEMS dies are tested over a temperature range from -40 to 150 °C.  These three specifications are first-order effects.

Temperature coefficient of offset (TCO): This is also known as temperature coefficient at zero pressure (TCZ). This indicates the offset changes at zero pressure as temperature changes.

Temperature coefficient of resistance (TCR): This indicates how the resistance changes at zero pressure as temperature changes. The bridge resistance does change significantly over temperature.

Temperature coefficient of sensitivity (TCS): This is also known as temperature coefficient of span. It indicates the deviation in full-scale output as temperature changes. As the temperature increases, sensitivity decreases. So at room temperature you might get a 100 mV output, but at 150 °C the output will decrease to around 75 mV.

The great news is that all the errors listed above are repeatable and consistent, which means they respond well to compensation. In addition to manufacturing MEMS dies, Merit Sensor also builds pressure-sensor packages and performs calibration over various temperature ranges.

The following two specifications, however, deal with errors that cannot be compensated: thermal hysteresis and long-term drift. Therefore, if you are trying to decide which MEMS die to buy, you will want to find a supplier that produces parts with good specs in these two areas. We, at Merit Sensor, know that our customers do not want their parts, which contain our MEMS dies, to fail in their customers’ applications; therefore, we take pride in producing MEMS dies with excellent thermal hysteresis values and long-term stability.

Thermal hysteresis: This is typically performed at zero pressure and shows the difference between the output when the temperature is at room temperature and then increased to 150 °C and then returned to room temperature and then decreased to -40 °C and then returned again to room temperature and so on. This testing characterizes the repeatability of the die over numerous thermal cycles. It would be ideal to get the same output every time the temperature returned to a given value.

S Series MEMS Die Accuracy with Thermal Hysteresis - Solid White Background

Long-term stability (or long-term drift): This specification indicates how stable the output of the die will remain, or, in other words, how little the offset will drift, over time and sustained temperature. We have tested parts, for example, at 150 °C for 300 hours.

S Series MEMS Die Long-Term Stability - Solid White Background

One thing to watch for is a data sheet advertising a MEMS die with an accuracy of ±0.25 %. Here’s the catch: That accuracy refers only to non-linearity at room temperature; it does not take the other errors that have been discussed into consideration. Hopefully this article has helped you to better understand the different performance characteristics of MEMS silicon dies and the specifications that are used to quantify the dies’ performance.

Finally, if you would like to learn more about the technology and performance of MEMS dies, we invite you to watch our recently broadcast webinar, which is now on demand.

Three Common Types of Pressure-Sensor Packages

At the heart of every MEMS pressure sensor is a MEMS silicon die. Merit Sensor owns and operates a wafer fab, where it produces all of its own MEMS die. Packaging a MEMS die requires specialized equipment and skills to handle the small and sensitive die and to perform delicate wire bonding. Therefore, many customers purchase pressure-sensor packages, in which the die have already been mounted and wire bonded. This article will discuss Merit Sensor’s three types of packages: uncompensated, passively compensated, and fully compensated.

Uncompensated

The most basic pressure-sensor package is uncompensated. In an uncompensated package the MEMS die has been mounted to a ceramic substrate with a special die-bond material, wire bonded to electrical traces on the ceramic, and covered with a protective cap or gel. Since each silicon die is inherently unique, the output for each one will be unique. Fortunately, silicon die have outputs that are very repeatable. This means the output can be compensated.

PMD Series Pressure Sensor Internal Components

PMD Series Pressure Sensor – Uncompensated

To get an accurate output, the customer will need to perform some degree of compensation. Certain applications lend themselves to compensation performed by the customer. Factors that will often determine whether the compensation is performed by Merit Sensor or the customer include the following:

  • Cost
  • Accuracy
  • Size
  • Output signal

Passively Compensated

A more ready-to-use version of a pressure-sensor package, especially for use at room temperature, is one with passive compensation. In this case the pressure-sensor package is basically the same as an uncompensated package; however, the thick-film resistors on the ceramic substrate have been laser trimmed, providing adequate compensation of the die’s output in operating temperatures between 10 °C and 40 °C.

Laser-Trimmed Thick-Film Resistor on AP Series Pressure Sensor

AP Series Pressure Sensor – Passively Compensated

For applications, such as invasive blood-pressure monitoring in a hospital room, compensation in this temperature range is sufficient. Other benefits of passive compensation are the pure analog signal with practically infinite resolution and frequency response times in microseconds.

Fully Compensated

In a fully compensated pressure-sensor package, signal conditioning (an on-board ASIC) is used to compensate the die’s output across a wide temperature span. A MEMS silicon die does not know the difference between pressure and temperature, so this level of compensation is especially critical in applications where the temperature of the sensing environment fluctuates drastically or reaches extremes highs or lows. Compensation through signal conditioning can provide a linear output and make that output as accurate as ±1 percent of the full-scale output (±1 %FS total error band) in operating temperatures between -40 °C and 150 °C.

Wheatstone Bridge with ASIC

Wheatstone Bridge on a MEMS Die with an On-Board ASIC

If we use fuel pumps in airplanes and fuel rails in vehicles for examples, it is typical for pressure sensors to be exposed to extreme temperatures; nevertheless, it is essential that these pressure sensors offer an accurate output. A fully compensated pressure sensor would be the appropriate solution.

TVC Series Pressure Sensor Internal Components

TVC Series Pressure Sensor – Fully Compensated

It is important to emphasize that each pressure sensor will require individual compensation, as each one will have a unique output inherent to its MEMS die. Many customers simply do not have the time or equipment to do this logistically or economically to each unit passing through their assembly line.

Merit Sensor has the experience and equipment to handle this necessary step for the customer. Furthermore, it often, although not always, makes sense for compensation to be done before the part leaves our facility. Nevertheless, we have left options for those customers who choose to do their own compensation. As always, our sales managers and technical team will be happy to answer any related questions.

Four Characteristics of Our Newest MEMS Sensing Element

Merit Sensor has owned and operated a wafer fab from its beginnings. Fabricating our own MEMS (micro-electro-mechanical systems) sensing elements, or die, is something that sets us apart from other pressure sensor manufacturers, many of whom source their MEMS die from foundries or suppliers. Producing our own wafers, which are diced into individual MEMS sensing elements, allows us to control our own technologies, development, and supply chain. To learn more about the advantages, read this AZoSensors interview with our director of engineering.

Since we continue to see interest worldwide in these MEMS sensing elements, we continue to develop MEMS die with superior performance at competitive cost. Our newest MEMS product on the market is the S Series, offering optimal size, sensitivity, and stability. Perhaps best of all is its excellent performance in regards to thermal hysteresis. Each of these characteristics will be discussed below.

Size

One remarkable feature of the S Series is its solid performance at a very small size: 1.5 mm x 1.5 mm x 0.9 mm. This size also makes it is possible to optimize the amount of die produced on each 150 mm (6 inch) wafer. The end result is a lower-cost die for the customer without any loss of superior performance.

S Series MEMS Die Dimensions

S Series MEMS Die Dimensions

Sensitivity

Silicon, which is the raw material of MEMS wafers, has piezoresistive properties, which means when pressure is applied, it is strained and its resistance changes accordingly. An output is based on the changes in resistance. Merit Sensor uses Wheatstone bridge technology to optimize the linearity of the output. It is challenging, however, to obtain an adequate output when the pressure is low. Nevertheless, through Merit Sensor’s proprietary MeritUltra technology the S Series provides a typical output at 5 psi / 34 kPa / 345 mbar of 100 millivolts (mV).

Stability

A stable part will remain accurate, i.e. it will not drift, over time and sustained temperature. The S Series data sheet specifies a long-term stability of ± 0.2 % of the full-scale output (% FSO). The chart below shows how stable and accurate the part has proven to be, demonstrating a typical offset drift of <0.05 % FSO at 300 hours.

S Series MEMS Die Long-Term Stability

Long-Term Stability of the S Series

Thermal Hysteresis

The characteristic we are really talking about here, once again, is accuracy. In addition to remaining accurate over time and sustained temperature, the S Series displays exceptional accuracy when exposed to thermal cycling. A MEMS sensing element is inherently sensitive to temperature. Its resistance and output will change when temperature changes. Fortunately, changes that are consistent are simple to compensate. The S Series die exhibits very consistent, accurate output when it is exposed to extreme temperatures and returned to room temperature. In thermal cycling tests it demonstrated a typical thermal hysteresis offset of <0.05 % FSO.

S Series MEMS Die Accuracy with Thermal Hysteresis

Accuracy of S Series with Thermal Hysteresis

If you have any questions about using the S Series in your application, contact one of our sales managers. You might also find the application note “Handling of Mounting of Pressure Die” useful.

How to seal an O-ring to a TR Series Pressure Sensor

Merit Sensor offers a fully calibrated, back side pressure, harsh media, pressure sensor for use with any media which are compatible with Silicon, glass, ceramic and solder. This sensor assembly (TR-Series) was designed to be used with an o-ring, creating a face seal to the back of the sensor.

 

There are many technical considerations that need to be evaluated when designing for an o-ring face seal. To ensure that a good design can be achieved during the first round of development, several factors must be clearly defined.   This information will be critical in subsequent material selections (for both the o-ring and the housing into which it will be inserted) and will be required in the subsequent dimensional and stress analysis.

 

Specifications

Temperature Specification

  • Identify the minimum and maximum end use temperatures for both the operation and the storage conditions.  Will the use temperature will be constant or fluctuating? Will the pressure be changing at the same time?

 

 

 

Pressure Specification

  • Identify the minimum and maximum use pressures. Will the pressures be all positive, all negative or a combination of both positive and negative? Will the pressures be fluctuating or constant? Will the temperature be changing at the same time?

 

 

Media Specification

  • Identify the media that will be in contact with the sensor. What chemistries do they contain? Are they compatible with Silicon, Borosilicate Glass, 96% Alumina Ceramic and Solder?   What will be the exposure conditions (temperature, pressure, duration, concentration, etc.) Be sure to think about both sides of the sensor. The backside will be exposed to the harsh media. The front side will be exposed to some other environmental conditions. Be sure that the “top side” is protected from the harsh media.

 

http://www.applerubber.com/src/pdf/chemical-compatibility.pdf

 

O-Ring Options

Material Options

  • The o-ring material should be selected based on the information specified above. The o-ring softness should be selected base on the maximum use pressure and the resulting packaging stresses. A soft o-ring will provide a very compliant seal which will result in very low induced packaging stresses but may not be able to seal well at high pressures. A hard o-ring conversely would seal well at high pressures but may also induce high packaging stresses. Different o-ring materials have different temperature handling capabilities. The glass transition temperature of the polymer will limit the lower functional operating temperature of the o-ring. The temperature at which the polymer begins to decompose or soften will limit the upper functional temperature of the o-ring. It is also important to look at the media compatibility of the different o-ring polymers. The longevity of the o-ring and the amount of swell that the o-ring will experience will be different depending on the o-ring material and the media. It may be difficult to find the exact right material to match all of the specification requirements.

 

http://www.applerubber.com/src/pdf/general-properties-of-orings.pdf

 

 

 

Geometry Options

  • After the material selection, the determination of the o-ring size (OD and cross-section) is the next thing to consider. The o-ring should accomplish several different goals. The o-ring must ensure that the media will not leak at minimum and maximum pressures. The o-ring must ensure that the media does not leak at minimum and maximum temperatures. The o-ring should be chosen to minimize package stress buildup during pressure and thermal cycles.
  • There are several different o-ring geometries that can be used for face sealing. Each of them has advantages and disadvantages. The most common and cost effective o-ring geometry is the standard circular cross-section. This geometry can be used for both positive and negative pressures. To assist with high pressure sealing, backer rings can be used to prevent issues with squeeze-out. In addition to the circular cross-section, there are “X” and “U” shaped o-ring cross-sections. The “U” shaped o-ring comes in two configurations that could work as a face seal (inward facing channel for positive pressure applications, outward facing channel for negative pressure applications). The “X” cross-section will work in either application.

 

 

O-Ring Gland Options

Counter Boar Gland

  • The counter bore gland is the most common o-ring gland. It is relatively simple to design and manufacture. The gland depth and width can be tailored to work with the specific application specifications. Items that need to be considered are the squeeze percentage, the swell and the coefficients of thermal expansion.

 

 

Dovetail Gland

  • The dovetail gland is the most complicated o-ring gland. It is difficult to design and is expensive to manufacture. The primary benefit of this gland design is that it will assist in holding the o-rings in place during assembly. It is not recommended for small o-rings. This design is even more sensitive to the squeeze percentage, the swell and the coefficients of thermal expansion.

 

Suggested Engineering Analysis and Verification

To ensure that the o-ring will seal properly over the full temperature and pressure use ranges, several different analyses should be carried out. It is important to look at static forces, dynamic forces and the effects of temperature on each.

 

Static and Dynamic Analysis

  • It is important to calculate the dimensional changes that will happen with temperature. The OD, ID and cross section diameters of the o-ring should be calculated at the Min and Max temperatures. The width and depth of the gland should be calculated for Min and Max temperatures.   The o-ring squeeze should be calculated at each of these extremes to ensure that the gland dimensions are adequate. Be sure to take into consideration the swell for the o-ring material base on the media in contact with the o-ring. Based on these dimensions, the zero pressure stresses on the package can be estimated.
  • The static model should then be used to evaluate the stresses during changes in both temperature and pressure. Based on the output of this analysis, a suitable combination of o-ring size, o-ring material and gland dimension can be selected to provide the optimal solution.

 

Because each application is a very unique combination of temperature, pressure and media, it is recommended that verification testing be carried out by the customer to ensure that the o-ring material, o-ring cross section and the gland dimensions will provide a robust solution in the final application.

TR Series Pressure Sensor Based Inches of Water Pressure Switch

In many situations there is a need to know the level of a liquid in a tank or the pressure inside of an air duct. Both of these cases are quite low pressure and can be difficult to measure. This can be accomplished in several ways. The simplest is a sight glass or sight tube, as shown below. This works on the premise that liquid in the tank will force liquid up the sight tube to the same level as what is in the tank, or the air pressure being measured will raise the liquid level equal to the pressure applied. A monometer is a commonly used device for measuring low air pressure/vacuum. This is particularly useful in a tank that is not transparent, or at least translucent. In the case of an air duct, there is nothing visible so an external device of some sort is required. Although this is a simple approach, it is not particularly convenient because it needs to be located at the tank or close to the duct being measured. This is not useful if remote monitoring is required, and even less useful if any sort of feedback is desired as it is completely manual.

Tank with Sight Glass

Tank with Sight Glass

If remote monitoring of the level is required, there are several more options. A common example is a float type resistive (potentiometer) sensor, as typically found in an automotive fuel tank level sending unit. These sensors work well, but have some drawbacks.

  • Located in the tank
  • Displace some volume in the tank
  • Moving parts
Tank with Float Type Level Sensor

Tank with Float Type Level Sensor

Tank with float type level sensor

Depending on the media being measured, and the design of the components, this type of sensor can fall victim to malfunctions caused be the media itself. A common issue is the float absorbing the media it is submerged in, which would result in an artificially low level reading because the float will lose some buoyancy.

Tank 3

In order to provide a reliable level sensor, one with no moving parts is very desirable. To accomplish this, a sensor such as Merit Sensor Systems’ TR series could be utilized. The TR series pressure sensor is a piezoresistive MEMS pressure sensing element paired with an ASIC on a ceramic substrate. The sensor is available in many pressure ranges, gage or absolute pressure measurement as well as custom calibration and output.

In order to realize the most accurate level reading possible a gage part should be used. This is preferred because the atmospheric pressure acting on the fluid in the tank will also act on the reference side of the MEMS pressure sensor providing the most accurate reading even while atmospheric pressure changes. In the case of a differential air pressure measurement, the reference can be atmospheric or another space. Some examples of this are:

  • Building duct static pressure measurement (atmospheric to pressurized duct (inches of water typically))
  • Building high duct static pressure (similar to regular duct static, but typically wired into the fan controller to turn off the fans if the pressure exceeds a safe level for the building and air supply system (inches of water typically))
  • Building air filter status (differential pressure across filter – larger differential as filters become blocked (inches of water typically))
  • Building space to space pressure (differential pressure between two spaces to ensure air flow is going in the right direction – common in clean rooms (tenths of an inch of water typically))

Graph 1

Measuring levels, for media such as water or air, is difficult as one inch of water measured at 39°F is a mere ≈ 0.0360911906567 PSI. Merit’s TR series is offered in a high sensitivity (low pressure) configuration that, when calibrated to 5 PSIG, could resolve to 1” of water. It is possible to achieve better resolution with a different calibration and/or custom MEMS device with higher sensitivity.

Below is a plot showing the difference between a 5PSIG calibration and a 1 PSIG calibration. There is a significant difference in the output of the sensor, giving much better resolution.

Once the sensor with an acceptable resolution has been selected there are options for the sensor interface. The TR series sensor provides a linear voltage output of 0.5 to 4.5V from minimum pressure to maximum pressure, and it is temperature compensated. This voltage can be monitored by a system controller or simply connected to a circuit, such as the one below, which would provide a variable level threshold via VR1. This could be used as a low level indicator/alarm or a overfill indicator alarm, depending on the configuration of the circuit.

Basic circuit for variable pressure level switch/indicator

Basic circuit for variable pressure level switch/indicator

Below is an example of the operation of the above circuit. This is an ideal, theoretical, example. With VR1 set to a calculated voltage for the desired level (10” W.C. in this case, or ~1.94V), the output of U1 will go high when the sensor voltage reaches the setpoint of VR1.

Graph 4

Whether a variable signal is desired, or a simple ON/OFF will suffice, a low pressure TR series sensor can be used. Several of the applications discussed here are currently handled by other sensor technologies, but could be handled very well by a properly designed, and implemented, MEMS based sensor.

The Wheatstone Bridge

Introduction to the Wheatstone Bridge

The heart of Merit Sensor’s pressure sensors is a Wheatstone bridge that is comprised of a group of four resistors on a silicon etched diaphragm. As pressure is applied to the diaphragm the resistors are stressed, changing their resistance.

In an ideal setting, all of the resistors would be perfectly matched and completely temperature independent.

In the real world, however, differences exist between the resistance values of each resistor. In addition, temperature also changes resistor values. The change to resistor values and the overall bridge output due to temperature is known as the Temperature Coefficient of Resistance, or TCR.

Many applications require that a pressure sensor operate independently of temperature. In these applications, the pressure sensor’s TCR must be compensated for.

There are two general methods for TCR compensation – passive and active.

In passive compensation, the individual bridge resistor values will need to be measured in order to determine values needed for the compensation resistors.

In active compensation, a microcontroller, signal conditioner or analog circuit records the bridge output across various temperature and pressure conditions and adjusts sensor outputs accordingly.

Bridge configurations

a. Closed – A bridge in which all resistors are connected (See Figure 1).

AN103 Bridge Configuration Options - AN103-001
Figure 1 – Closed bridge

In a closed bridge there is no way to measure individual resistors as there will always be influences from the other three resistors of the bridge.

b. Half Open – A bridge that is divided into two branches and connected at one end (See Figure 2).

AN103 Bridge Configuration Options - AN103-002
Figure 2 – Half open bridge

In contrast to the closed bridge, a half open bridge allows measurements to be taken for each resistor, which is a benefit if the sensor’s performance needs to be determined. A half open bridge also allows for either active or passive compensation to be added as needed.

A half open bridge requires an additional electrical connection.

c. Full Open – A bridge that is divided into two branches, which are open at both ends (See Figure 3).

AN103 Bridge Configuration Options - AN103-003
Figure 3 – Full open bridge

Similar to the half open bridge, the full open bridge allows for each resistor to be measured. In addition to being able to use either active or passive compensation, each half of the bridge can be powered and measured independently. This is beneficial as some signal conditioners commonly used in pressure sensor applications require two independent branches.
However, the full open bridge configuration requires an additional electrical connection beyond that required by the half open configuration.

Examples of Implementations

a. Closed – Since individual resistors can’t be measured in a closed bridge, a closed bridge can be used with active compensation or in an application where sensor output fluctuations due to temperature changes are acceptable.

Figure 4 depicts a closed bridge with active compensation.

AN103 Bridge Configuration Options - AN103-004Figure 4 – Closed Bridge with Interface device (Signal Conditioning ASIC, Microcontroller, Analog circuitry, Etc.)

One example of a suitable application for a closed bridge where temperature independence is not critical is a pressure switch, where the absolute pressure measurements are not as important as knowing you’ve reached a pressure threshold.

b. Half Open – Active compensation can be applied to the half open bridge as in Figure 4 shown above. Passive compensation can also be applied to the half open bridge as seen below in Figure 5.

AN103 Bridge Configuration Options - AN103-005

Figure 5 – Half open bridge with passive compensation

The implementation of a half open bridge with passive compensation in Figure 5 shows the added components as well as the extra electrical connection (Vin+) required to close the bridge. The additional resistors, as they are named, accomplish span, zero and output impedance compensation. These components need to be added after open bridge measurements have been taken at the required conditions.

C. Full Open – The full open bridge has a wide variety of implementations. In addition to being used as a full open bridge it can be used as a half open (Figure 5) or a closed bridge (Figure 4). Figure 6 is an illustration of how a full open bridge could be used for two functions – temperature and pressure.

AN103 Bridge Configuration Options - AN103-006

Figure 6 – Full open bridge with two functions

In this implementation, half of the bridge is being used as a temperature sensor and the other is being used as a pressure sensor. Only half the pressure output signal will be present because there is only the voltage swing of half the bridge. However, this allows for the added benefit of a means to measure actual die temperature. The temperature measurement will allow for a more accurate input for temperature compensation than an ambient temperature measurement would.

Choosing the Appropriate Configuration for your Application

The entire sensing system should be taken into consideration when making decisions about the bridge configuration. First, the user must decide if temperature independence is important and if it is, whether active or passive compensation will be used. If active compensation is chosen and a signal conditioner or other electronic device will be used, that device’s requirements must be met.

Use caution as devices with similar functions may have very different requirements.

As previously discussed, each configuration has its own benefits and drawbacks. The added electrical connections of a full open bridge add to the complexity of assembly but allow for more flexibility as well as the ability to more easily troubleshoot bridge issues.

Ultimately, the choice of bridge configuration should be based on a thorough analysis of the system.

 

 

 

 

 

 

 

Disclaimer Notice
Merit Sensor Systems produces high quality products that perform within the parameters of the data sheet. Typical pressure and temperature performance values are not tested 100% but they have been validated during qualification. Merit Sensor cannot guarantee that the product will function properly after mounting and post processing by the customer. It is the responsibility of the customer to test and to qualify the function of the Merit Sensor product in the final package. Customer is responsible for the required knowledge to handle product and Merit Sensor assumes no liability for consequential damages that may result in yield loss or field failures in the final application.
THE INFORMATION IN THIS DOCUMENT IS PROVIDED TO YOU “AS IS” WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, WHETHER EXPRESS OR IMPLIED. Merit Sensor Systems reserves the right to modify this document and assumes no responsibility for any failures that result from the use of this information. All information provided in this document is for illustration purposes only.

Pressure Sensors for Harsh Environments and Extended Temperatures

Many pressure sensor applications in rugged environments such as aerospace, automotive, industrial, and even medical equipment present developers with conflicting requirements that lead to expensive compromises. These sensors are often used to measure pressure, level, and flow of harsh fluids such as gas, oil, refrigerant, or other caustic solvents that can damage the sensor element. Extended temperature requirements offer additional challenges, even beyond compensation for accurate pressure readings. Automotive and aerospace specifications are especially stringent, with operating temperature ranges as broad as ‑40°C to +150°C. And these rugged applications typically have high reliability and accuracy requirements, since component failure can lead to product recalls or safety risk. To respond, equipment manufacturers rely on expensive ongoing maintenance and component replacement to work around the sensor’s inherent short lifespan.

The packaging of the sensor component is critical in solving this issue, but is a challenge that has, until now, eluded sensor manufacturers. Consider a typical use case. An automotive application such as gas or diesel fuel-line sensing requires a sealed sensor element that can be installed inside the fuel line to detect pressure changes that indicate a plugged fuel filter, which provides a feedback signal to the car’s computer to notify the driver. Similar requirements occur in airplane engine, gear, and valve controls, and in measurement and control of compressors or leak-detection systems in industrial equipment. While medical applications may not require the pressure sensor to operate in fluids as harsh as gasoline, even saline solution can be corrosive over time, and the cleaning and sterilization process often requires repeated contact with caustic chemicals such as bleach.

The problem is that the adhesives that are used to create the pressure seal and protect the sensor die and related circuitry eventually soften in harsh fluids. Once the seal breaks, the sensor circuitry is damaged, creating a common reliability failure that can be expensive if it causes a product recall or demands regular maintenance and replacement of the sensing subsystem.

1-blog-1

 

Figure 1. Sensor package showing back-side entry to protect electronic circuitry from harsh media. [Source: Merit Sensor Systems]

Extended temperature requirements increase the packaging challenge. While some new adhesives may withstand higher temperatures than was possible in the past, they still risk die detachment at pressures of 300 psi, and humidity can destroy the bond strength of most adhesives. There are exotic epoxies that can withstand some temperature and humidity extremes, but storage and application raise additional manufacturability issues, and these epoxies can impact accuracy of the sensing element in extended temperature applications.

Extended temperature requirements increase the packaging challenge. While some new adhesives may withstand higher temperatures than was possible in the past, they still risk die detachment at pressures of 300 psi, and humidity can destroy the bond strength of most adhesives. There are exotic epoxies that can withstand some temperature and humidity extremes, but storage and application raise additional manufacturability issues, and these epoxies can impact accuracy of the sensing element in extended temperature applications.

For a pressure sensor to perform well from ‑40°C to +150°C, it requires a stable MEMS element as well as stable packaging and manufacturing processes. But instability typically occurs because of differences in the thermal coefficients of expansion (TCE) of the MEMS die and the substrate on which it is mounted. Stainless steel might be considered an ideal substrate, but its TCE is much higher than that of silicon. As temperature changes, the metal expands and contracts while the silicon elements soldered onto it experience much smaller changes. The MEMS element reacts to the stresses caused by the TCE differences, triggering errors that look to the system like pressure changes—presenting system designers with a new reliability issue.

An innovative new pressure sensor packaging approach creates a eutectic die bond on ceramic substrate using a gold-tin soldering alloy for a hermetic seal even in harsh fluids, at high pressure, and at extremely wide temperature ranges. The ceramic substrate has a TCE that is close to silicon so there’s no significant thermal mismatch, and gold and tin are common soldering elements that hold up well to harsh fluids. While their high individual melting points can impact manufacturability, a gold-tin soldering bond with an 80:20 ratio creates an alloy with a much lower melting point, which improves manufacturability while still maintaining the benefits of both metals in harsh environments. And while this gold-tin solder is more expensive than adhesive, the cost differential is minute in comparison to the significant improvement in long-term reliability and maintenance costs.

1-blog-2

Table 1. Comparison of pressure sensor package types across harsh application requirements.

An additional element to consider in comparing sensor packaging approaches is whether the pressure media comes in at the top or back side of the sensor. If the pressure is on the top side of the sensor, the circuitry must be protected from shorts or corrosion, which is typically accomplished with a protective gel. But a gel that is stiff enough to withstand corrosive fluids and high temperatures can also be stiff enough to cause stress to the MEMS element, which, again, generates sensing errors. In contrast, back-side entry exposes only the silicon, glass, and the eutectic die attach to the pressure medium—and these elements have been proven to withstand these harsh environments.

Figure 2. Merit Sensor harsh-media, extended-temperature sensors are available with optional ferrule (left), polycarbonate pressure ports (middle) for radial pressure sealing and standard face seal (right). Solder pins are also available to simplify system design. [Source: Merit Sensor Systems]

System developers who require pressure sensing capabilities in harsh-media, extended-temperature applications have discovered that packaging is critical to improve product lifetime reliability and reduce cost of ownership. It is a challenge that has finally been solved.

Scott Sidwell is sensors engineering manager for Merit Sensor Systems. He can be reached at [email protected]. For more information, see https://meritsensor.com.