Machine Design

Measuring out fluids microdrop by microdrop

Combining innovative sensors, precise pumps, valves with no dead volumes, and some elegant engineering lets designers deliver drugs and other fluids in small but accurate doses.

By Andrew Harris
Vice President of Technical Marketing
Burkert Irvine, Calif.


Burkert's Model 2203, a 2/2 way valve, and 6604, a 3/2-way valve, can be integrated with a compensated flow sensor to deliver nanoliter quantities of fluid with ±1% accuracy and 20 msec response times.

The compensated sensor measures flow rates up to 5.5 l/sec with an accuracy better than 2% of the flow rate over a 20 to 50°C temperature range.

Compensating for temperature lets the sensor and valve accurately deliver extremely small quantities of liquid.

More and more lab equipment is being built for medical re-search and diagnosis. And researchers want machines that can handle thousands of samples quickly, precisely, and efficiently. One way to do this is to design machines that use extremely small samples and amounts of reagents — in the submicroliter to nanoliter range. This lets chemicals react faster throughout the sample and reduces the amount of reagent and sample used, both of which can be expensive.

Delivering microdoses of liquids requires accurate flow measurement, small-volume pumps, and valves with no dead volume. Valves with significant dead volumes — the amount of fluid trapped in the valve when there is no flow — could dispense differing fluid volumes if the valve doesn't operate consistently. And fluid trapped in the valve can contaminate or "carry over" into the next sample. Until now, however, reliable interaction between these components has been a challenge in both lab and industrial applications. Engineers have answered this challenge with a microflow sensor that measures pressure differences across a fluidic restriction and controls a miniature valve for ultra-precise fluid control.

The sensor is based on a modified commercial low-pressure microchip sensor. A flow restrictor connects two adjacent piezoresistive low-pressure transducers, and all are part of a micromachined channel. The sensor measures differential pressure across the flow restriction. The pressure difference caused by viscous liquid flowing at low Reynolds numbers across the micromachined channel is expressed as:

in which ΔP = pressure drop, Pa; Qv = the volumetric flow rate, m 3 /sec; C = friction factor; = temperature-dependent, fluid-dynamic viscosity, Pa sec; L = channel length, m; A = channel cross section, m 2 ; and DH = equivalent hydraulic diameter, m. Changing the geometrical parameters A, L and DH affects the sensor's full-scale flow sensitivity, while adjusting the restrictor will ultimately let this sensor measure full-scale flow rates down to 50 l/hr.

The 12 9 3-mm sensor also completely isolates fluids from the electrical side of the sensor, a key feature for microfluidic components.

The valve body is made of polyetheretherketone (PEEK) plastic, but polyvinylideneflouride (PVDF) and ethylenetetraflouroethylene (ETFE) are viable substitutes. It has a Simriz isolating diaphragm that hermetically isolates the actuator from the fluid. A stainless-steel plate may be used to separate the actuator from the coil to prevent dispensed fluids from being heated. These features let the sensor get accurate temperature and viscosity compensation. The rocker or flipper-style valve is frictionless and tight, so it sists clogging and sticking while delivering millions of reliable cycles. The valves can be modified to allow for coil replacement or rotation online without contaminating the process.

The flow rate of this valve/sensor system is a function of the pressure difference across the micromachined channel and the fluid's viscosity, and both are affected by temperature. As these two temperature effects are independent, they must be addressed on different levels: on the chip level for pressure sensor parameters, and on the system level for viscosity.

Pressure-sensor sensitivity and offset variation due to temperature is well known and can easily be compensated so that temperature induced error is less than 1% of full-scale output. An ASIC (application-specific integrated circuit) handles compensation within the valve and sensor. Pressure-temperature calibration is performed by programming in compensation coefficients. The ASIC mounts onto a ceramic substrate, along with the sensor.

Correcting the output flow for viscosity of the pumped fluid requires accurate measurements of the liquid's temperature in the sensor. For that reason, piezoresistors in the pressure-sensor membranes serve also as temperature sensors. The fluid-temperature signal is used externally, along with data on fluid viscosity over temperature (which can be found in the literature), to correct the flow for liquid viscosity. For solvents or mixtures that don't routinely have tables of viscosity data, the sensor is manually calibrated. Using external viscosity correction lets the flow sensor accurately measure flow rates with an overall accuracy better than 2% of the full-scale flow. Response time of the valve and sensor to a flow pulse is less than 2 msec. Total delivered volume may be within 1% of target volume within 20 msec.

The integrated and fully compensated flow sensor and zero-dead-volume valve is suited for precise, high-speed dosing. This system can be used in high-speed liquid handling instruments in which speed, small size, and accuracy are of prime importance. Increasing speed and precision dosing by reducing dead volume makes the system ideal for new pharmaceutical and medical diagnostics equipment such as DNA sequencers, blood analyzers and drug detectors.

The following people contributed to this article: Marc Boillat, Bart van der Schoot, and Bastien Droz, all from Seyonic SA Switzerland.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.