BACKGROUND ART

[0001 ] The present invention relates to calibrating flow sensors, particularly for controlling a mobile phase drive to operate based on sensor signals of calibratable flow sensors, more particularly to a sample separation system such as a high performance liquid chromatography application. ©

[0002] In high performance liquid chromatography (HPLC, see for instance http://en.wikipedia.org/wiki/HPLC), a liquid has to be provided usually at a very controlled flow rate (e.g. in the range of microliters to milliliters per minute) and at high pressure (typically 20-100 MPa, 200-1000 bar, and beyond up to currently 200 MPa, 2000 bar) at which compressibility of the liquid becomes noticeable. For liquid separation in an HPLC system, a mobile phase comprising a sample fluid with compounds to be separated is driven through a stationary phase (such as a chromatographic column), thus separating different compounds of the sample fluid.

[0003] For controlling a pump in a HPLC, the flow rate (i.e. displaced fluid volume per time interval) of fluids may be measured at various positions along a fluidic path. This can be done by flow sensors providing data indicative of the fluid flow and being used for controlling the pump. The signal of the flow sensors may depend on the temperature.

[0004] WO2005/1 13457 A2 discloses a method and an apparatus for monitoring and controlling nano-scale flow rate of fluid in the operating flow path of a HPLC system to provide fluid flow without relying on complex calibration routines to compensate for solvent composition gradients typically used in HPLC. The apparatus and method are used to correct the flow output of a typical, analytical-scale HPLC pump to enable accurate and precise flow delivery at capillary and nano-scale HPLC flow rates.

[0005] US 6,779,712 B2 discloses a flow sensor comprising a substrate with integrated heat source and temperature sensors. Solder bumps are arranged on the heat source and the temperature sensors and the substrate is attached to the outside of a tube using flip chip technology. Preferably, the outside of the tube is structured for being wetted at appropriate positions by the solder. This allows to assemble the sensor easily and accurately.

[0006] However, ensuring a proper performance of flow sensors and may still be a challenge. The accuracy of the sensors may be compromised if such sensors are not properly calibrated.

DISCLOSURE

[0007] It is an object of the invention to enable a reliable operation of a flow sensor in a fluid handling system. The object is solved by the independent claims. Further embodiments are shown by the dependent claims.

[0008] According to an embodiment of the present invention, a calibration device for calibrating at least one flow sensor of a fluid handling system for handling a fluid is provided. The calibration device comprises a fluid drive control unit configured for controlling a fluid drive unit to conduct a fluid through a fluidic conduit in accordance with a predefined fluid conduction protocol. The fluidic conduit is in fluid communication with the at least one flow sensor (i.e. fluid flowing along the fluidic conduit will be detectable by the flow sensor). A calibration data determining unit is configured for determining calibration data for calibrating the at least one flow sensor based on an analysis of a course of flow sensor data measured by the at least one flow sensor when applying the predefined fluid conduction protocol.

[0009] According to another embodiment of the present invention, a fluid handling system for handling a fluid is provided, wherein the fluid handling system comprises a fluid drive unit configured for conducting a fluid through a fluidic conduit, at least one flow sensor configured for sensing a flow rate of the fluid flowing through the fluidic conduit, and a calibration device having the above mentioned features for calibrating the at least one flow sensor.

[0010] According to still another embodiment of the present invention, a method of calibrating at least one flow sensor of a fluid handling system for handling a fluid is provided. The method comprises conducting a fluid through a fluidic conduit in accordance with a predefined fluid conduction protocol, wherein the fluidic conduit is in fluid communication with the at least one flow sensor. The method further comprises determining calibration data for calibrating the at least one flow sensor based on an analysis of a course of flow sensor data measured by the at least one flow sensor when applying the predefined fluid conduction protocol.

[001 1 ] According to still another embodiment of the present invention, a software program or product is provided, preferably stored on a data carrier, for executing a method having the above mentioned features, when run on a data processing system such as a computer.

[0012] Embodiments of the invention can be partly or entirely embodied or supported by one or more suitable software programs, which can be stored on or otherwise provided by any kind of data carrier, and which might be executed in or by any suitable data processing unit. Software programs or routines can be preferably applied in the context of flow sensor calibration and operation. The flow sensor calibration and operation scheme according to an embodiment of the invention can be performed or assisted by a computer program, i.e. by software, or by using one or more special electronic optimization circuits, i.e. in hardware, or in hybrid form, i.e. by means of software components and hardware components.

[0013] In the context of this application, the term "fluid" may particularly denote any liquid, any gas, any mixture of liquid and gas, optionally comprising solid particles. Particularly, analytes in liquid chromatography are not necessarily liquids, but can be dissolved solids or dissolved gases. It is also possible to use fluids such as super critical carbon dioxide (CO2).

[0014] In the context of this application, the term "fluid handling system" may particularly denote any fluidic system having the capability of processing a fluid. For instance, a fluid may be conducted from a fluid inlet towards a fluid outlet. Optionally, at least one fluid processing element (such as a fluidic channel or a chromatographic separation column) may be arranged between the fluid inlet and the fluid outlet. A fluid handling system may be, for instance, a fluid conveying system, a fluid purification system, or a fluid separation system.

[0015] In the context of this application, the term "flow sensor" may particularly denote any sensor being capable of providing information indicative of a flow rate of a fluid flowing through a conduit, particularly in terms of flowing fluid volume per time interval (ml/min) or flowing mass per time interval (g/min). The flow sensor may either directly or indirectly allow to determine the value of the flow rate, i.e. may directly measure the flow rate or may measure another parameter which may then be evaluated so as to derive the flow value. Particularly, such a flow sensor may have a temperature dependent sensing characteristic.

[0016] In the context of this application, the term "fluid conduit" may denote any capillary, groove in a substrate or other lumen defining a path along which a fluid may flow along the fluid handling system.

[0017] In the context of this application, the term "predefined fluid conduction protocol" may particularly denote a specification of a fluid conduction procedure and may particularly denote a workflow, an algorithm or a set of operation parameters defining as to how one or more fluids are to be conducted through the fluidic conduit.

[0018] In the context of this application, the term "a course of flow sensor data measured by the at least one flow sensor when applying the predefined fluid conduction protocol" may particularly denote that the time dependence or development of the value as actually measured by the flow sensor(s) is detected and its value over time may be analyzed.

[0019] According to an exemplary embodiment of the invention, flow sensors or fluid handling systems may be efficiently calibrated or initialized. Particularly, such a sensor may suffer from a change of its default position (such as a zero position, i.e. a sensor signal at zero flow) over time which may deteriorate the sensor performance. By performing the calibration, a stable default position (for instance a zero position at which there is no flow) can be precisely adjusted regardless of external influences such as pressure, temperature, aging effects and the like. For that purpose, a certain defined fluid conduction protocol is applied to the flow sensor which may allow to estimate or calculate an actual flow rate based on the sensor response when executing the fluid conduction protocol. Particularly, specific features (such as points of time at which the sensor signal is non-continuous or changes suddenly) in the course of the flow sensor data can be detected which may depend on a flow rate but also different solvents included in the fluidic conduits in which the flow sensor is arranged. Evaluating such features allows to derive actual flow rate data, particularly without the need of an evaluation of absolute flow sensor values. Such an actually derived flow rate may then be compared with an absolute flow rate which can also be taken from the flow sensor data. Since the latter actually derived flow rate may include artifacts when the flow sensor is not in a properly calibrated state, the difference between the actual flow rate and the measured flow rate is a quantitative measure for the detuning of the flow sensor. However, as an alternative to the comparison of the actual flow rate with the measured flow rate, it is also possible to derive this difference from the flow sensor data only, for example by forcing a solvent to flow firstly in the first direction and subsequently in an opposite second direction. The difference in the sensor performance, as derivable from the course of the flow sensor data, may then also allow to derive an offset value being indicative of the detuning of the flow sensor and being a meaningful value for calibrating or initializing the flow sensor. However, it is common in all embodiments that the course of the flow sensor data is evaluated for the sake of calibration.

[0020] In the following, further exemplary embodiments of the method will be explained. However, these embodiments also apply to the calibration device, to the fluid handling system, and to the software program or product.

[0021 ] In an embodiment, the calibration data is determined by a comparison between a measured flow rate derived from the measured flow sensor data and a real flow rate being determined (particularly calculated by applying a predefined mathematical model of the procedures according to which the fluid is displaced within the fluidic conduit) based on the course of flow sensor data. Hence, the actual or real flow rate may be determined by an analysis of features of the course of the flow rate data such as a discontinuity in the time dependence of the flow rate. Thus, such a real, artifact-free flow rate being derivable from a qualitative, model-based analysis of the course of the sensor data, if desired in combination with other information such as the interior volume of the fluidic conduit or the like, allows to derive the real flow rate. This can be compared to a measured, artifact-including flow rate as can be taken from a quantitative analysis of the measured signal.

[0022] In an embodiment, applying the predefined fluid conduction protocol comprises constituting the conducted fluid from a first solvent and a second solvent, wherein the first solvent and the second solvent separate from one another at a solvent boundary and differ regarding at least one property so as to be distinguishable by the at least one flow sensor. Conducting the fluid through the fluidic conduit may then be performed so as to displace the solvent boundary to pass one of the at least one flow sensor, thereby triggering a sudden change in the flow sensor data upon passing the one of the at least one flow sensor. This embodiment is based on the cognition that at a solvent boundary between different solvents (such as water and an organic solvent like acetonitrile, ACN), there will be a discontinuity in the measured flow rate. This effect results from the fact that the signal of flow sensors also depends on the flowing medium, particularly its heat capacity (heat capacity is, in many cases, the dominating parameter for these type of measurements) and other parameters. Thus, an analysis of the time dependence of the position of the solvent boundary also allows to derive information about the real flow rate, particularly if geometrical parameters such as length and cross-section of the fluidic conduit are known and can be used for the analysis.

[0023] In an embodiment, conducting the fluid through the fluidic conduit is performed so as to displace the solvent boundary to subsequently pass another one of the at least one flow sensor, thereby triggering another sudden change in the flow sensor data upon passing the other one of the at least one flow sensor. Each time the solvent boundary passes a respective flow sensor can be detected in the form of a discontinuity in the sensor signal. Therefore, information regarding the flow rate of the flowing fluid can also be derived from such an analysis of the time dependence of the sensor signals.

[0024] In an embodiment, conducting the fluid through the fluidic conduit in accordance with the predefined fluid conduction protocol comprises firstly conducting the fluid in a first direction and subsequently conducting the fluid in a second direction opposite to the first direction, thereby triggering a sudden change in the flow sensor data (from a positive value to a negative value, or vice versa) upon switching from the conduction in the first direction to the conduction in the second direction. Such an embodiment may make use of the effect that the sign (and sometimes also an amplitude) of the measured flow rate changes when the flow direction is inverted.

[0025] In an embodiment, the predefined fluid conduction protocol comprises a first protocol section and a second protocol section applied after having applied the first protocol section, wherein the first protocol section and the second protocol section differ, particularly only differ, regarding a velocity (for instance adjustable by adjusting a pump pressure) with which the fluid is conducted through the fluidic conduit. Changing the speed according to which the fluid is conducted through the fluid conduits of the system also allows to derive independent information allowing to calculate the real flow rate or at least an offset between the real flow rate and a measured flow rate. Therefore, the fluid flow velocity is another parameter which can be adjusted so as to derive flow rate values.

[0026] In an embodiment, determining calibration data comprises determining a flow rate offset value indicative of a discrepancy between a measured flow rate measured by the flow sensor data and a real flow rate of the fluid conducted through the fluidic conduit. Such an offset value can be directly taken for correcting a measured sensor signal, because it is in fact the quantitative measure indicating to which extent the sensor has to be calibrated.

[0027] In an embodiment, the method comprises calibrating the at least one flow sensor using the calibration data so that a measured flow rate to be determined based on flow sensor data under reference conditions is adjusted to be equivalent to a real flow rate of the fluid conducted through the fluidic conduit under the reference conditions. Particularly, the reference conditions may correspond to a zero-flow state in which the flow rate of the fluid through the fluidic conduit is in fact zero. In such an embodiment, the sensors are calibrated or initialized so as to drive back a flow sensor value, which is detected when a zero flow of fluid is present, to zero. In other words, the sensor value which is obtained when there is no flow can be adjusted based on the determined offset. However, alternatively it is also possible that the reference conditions are different from a zero flow state, for instance are at a certain standard flow rate.

[0028] In an embodiment, the analysis of the course of the flow sensor data comprises determining at least one point of time at which a sudden change in the flow sensor data occurs, particularly determining at least one time interval between two points of time at each of which a sudden change in the flow sensor data occurs. Such an inconsistency or discontinuity in the measurement data is easily detectable by a software by pattern analysis of the time dependence of the detected sensor signal.

[0029] In an embodiment, determining calibration data comprises correlating an event during the course of the predefined fluid conduction protocol with at least one point of time at which a sudden change in the flow sensor data occurs. Such an event may for instance be a switching event from a forward fluid flow direction to a backward fluid flow direction, or vice versa. Another event may be the point of time at which a boundary between two different solvents passes a sensor. Still another event is when the fluid drive unit is operated to change the velocity of the fluid flow from a first value to another second value. [0030] In an embodiment, determining calibration data comprises the analysis of the course of flow sensor data under consideration of preknown information (such as geometric information) regarding the fluidic conduit, particularly a fluid accommodation volume of the fluidic conduit or of a reference section of the fluidic conduit. Since the flow rate can be calculated as a displaced fluid volume (or fluid mass) per time interval, knowledge of the interior volumes of the fluidic conduits which are filled with the fluid or fluids can provide information serving as a support for precisely calculating the real flow rate. [0031 ] In an embodiment, the method comprises conducting, as the fluid, a first solvent and a second solvent through the fluidic conduit along a fluid flow direction, wherein the first solvent and the second solvent separate from one another at a solvent boundary and differ regarding at least one property so as to be distinguishable by the at least one flow sensor, inverting the fluid flow direction of the second solvent (but in an embodiment not inverting the fluid flow direction of the first solvent) so as to move the solvent boundary from a determinable initial position towards a sensor position of the at least one sensor, determining, based on the analysis of the course of flow sensor data, a time interval which the solvent boundary takes for moving from the initial position towards the sensor position, determining the calibration data based on the determined time interval, based on a fluid accommodation volume of the fluidic conduit between the initial position and the sensor position, and based on the measured flow sensor data. Such an embodiment is shown in Fig. 2 to Fig. 6 and is based on a reproducible formation of a solvent boundary in a binary pump arrangement and by the moving of such a solvent boundary from a mixing point towards a known sensor position. In combination with geometrical knowledge regarding the fluid conduit, this embodiment allows to calculate the actual flow rate of the fluid which can then be compared with a measured flow rate (which can be directly taken from the measured flow rate data).

[0032] Still referring to the previous embodiment, the initial position may relate to a mixing point (such as a T-point) at which the first solvent and the second solvent are to be mixed. It is highly advantageous that a mixing point at which usually two solvents are mixed to form a solvent composition for a chromatographic gradient run or the like is used as a reference position at which, in an equilibration state of a forward pumping mode, the solvent boundary is positioned. Thus, starting from such a position allows to start from an accurately reproducible state.

[0033] In another embodiment, the method comprises conducting, as the fluid, a first solvent and a second solvent through the fluidic conduit along a fluid flow direction, wherein the first solvent and the second solvent separate from one another at a solvent boundary and differ regarding at least one property so as to be distinguishable by the at least one flow sensor, inverting the fluid flow direction of the second solvent and adjusting a first flow rate so as to move the solvent boundary from a determinable initial position towards a sensor position of the at least one sensor, determining, based on the analysis of the course of flow sensor data, a first time interval which the solvent boundary takes for moving from the initial position towards the sensor position with the first flow rate, repeating the conducting and the inverting, wherein a second flow rate is adjusted which differs from the first flow rate, determining, based on the analysis of the course of flow sensor data, a second time interval which the solvent boundary takes for moving from the initial position towards the sensor position with the second flow rate, and determining the calibration data based on the determined first time interval and the determined second time interval. Thus, performing one and the same predefined fluid conduction protocol with however two different flow rates (such as "fast" and "slow") of the fluid, it is possible to derive an offset value between a real flow rate and a measured flow rate by an analysis of the course of the flow sensor data with two different flow rates. Such an embodiment also relates to Fig. 7 and Fig. 8.

[0034] Still referring to the previous embodiment, the calibration data may be determined by extrapolating flow rate values according to the determined first and second time intervals to an infinitely large time interval (Δί->∞). Still referring to the above embodiment, two (or more) points on an assumed linear function can be used to derive the corresponding information at a zero flow condition.

[0035] In another embodiment, the method comprises conducting, as the fluid, a first solvent and a second solvent through the fluidic conduit along a fluid flow direction, wherein the first solvent and the second solvent separate from one another at a solvent boundary and differ regarding at least one property so as to be distinguishable by the at least one flow sensor, determining a first time interval which the solvent boundary takes for moving from a determinable initial position via a sensor position of the at least one sensor to a determinable final position, when the solvent boundary has reached the final position, inverting the fluid flow direction of the fluid, determining a second time interval which the solvent boundary takes for moving from the final position via the sensor position back to the initial position (or back to a further reference position which may also differ from the initial position), and determining the calibration data based on a difference between the determined first time interval and the determined second time interval. Still referring to the previous embodiment, an absolute value of the flow rate of the fluid during the conducting and after the inverting may be the same (it may however also be different). Such an embodiment, which is shown in Fig. 9 to Fig. 1 1 , allows to derive the calibration data by firstly moving a fluid forwardly and subsequently backwardly and determining a difference in the flow sensor data.

[0036] In an additional or alternative embodiment, determining the calibration data comprises calculating an integral of a flow rate, measured by the at least one flow sensor, over the time for the first time interval and/or for the second time interval. The absolute value of the flow rate needs not necessarily be the same value. In practice it may even be hard to adjust the same absolute flow rate. Relevant is the knowledge of Integral (flow ^* dt). This value can be calculated from the observed flow over time.

[0037] In another embodiment, the method comprises conducting, as the fluid, a first solvent and a second solvent through the fluidic conduit along a fluid flow direction, wherein the first solvent and the second solvent separate from one another at a solvent boundary and differ regarding at least one property so as to be distinguishable by the at least one flow sensor, moving the solvent boundary from a first sensor position of a first flow sensor of the at least one sensor to a second sensor position of a second flow sensor of the at least one sensor, determining a time interval which the solvent boundary takes for moving from the first sensor position to the second sensor position based on the analysis of the course of flow sensor data of the first flow sensor and of the second flow sensor, and determining the calibration data based on the determined time interval, based on a fluid accommodation volume of the fluidic conduit between the first sensor position and the second sensor position, and based on the measured flow sensor data. In such an embodiment, it is possible to use multiple sensors and to detect discontinuities in the various sensor signals each time that the solvent boundary passes one of the sensors. Therefore, in such an embodiment an initial and a final position may be defined by a position of two sensors along a fluidic conduit, particularly along a dislocated fluidic conduit. Such an embodiment also relates to Fig. 12 to Fig. 16. [0038] In the following, further exemplary embodiments of the calibration device will be explained. However, these embodiments also apply to the method, to the fluid handling system, and to the software program or product.

[0039] In an embodiment, the calibration device comprises a calibrating unit configured for calibrating the at least one sensor using the determined calibration data. Such a calibrating unit may be a processor such as a microprocessor or a central processing unit, CPU. It may be configured as a single processor which also performs the function of the fluid drive control unit and the calibration data determining unit. Alternatively, it is also possible that the calibration unit, the fluid drive control unit and/or the calibration data determining unit are arranged as different processors.

[0040] In the following, further exemplary embodiments of the fluid handling system will be explained. However, these embodiments also apply to the calibration device, the method, and to the software program or product.

[0041 ] In an embodiment, the fluid handling system is configured for supplying, as the fluid, a solvent composition of at least two different solvents having different values of heat capacity. This allows to distinguish the different solvents in a detection signal of a flow sensor.

[0042] In an embodiment, the fluid conduit is a bifurcated fluid conduit. At the position of a bifurcation, a mixing point for mixing two different solvents may be present.

[0043] In an embodiment, at least a part of the plurality of flow sensors is configured for sensing a flow rate in a range between about 1 nl/min and about 100 μΙ/min, particularly in a range between about 10 nl/min and about 10 μΙ/min. Therefore, the sensor array may be specifically configured for microfluidic or nanofluidic applications. The term "microfluidic" may particularly relate to a fluidic device as described herein which allows to convey fluid through microchannels having a dimension in the order of magnitude of less than 500 pm, particularly less than 200 pm, more particularly less than 100 pm or less than 50 m or less. In one embodiment, a corresponding device may have a dimension of 50 pm x 20 pm. The term "nanofluidic" may particularly denote a fluidic device as described herein which allows to convey fluid through nanochannels having even smaller dimensions than the microchannels. In these dimensions, measurement of fluid properties is very difficult and conventional ways of operating a pump may be no more applicable. In the low flow regime of liquid chromatography the flow is controlled with flow sensors. These sensors control in a feedback loop the operation of a pump. If the solvent delivery system has two or more solvent channels, two or more flow sensors may be used.

[0044] In an embodiment, the pumping system is a for instance binary pump (or other, for instance quaternary pump) having a first pump unit and a second pump unit, each having a reciprocatable elements configured for reciprocating within a respective pump chamber, to pump a first solvent supplied from a first solvent container via the fluid conduit and to pump a second solvent supplied from a second solvent container via the fluid conduit, wherein the first pump unit is located upstream of a first flow sensor of the plurality of flow sensors and downstream the first solvent container, and the second pump unit is located upstream of a second flow sensor of the plurality of flow sensors and downstream of the second solvent container. A typical embodiment has the flow sensor after the pump. However, other geometric arrangements are also possible.

[0045] In an embodiment, the fluid handling system is configured as a fluid separation system for separating compounds of a sample fluid in a mobile phase, the mobile phase being constituted by the fluid. The fluid separation system comprises a mobile phase drive as the pumping system, configured to drive the mobile phase through the fluid separation system, and a separation unit, particularly a chromatographic column, configured for separating compounds of the sample fluid in the mobile phase. Such sample separation systems may for instance be chromatographic separation systems or other separation systems such as electrophoresis device. However, in other embodiment, the fluid handling system may be configured as a fluid purification system. [0046] In an embodiment, the fluid handling system is configured as a liquid chromatography apparatus. For instance, the fluid handling system may be a HPLC. A HPLC may be operated at high pressures of for instance between 800 bar and 1500 bar requiring a specifically precise control of a corresponding pump. Embodiments of the invention utilize the sensor signals measured by the sensor array as a basis for controlling such a pump pumping the fluid.

[0047] In an embodiment, a detector is provided which is configured to detect separated compounds of the sample fluid. Such a detector may include a flow cell having an electromagnetic radiation based detection principle.

[0048] In an embodiment, a collection unit is provided which is configured to collect separated compounds of the sample fluid. Such a collection unit may be a fractioner collecting the different separated components of the fluidic sample in different vials or fluid containers.

[0049] In an embodiment, a data processing unit is provided which is configured to process data received from the fluid separation system. Such a data processing unit, for instance a microprocessor or a central processing unit (CPU) may ensure that all the components are properly synchronized during a sample separation procedure.

[0050] In an embodiment, a degassing apparatus is provided which is configured for degassing the mobile phase. Such a degassing apparatus may remove bubbles from the solvents which can be disturbing for the sample separation procedure.

[0051 ] One embodiment comprises a pumping apparatus having a piston for reciprocation in a pump working chamber to compress liquid in the pump working chamber to a high pressure at which compressibility of the liquid becomes noticeable. [0052] The separating device preferably comprises a chromatographic column (see for instance http://en.wikipedia.org/wiki/Column chromatography) providing the stationary phase. The column might be a glass or steel tube (for instance with a diameter from 50 pm to 5 mm and a length of 1 cm to 1 m) or a microfluidic column (as disclosed for instance in EP 1577012 or the Agilent 1200 Series HPLC-Chip/MS System provided by the applicant Agilent Technologies, see for instance http://www.chem. agilent.com/Scripts/PDS. asp?IPage=38308). For example, a slurry can be prepared with a powder of the stationary phase and then poured and pressed into the column. The individual components are retained by the stationary phase differently and separate from each other while they are propagating at different speeds through the column with the eluent. At the end of the column they elute one at a time. During the entire chromatography process the eluent might be also collected in a series of fractions. The stationary phase or adsorbent in column chromatography usually is a solid material. The most common stationary phase for column chromatography is silica gel, followed by alumina. Cellulose powder has often been used in the past. Also possible are ion exchange chromatography, reversed-phase chromatography (RP), affinity chromatography or expanded bed adsorption (EBA). The stationary phases are usually finely ground powders or gels and/or are microporous for an increased surface, though in EBA a fluidized bed is used.

[0053] The sample fluid might comprise any type of process liquid, natural sample like juice, body fluids like plasma or it may be the result of a reaction like from a fermentation broth.

[0054] The pressure in the mobile phase might range from 2-200 MPa (20 to 2000 bar), in particular 10-150 MPa (100 to 1500 bar), and more particular 50- 120 MPa (500 to 1200 bar).

BRIEF DESCRIPTION OF DRAWINGS [0055] Other objects and many of the attendant advantages of embodiments of the present invention will be readily appreciated and become better understood by reference to the following more detailed description of embodiments in connection with the accompanied drawings. Features that are substantially or functionally equal or similar will be referred to by the same reference signs. [0056] Fig. 1 shows a liquid chromatography sample separation system according to an exemplary embodiment of the invention.

[0057] Fig. 2 to Fig. 5 show a bifurcated fluidic conduit with a flow sensor arranged therein, wherein the flow sensor is calibrated by an analysis of the time dependence of the position of a solvent boundary moving over time. [0058] Fig. 6 shows a course of flow sensor data during carrying out the method according to Fig. 2 to Fig. 5.

[0059] Fig. 7 is a diagram showing a time dependency of a measured flow rate for different fluid velocities.

[0060] Fig. 8 corresponds to Fig. 7 and shows a time dependency of the flow rate.

[0061 ] Fig. 9 to Fig. 1 1 relate to a method of calibrating a flow sensor according to another exemplary embodiment of the invention in which the calibration is performed by inverting the fluid flow direction.

[0062] Fig. 12 to Fig. 15 show another bifurcated fluidic conduit having arranged therein multiple flow sensors, wherein the calibration is performed by evaluating a time dependence of flow sensor data when a solvent boundary is moved along the bifurcated conduit.

[0063] Fig. 16 shows a time dependency of the measured flow rate in accordance with the fluid conduction protocol relating to Fig. 12 to Fig. 15. [0064] Before exemplary embodiments of the invention will be described referring to the figures, some basic considerations will be summarized based on which exemplary embodiments of the invention have been developed.

[0065] Flow sensors are used to measure and to regulate flow. One of the limitations of flow sensors is the change in response over time. The accuracy is deteriorated or gets even lost. For low flow applications, the change of the zero-flow value can become a dominating error. To overcome this deficiency, the sensor can be re-calibrated from time to time. Recalibration requires the application of zero-flow or a well known flow. In low flow applications (such as a flow rate of nanoliter/min) this becomes increasingly difficult. Temperature changes and tiny leaks may also falsify the result. To achieve higher precision the sensor has to be disconnected in a conventional approach. Disconnecting the sensor is time consuming and the tightness of the connection in the nanoliter/min scale is not always guaranteed when reconnecting the sensor. A real zero-flow while the original system is still connected is difficult to achieve.

[0066] In contrast to this, an exemplary embodiment of the invention allows to calibrate the zero-flow value of the flow sensor without hydraulic disconnection or applying a well known flow. An exemplary embodiment of the invention uses at least two input flow passes, a mixing point and an outlet. This a typical setup in HPLC (High performance liquid chromatography), particularly in FPLC (Fast protein liquid chromatography) . The two-flow passes shall supply two solvents with significant thermal differences. The differences shall be in a way that the flow sensor response is different for these two solvents. Generally, different methods of achieving a corresponding calibration are possible. Some exemplary of them will be summarized in the following:

[0067] Method 1 :

[0068] Step 1 : There is a positive flow in channels A and B.

[0069] Step 2: Channel A applies a small back flow while channel B has a larger positive flow. The solvent of channel B will reach the sensor in A and will cause a change in the detected signal. The volume between mixing point and sensor A divided by the time needed while give a reference flow to calibrate sensor A.

[0070] Method 2: [0071 ] Step 1 : Method 1 is repeated several times at different flow rates. Between the steps the channels are flushed forward to re-establish the initial condition.

[0072] Step 2: The time needed (t) is drawn as a function of flow (Q). The line is fitted and the abscissa crossing of the fit is equivalent to the sensor offset. [0073] Method 3:

[0074] Method 3 can be applied when more than one sensor is present in the system. The solvent front is always generated at the mixing point. Then it is redrawn to one of the inlet sensors until it is detected there and then it is pushed forward to the outlet sensor. The time need is a measure for the real flow. This can be used to calibrate the sensor. If the volume between the sensors is not exactly known, the process can be repeated at different flow rates.

[0075] The skilled person will understand that other than the explicitly described methods can be applied following basic principles of the present invention. [0076] Referring now in greater detail to the drawings, Fig. 1 depicts a general schematic of a liquid separation system 10. A pump 20 receives a mobile phase from a solvent supply 25, typically via a degasser 27, which degases and thus reduces the amount of dissolved gases in the mobile phase. The mobile phase received from the solvent supply 25 is composed of a first solvent A contained in a first solvent container 74 and of a second solvent B contained in a second solvent container 76. The pump 20 - as a mobile phase drive - drives the mobile phase through a separating device 30 (such as a chromatographic column) comprising a stationary phase. A sampling unit 40 can be provided between the pump 20 and the separating device 30 in order to subject or add (often referred to as sample introduction) a sample fluid into the mobile phase using a switchable fluid valve 90. The stationary phase of the separating device 30 is configured for separating compounds of the sample liquid. A detector 50 is provided for detecting separated compounds of the sample fluid. A fractionating unit 60 can be provided for outputting separated compounds of sample fluid. [0077] While the mobile phase can be comprised of plural solvents, it may also be mixed from one solvent only. Such mixing may be a low pressure mixing and provided upstream of the pump 20, so that the pump 20 already receives and pumps the mixed solvents as the mobile phase. Alternatively, the pump 20 may be comprised of plural individual pumping units, with plural of the pumping units each receiving and pumping a different solvent or mixture, so that the mixing of the mobile phase (as received by the separating device 30) occurs at high pressure und downstream of the pump 20 (or as part thereof). The composition (mixture) of the mobile phase may be kept constant over time, the so called isocratic mode, or varied over time, the so called gradient mode.

[0078] A data processing unit 70, which can be a conventional PC or workstation, may be coupled (as indicated by the dotted arrows) to one or more of the devices in the liquid separation system 10 in order to receive information and/or control operation. For example, the data processing unit 70 may control operation of the pump 20 (for instance setting control parameters) and receive therefrom information regarding the actual working conditions (such as output pressure, flow rate, etc. at an outlet of the pump). The data processing unit 70 may also control operation of the solvent supply 25 (for instance setting the solvent/s or solvent mixture to be supplied) and/or the degasser 27 (for instance setting control parameters such as vacuum level) and may receive therefrom information regarding the actual working conditions (such as solvent composition supplied over time, flow rate, vacuum level, etc.). The data processing unit 70 may further control operation of the sampling unit 40 (for instance controlling sample injection or synchronization sample injection with operating conditions of the pump 20). The separating device 30 may also be controlled by the data processing unit 70 (for instance selecting a specific flow path or column, setting operation temperature, etc.), and send - in return - information (for instance operating conditions) to the data processing unit 70. Accordingly, the detector 50 may be controlled by the data processing unit 70 (for instance with respect to spectral or wavelength settings, setting time constants, start/stop data acquisition), and send information (for instance about the detected sample compounds) to the data processing unit 70. The data processing unit 70 may also control operation of the fractionating unit 60 (for instance in conjunction with data received from the detector 50) and provides data back.

[0079] In the following, the detail of the pump 20 and the portion of the liquid chromatography system 10 surrounding the pump 20 will be described in further detail.

[0080] The first solvent container 74 includes solvent A such as water. A second solvent container 76 includes the other solvent B, particularly an organic solvent such as acetonitrile (ACN). Fluid conduits 88 such as capillaries conduct the respective solvents through the fluidic system of Fig. 1. Furthermore, there is a first flow sensor 82 downstream of first solvent container 74, wherein a first pump unit 64 comprising a piston reciprocating in a pumping chamber is arranged downstream of the first solvent container 74 and upstream of the first flow sensor 82. Correspondingly, a second flow sensor 84 is provided downstream of second solvent container 76, wherein a second pump unit 65 comprising another piston reciprocating in another pumping chamber is arranged downstream of the second solvent container 76 and upstream of the second flow sensor 84. Downstream of a mixing unit 78, the solvent composition is guided through a third flow sensor 86 capable of determining the flow of the solvent composition. Reference numeral 80 denotes all the components forming parts of a calibratable sensor array according to an exemplary embodiment of the invention.

[0081 ] A detailed view in Fig. 1 shows schematically how a flow sensor such as the flow sensor 86 can be constituted according to an exemplary embodiment of the invention. However, the other flow sensors 82, 84 can also be implemented as shown in this detailed view or in a different manner. The detailed view shows fluidic conduit 88 through which a fluid flows. A parabolic or other fluid profile 706 may be formed within the tubular fluid conduit 88 due to friction effects or the like. A fluid flow direction is denoted with reference numeral 708. At a first position within the fluidic conduit 88, a first temperature sensor 700 is provided for detecting the fluid temperature here. Further downstream, a second temperature sensor 702 is provided for detecting the fluid temperature there. Between the temperature sensors 700, 702, a heating unit 704 is provided for supplying a defined amount of heat to the fluid in the conduit 88. Upon activating the heating unit 704, a thermal energy supply pattern denoted with reference numeral 710 is formed within the conducted fluid. A time-dependence of the change of the signal of the temperature sensors 700, 702 will then allow to calculate the flow of the fluid. The larger the flow, the faster will be the transfer of the additional thermal energy supplied by the heating unit 704 to the downstream temperature sensor 702.

[0082] Several components of Fig. 1 also serve as a calibration device for calibrating the flow sensors 82, 84 or 86 of the fluidic system 10. A fluid drive control unit 99 of the calibration device and is configured for controlling the pumps 64, 65 as to how to conduct the respective solvents A, B through the fluidic conduit 88. The fluid drive control unit 99 applies a predefined fluid conduction protocol, more precisely instructions indicative thereof, to the pumps 64, 65 so that the solvents A, B are pumped through the fluidic conduits 188 in accordance with this fluid conduit protocol. As a result of the execution of the fluid conduction protocol, the solvents A, B are pumped through the system 88 and will also form a solvent boundary close to the mixing point 78 at which they are usually mixed. As a result, flow sensor data will be generated by the flow sensors 82, 84, 86. This flow sensor data, i.e. measurement data indicative of the time dependence of the flow rate as detected by the flow sensors 82, 84, 86, can be supplied to a calibration data determining unit 66. The calibration data determining unit 66 is in bidirectional communication with the flow sensors 82, 84, 86 for receiving flow sensor data from them as well as transmitting calibration data to them. Furthermore, the calibration data determining unit 66 is in bidirectional data communication with the fluid drive control unit 99 to exchange signals between these two units 66, 99. The calibration data determining unit 66 determines calibration data for calibrating one or more of the flow sensors 82, 84, 86 based on an analysis of the time dependence of the flow sensor data measured by one or more of the flow sensors 82, 84, 86 when the fluid conduction protocol is supplied to them. The flow sensors 82, 84, 86 may then be calibrated or initialized accordingly.

[0083] Apart from determining calibration data, the calibration data determining unit 66 simultaneously operates, in the shown embodiment, for calibrating the sensors 82, 84, 86 using the determined calibration data. Alternatively, a calibration unit for actually performing the sensor calibration may be foreseen separately from the calibration data determining unit 66. [0084] Summarizing, the predefined fluid conduction protocol is first applied to the system. Then, the flow sensor data is measured. Based on the flow sensor data, one or more of the sensors 82, 84, 86 is calibrated by determining a difference between a target sensor behavior (as can be derived from the real flow rate calculated from the flow sensor data) and a measured flow rate (as can be taken directly from the quantitative evaluation of the amplitude of the flow sensor data). This comparison may result in a discrepancy which allows to initialize the flow sensors 82, 84, 86, i.e. to drive them back to output a defined sensor value under reference conditions (such as a 0-point condition).

[0085] Fig. 2 to Fig. 16 show four examples of fluid conduction protocols each of which may be applied to the system of Fig. 1 for performing flow sensor calibration. However, it should be understood by a skilled person that alternatives are possible which are based on the same or similar principles.

[0086] Referring to Fig. 2 to Fig. 6, a calibration method according to a first exemplary embodiment of the invention will be explained. [0087] Fig. 2 shows a bifurcated fluidic conduit 88 wherein a first solvent 202 (such as water) is conducted through a first inlet channel towards a mixing point 78, and a second solvent 204 (such as an organic solvent like acetonitrile, ACN) is supplied via a second conduit upstream of a mixing point 78. At the mixing point 78, the two solvents 202, 204 are brought in interaction and form a solvent boundary 206. The solvents 202, 204 differ regarding physical parameters such as thermal conductivity, heat capacity, etc. which results in different absolute sensor values at the same flow rate depending on whether the first solvent 202 or the second solvent 204 is present at a position of a flow sensor 84. This can be understood by recalling the detection principle of the flow sensors (see reference numerals 700, 702, 704, 706, 708, 710 in Fig. 1 ). [0088] Starting from the forward pumping mode of Fig. 2 in which both solvents 202, 204 are pumped in a downward direction, it is possible to invert the fluid flow direction for the second solvent 204 from an initial direction 208 towards a final direction 300. Upward inverting the fluid flow direction from the forward direction 208 to the backward direction 300 for the second solvent 204 only, the solvent boundary 206 moves from a determinable initial position (in this example the mixing point 78) towards a position of the flow sensor 84. Fig. 3 shows a scenario shortly after the beginning of the backward pumping procedure of the second solvent 204, and Fig. 4 shows a subsequent point of time at which the solvent boundary 206 reaches the position of the flow sensor 84. After having passed the flow sensor 84 the fluid boundary 204 is moved further upwardly, see Fig. 5.

[0089] Fig. 6 is a diagram 600 being indicative of the impact of the procedure described referring to Fig. 2 to Fig. 5 on the sensor signal. The diagram 600 has an abscissa 602 along which the time t is plotted. Along an ordinate 604, the measured flow rate Φ ₅ is plotted as detected by the flow sensor 84. Fig. 6 also shows a time t1 and a time t2. t1 relates to the point of time at which the flow direction is inverted from 208 to 300. t2 relates to the point of time at which the solvent boundary 206 passes the flow sensor 84 (Fig. 4). Discontinuities 608 of the measured flow rate are detected at these points of time t1 , t2. Particularly, based on the analysis of the course of the flow sensor data 606, a time interval At=t2-t1 is determined which the solvent boundary 206 needs for moving from the initial position towards the sensor position. Calibration data for calibrating the flow sensor 84 is then determined based on the determined time interval At and based on a fluid accommodation volume V| of the part of the fluidic conduit 88 between the initial position and the sensor position (both positions are known so that only the cross-sectional area of the fluidic conduit 88 is needed to determine V|). This can be taken from a detail shown in Fig. 2. The real flow rate is than calculated as

[0091 ] A difference (called "offset") between the actual flow rate 0 _rea i (i.e. the correct flow rate at the position of the flow sensor 84) and a measured flow rate sensor (i.e. the false flow rate detected by the uncalibrated flow sensor 84) is then determined by:

[0092] Offset= Osensor - <t>real

[0093] The value "offset" can be directly used for calibrating the flow sensor 84.

[0094] Referring to Fig. 7 and Fig. 8, a calibration method according to a second exemplary embodiment of the invention will be explained.

[0095] Fig. 7 shows a diagram 750 having an abscissa 752 along which the inverse of At is plotted. Along the ordinate 754, the measured flow rate Φ ₅ is plotted. By firstly performing the procedure shown in Fig. 2 to Fig. 5 with a relatively slow ("s") solvent speed and subsequently with a faster ("f ) solvent speed allows to derive two different values 1/At _s and 1/At _f . In other words, the steps of conducting, inverting und determining, as described above referring to Fig. 2 to Fig. 5, can be repeated twice with different speed values of the fluid. The calibration data may then be determined by extrapolating the flow rate values according to the determined values 1/At _s and 1/At _f to an infinitely large time interval At->∞ which corresponds to1/At->0. It can be taken from Fig. 7 as to how the value "offset" is then determined.

[0096] Fig. 8 shows a diagram 600 being similar to the diagram shown in Fig. 6. As can be taken from Fig. 8, curve 600 relating to the fast speed and 606' relating to the slow speed can be captured. The parameters At _s and At _f can be determined based on the time values t-i , t ₂ f _as t and t _2s iow plotted in Fig. 8.

[0097] Therefore, the calibration according to the embodiment of Fig. 7 and Fig. 8 can be performed without the knowledge of any geometrical parameters of the system (such as internal fluid volumes or the like) but can be derived solely on the basis on an analysis of the flow sensor data shown in the diagram 600 of Fig. 8.

[0098] Fig. 9 to Fig. 1 1 show a third embodiment of the method in which, starting from an initial position shown in Fig. 9, a solvent boundary 206 is firstly moved forwardly (see flow direction 208 in Fig. 10), wherein subsequently the fluid flow direction is inverted (see reference numeral 300 in Fig. 11). In the absence of any offset, the flow rates measured in the forward direction 208 and in the backward direction 300 should is identical. However, since this is not the case if the flow sensor 84 is uncalibrated, there will be a difference in the measured flow rate which allows to derive the required calibration data.

[0099] For performing the experiment according to Fig. 9 to Fig. 1 1 , it is possible that the absolute value of the flow rate of the fluid in forward and backward direction is identical so that differences in the measured flow rate only result from the calibration offset. However, it is not a must that the flow rate is absolutely constant in forward and backward direction. Different flow rates can also be compensated for by a mathematical compensation procedure.

[00100] Referring to Fig. 12 to Fig. 16, a calibration method according to a fourth exemplary embodiment of the invention will be explained. [00101 ] In the embodiment as shown in Fig. 12 to Fig. 16, two flow sensors 84, 86 are used for detecting a point of time at which a solvent boundary 206 passes a respective flow sensor 84, 86. Starting from the state of Fig. 12, the solvent boundary 206 is moved downwardly and passes in Fig. 13 flow sensor 84. This corresponds to the point of time t1 in diagram 600 shown in Fig. 16. As can be taken from Fig. 14, continuing the pumping of the fluid will move the solvent boundary 206 downwardly so that, as can be taken from Fig. 15, the solvent boundary 206 then passes the flow sensor 86. This corresponds to the point of time t2 in Fig. 16. By calculating the time interval At-i ₂ =t2-t1 , compare Fig. 16, and knowing the internal volume V| (see Fig. 12) between the positions of the two sensors 84, 86, it is again possible to determine the offset value.

[00102] It should be noted that the term "comprising" does not exclude other elements or features and the "a" or "an" does not exclude a plurality. Also elements described in association with different embodiments may be combined. It should also be noted that reference signs in the claims shall not be construed as limiting the scope of the claims.