GB2394191A - Detecting the position of a selected sample component in a chromatography column using ultrasonic signals. - Google Patents

Abstract

A chromatography process comprises detecting the position of a selected component of a process liquid passing through a chromatography column 1 by transmitting one or more ultrasonic signals through the bed space of the chromatography column, the speed and/or the attenuation of the one or more ultrasonic signals being affected by the presence in its path of the component of the process liquid. The selected component may pass through the column as a band 6 and a plurality of ultrasonic signals may be transmitted through the column so as to track the progress of this band along the length of the column. A plurality of ultrasonic transmitters 21 and receivers 20 may be distributed along the length of the column to allow ultrasonic signals to transmitted through the column along a variety of paths. The selected component may be a contaminant, and detecting its movement along the column allows a user to assess the condition of the packed bed material within the column. The chromatography column may be a vertical cylinder with the ultrasonic signals being transmitted radially across the cylinder.

Description

2394191 1

( CHROMATOGRAPHY METHODS AND CHROMATOGRAPHY APL'ARATUS

FIELD OF THE INVENTION

_ _. _ _ _

this invention has to do with methods and apparatus s used in chromatography. The new proposals are concerned with finding out the condition of a bed of a particulate medium in the bed space of a chromatography column. This assessment might be during the,oac:kiny of the column or afterwards i.e. during or after a chromatography process 1() using the packed bed.

BACKGROUND OF THE INVENTION

The conventional principles and apparatus for packed-bed chromatography are well-established. The present invention is concerned with issues arising in 15 c.ommercial-scale preparative chromatography, where columns are often very large and any one or more of the particulate packing meclitm, the product to be separated and the time taken up in arranging the running of the procedure is/are of very high value.

2() IL is well known that a packed column has to be continuously, uniformly anc1 entirely filled with Lee relevant medium under an appropriate degree of compression of its particles.

Conventional column packing, an operation requiring 25 -treat care and experience, involves removing the top plate or "cell" of Lyle column and pouring in a slurry of

the relevant medium in a liquid carrier. More recently the use of a packing port has become favoured because it obviates taking the lid off. Various kind of valved ports have been proposed for this; see for example GB-A 5 2258915, WO 96/10451 and WO 99/64130. The column generally has upper and lower restricted-permeability elements - mesh or sir-ter layers - to retain the media particles, and the packing port provides communication directly into the bed space past i.e. penetrating one or 10 both of these restricted-permeability layers. To pack, typically the column is filled with liquid and the medium slurry pumped in through the packing port at one end, liquid carrier leaving the column via the permeable element at the other. The particulate medium is retained 15 and gradually accumulates until the column is full. The accumulating bed is compressed by the controlled pressure from the pumped liquid and - in the usual case when packing from the Lop - also by the weight of the upper part of the bed. There is a subtle, continuous variation 2() in the conditions experienced by the medium. A c:cmrno procedure is for the operator to continue pumping until the column is essentially full, whereupon the pump usually cuts cut spontaneously. As the liquid flow :'ressure is relieved, the packed bed then "relaxes" to 25 fill the column with a more uniform compression profile over its length.

( Even with packing ports, the packing of chromatography columns is a skilled job requiring training and experience.:'acking nozzle helps to achieve consistent results, but only experience and time 5 consuming testing of trial packs c-an indicate the optimal slurry concentrations and primp pressure profiles for a given medium in a given column. Since many columns are of steel the operator can have very little idea of what is happening inside.

10 SUMMARY OF THE INVENTION

We have found that by transmitting periodic mechanical vibratory signals, in particular sonic or ultrasonic signals, through the bed space of a column, one can obtain useful information about the condition 15 and/or position of a bed of particulate medium in that space by detecting the transmitted signal. We have confirmed that apparatus to implement these findings is practical to make and use, and that useful results are obtained 2() One aspect of the invention is to use the detection of a sonic or ul.trasoni.c transmission through the bed space as a mcar-rs of improving the process of packing the column, by marking the passage of the advancing "front" of the accumulating bed and/or to observe conditions in 25 the body of the bed which may be of practical importance.

Other aspects relate Lo detecting the properties of

sonic/ultrasonic transmissions through the bed space as a means of noting the presence of other materials in the red, e.g. components which are elating through the column, or impurities which may be fixedly bound in the 5 column. OLher aspects are chromatography apparatus adapted for the performance of the various method aspects, comprising one or more sonic/ultrasonic transmitters/detectors. 10 Further aspects relate to the provision of chromatography packing apparatus enabling a packing procedure to be monitored and to assess its conformity with a predetermined packing profile, and for using automatic or operator- controlled feedback to approximate 15 the actual packing process to the predetermined profile.

Aspects of the invention are set out in the claims.

We have determined that a sonic/ultrasonic transmission through the bed space of a chromatography column is attenuated by the presence of the bed and its 2() speed increased, i.e. its time c-' f flight reduced, compared with when the bed is not present. Lither or both of these effects may be used to obtain data about the bed in the column. Also, either or both of these properties (amplitude, speed) can be influenced by the 25 presence of adventitous substances in the bed.

Thus, the present proposals enable any one or more

of - determining a height to which packing medium has accumulated during packing; - determining a packing density/degree of 5 compression of packed medium; - determining - by means of plural transmissions along different paths - a rate of advance of the bed front during packing; - determining the presence/position/extent of 10 contaminant materials or elating product components in the bed from time to time, e.g. as chromatography proceeds. Preferably the state of packing of materials within the chromatography column can be assessed by measuring 15 the time taken between transmission and reception of a transmission, i.e. the speed.

Alternatively or additionally, measuring the attenuation of a transmission can be particularly useful in determining the height of a slurry bed within the 20 chromatography column.

Preferably the transmission is directed through the interior of the column in a direction substantially parallel to a base of the column, e.g. in a substantially radial direction so that the transmission crosses through 25 or close to the axis of the column. The transmission may be focussed by a sonic lens.

The transmission is preferably in a pulsed form.

This aids detection.

The frequency of the transmission is selected to obtain a suitable detectable signal for the Purpose in 5 hand. Preferred frequencies are ultrasonic, e.g. 2 0.5MHz, > l.OMHz, 2 l.r'MHz. For example a transmission of 2 MHz was Lourld to give a suitable detectable signal.

The attenuation suffered by a 2.5MHz transmission passing through a slurry bed was found to be greater than that at 10 2MHz. When measuring the speed of a transmission minimum attenuation may be desired. When determining the height of a slurry bed within the chromatography column it may be desirable to use a sonic transmission which is more 15 significantly attenuated by passing through the slurry bed. A sonic/ultrasonic transmitter is used to generate the transmissior-r. The transmitter may be attached to a side wall of the chromatography column, preferably on the 20 outside. The transmitter may be a transceiver which is capable of detecting as well as transmitting sonic transmissions. Pretcrably the sonic transceiver comprises a piezoelectric material. Preferably the sonic 25 transceiver is used to transmit a sonic transmission and to receive an echo of this transmission. This technology

( is known per se e.g. for measuring fluid velocities in pipes. The time delay of the echo together with the dimensions of the column can be analyzed to find the 5 speed of the sonic transmission and thus obtai information about the packing state of the materials therein. The thickness and material of the chromatography column walls can be taken into account as necessary. 10 Alternatively sonic transmission may be generated by a transmitter and received separately by a receiver. 'I'he transmitter and receiver are then preferably on opposite sides of the column.

In the following proposals it is generally possible 15 to use either a sonic transceiver or separate transmitter and receiver.

Preferably the transmitter is attached to the exterior of the side wall of the column. A lens may b placed between the transmitter and the column surface.

2() Alternatively a transmitter and/or sonic lens may be embedded within the side wall of the chromatoclraphy column or even placed within the column.

The transmitter may be placed at any appropriate height. To detect when the slurry bed has reached a 25 given height in the column then one may place a transmitter at this height and to direct the transmissio

( substantially parallel to the base of the column.

Preferably more than one transmission path is provided, at different parts of the interior of the chromatography column. This enables e.g. spatial profile 5 of the packing state of materials within the column, useful in checking the uniformity of packing.

For example transmission paths may be at different axial heights.

Each transmission path may have a respective 10 transmitter/detector pair, or a single sonic transmitter/detector may change its position e.g. by sliding to different positions on the exterior of the column=s side wall.

A further aspect of the present invention is a IS method of packing a chromatography column involving pumping a slurry of particles into the column, using a sonic transmission to monitor the height of said slurry within the column and stopping the pumping when th desired height of said slurry within the column is 2() monitored. Preferably the stopping of the pumping is automated.

Preferably a switching device takes an input based on the amplitude of the monitored scenic transmission and switches off tne pump when said input indicates that the 25 desired height of slurry within the column has been reached.

( A further aspect is a method of packing a chromatography column involving using a pump to pump a slurry of particles into the column, using a sonic trar-smission to monitor the state of packing of said 5 slurry within the column, comparing the state of packing monitored to a desired state of packing and when the monitored state of packing is different from said desired state of packing adjusting the packing parameters so as to achieve the desired state of packing.

10 This aspect of the invention may be used in combination with the aspects above.

The packing parameters are any adjustable parameters which affect the packing process, for example the pumping pressure (which may be adjusted by altering the pumping 15 speed). It may also be possible to adjust the concentration of slurry being pumped into the chromatography column.

Preferably the packing parameters are adjusted by means of feedback to the packing apparatus (e.g. the 20 pump), said feedback being based upon the desired state of packing and the state of packing monitored by the sonic transmission.

Preferably the feedback is automated by the use of a computer or appropriate electronic circuits, but it may 25 be possible for the method to be implemented manually.

The feedback may be automated by use of a I to P

( converter which converts an electric current to a desired pumping pressure. The feedback may be proportional to the difference between the desired state of packing and the state of packing monitored.

Preferably the comparison between the state of packing monitored and the desired state of packing is achieved by comparing the measured speed of the sonic transmission with a desired value for this speed.

Preferably this comparison is carried out 10 continuously during the packing process and continuous feedback to the packing apparatus is generated on the basis of this comparison.

The feedback may enable an increase or decrease in the pumping pressure. For example if the speed of the 15 sonic trar-rsmissiorl is lower than the desired value then the density of the slurry is too low and the pumping pressure will need to be increased so as to increase the packing compression and density of the slurry. Equally if the speed of the sonic transmission is higher than the 2( desired valor then the packing compression is too high and the pumping pressure will r-rQed to be decreased.

Preferably information relating to the desired state of packing is provided in the form of a packing profile which details the transmission properties at various 25 stages during the packing process. These stages may be various points in time since the packing process started.

( They may be defined by the duration of time for which the pump has been active, a height in the column which the slurry has reached or a volume of slurry whi ch has been pumped. 5 Tie packing profile may for example give a profile of the desired speed of the Lransmission against a time from the start of the packing process.

IL is envisaged that the profile may be derived from measurements carried out during a successful I- dial 1() packing of a chromatography col umn.

The profile may give a range of acceptable values, rather than an exact value, for the property e. g. speed of the transmission at each stage during the packing prone s s.

15 The profiles may differ for different combinations of chromatography col umn type and slurry type. Therefore it is envisaged that a different packing profile may be provided for each such combination.

The packing profile is also expected to vary 20 according to Lie path of t he transmission in the chromatography column.

Preferably the packing parameters are adjusted on the basis of the packing st at e monitored by respective ones of plural transmissions and the desired packing 25 state for each respective transmission.

further aspect of the present invention is a

( system for packing a chromatography column, the system comprising a pump for pumping particulate slurry into the column, signal generating means for generating a signal to control a sonic/ultrasonic transmitter, receiving 5 means for receiving a signal from a corresponding receiver, analyzing means for anal.ysing a signal received by the receiving means and outputting a result based on analysis of said received signal and pump controlling means for controlling said pump on the basis of the I() output from said analysirig means.

Preferably the system has data means for storing or for reading packing profile data. Preferably said packing profile data contains chromatography column dimension data. 15 Preferably the analysing means is capable of detecting a signal received by the receiving means which corresponds to a signal previously generated by the signal generating means, calculating the time delay between the generation and detection of said signal, 2() analyzing said time delay along with chromatography column dimer-siordata and thus computing the speed of a sonic transmissior- corresponding to said generated and detected signal and generating an output based on said computed speed of said sonic transmission.

25 Alternatively or additionally the analysing means is capable of detecting a signal received by the receiving

( means which corresponds to a signal previously generated by the signal generating means and generating an output based on the amplitude of said received signal.

Preferably the packing profile data also contains 5 data relating to the type of slurry with which the chromatography column is to be packed. Preferably the packing profile data contains data relating to the desired output from the analysing means during the packing process. Preferably the packing profile data 10 contains data relating to the desired speed of said sonic transmission corresponding to said generated and detected signal. Preferably the pump control means is capable of controlling the pump on the on the basis of said packing 15 profile data and said outputted result from said analysing means.

Preferably the pump control means is configured to increase the pumping pressure if the result from the analysing means indicates that the velocity of the sonic 2() transmission is lower than desired and to decrease the pumping preC,c;ure if the result from the analysing means indicates that the speed of the sonic transmission higher than desired.

It is envisaged that the packing profile data 25 relating to said desired output will detail the desired temporal profile of the output from the analyzing means

( during the packing process.

It is envisaged that it will be possible to provide tile packing profile data on a portable data device such as a smart card or a floppy disk.

s Preferably the system for packing the chromatography column is capable of taking inputs from a plurality of receivers/detectors. It is envisaged that each may be located at a different height on the chromatography column. 10 Preferably the system has a plurality of analysinc; means, each respective analysing means for analysing a signal from a respective receiving means and generating a result based upon said analysis. Alternatively the system may have an analysing means capable of analysing a 15 plurality of signals received by a plurality of receiving means and capable of generating an output or a plurality of outputs based on analysis of said received signals.

Preferably the pump control means is capable of receiving a plurality of outputs from one or more analysing means 2() and controlling said pump on the basis of said outputs.

A further aspect of the present invention is a chromatography column with an attached sonic transmitter or transceiver for transmitting a sonic transmission into the interior of the chromatography column.

25 A further aspect in the present application has to do with packing a chromatography column. In our work we

l, have found that although a sonic transmission can be affected by the packing density and pack quality of a bed, these effects are more difficult to detect reliably than the simple large effect on the transmission due to 5 the presence, as opposed to the abserlce, of the bed (settled or packed media). The latter is a strong change and easily detected, so that the position of the bed "front" can be determined with confidence using such sonic transmissions, even when (as is strongly preferred) 10 the sonic transmitters/receivers are outside the column interior and must act via the column wall. At the same time we have made a new and useful finding that the rate of advance of the bed front during packing is a significant parameter correlating with the quality -- i IS particular the quality in terms of plate value i.e. chromatography efficiency - of the resulting pack. This is of practical importance because of the difficulty of standardizing pack control parameters. For practical reasons the conventional primary control parameter is the 2() packing pressure applied by the pump at the packing port, and packing operatives are accustomed to adjusting this packing pressure during the procedure to achieve desired results. However the absolute values and profiles of the packing pressure for packing a given medium into a given 25 column cannot usefully be prescribed. This is because different colurnr-r setups, even with essentially identical

( columns and media, generate significantly different back pressures associated with variations in slurry corlcer-rtration, bed support type, buffer viscosity, temperature, column expansion and the flow systems 5 downstream of the bed, e.g. length, diameter and the number and acuity of bends in pipe work. In practice, several trial packs followed by plate value assessments are needed before an optimum packing pressure profile can be settled on for a given column set-up.

10 A constant packing pressure is not a useful control.

In general the flow will be found too high in the early stages. Conversely if fluid flow is held constant the pressure at the end of the procedure is too high for good results. These subtleties are peculiar to closed-column, 15 injected slurry techniques; they do not arise in the conventional oper-column pack where essentially all the medium needed for the pack is present in or above the column space at the outset.

Thus we propose a method of packing a chromatography 2() column in which the rate of advance for the bed front is measured and ore or more packing parameters -typically pump speed ancl/or slurry concentration - controlled in dependence on the measurement to approximate the ongoing rate of advance to a target value, or to keep it in or 25 bring it into a target range.

Ire line with the aspects above, the preferreci method j! l

of measuring the rate of advance of the bed front is by detecting successive positions of the bed front using sonic/ultrasonic transmissions through the column interior. Typically the preferred (target) rate of bed front advance will vary during the pack. The method may involve measuring that rate at plural positions distributed axially (i.e. in the direction of accumulation of the bed) along the column. Control 1() feedback can be arranged by means of conventional processing technology, feeding the output from the relevant sensors, e.g. ultrasound receivers to a control processor for calculation of the real-time rate of advance, comparison with a target value representative of 15 a desired rate of advance or "profile" (variation with time, or with axial locatior- ) of the rate of advance, and control signals sent to a pump to determine or vary the rate of pumping accordingly.

Because this method takes direct account of the 2() actual accumulation of the boa, it can avoid some of the triad and error preparation which (for reasons explained previously) is associated with control via monitoring the packing pressure.

While the optimum advance rates and advance rate 25 variation patterns can be determined previously for given columns, liquid and media, it can be said in general that

( a preferred "rate of advance" profile will usually have a first phase, corresponding to an initial build-up of medium or the bed support (mesh or winter), which is slow by comparison with a subsequent main phase which is l 5 faster. There may be a gradual increase between the two.

This appears ire general to lead to better packing results. The difference if any between the rates of advance during the main phase of packing and at the final phase of packing (where the bed approaches the top 10 permeable retainer e.g. mesh or sinker) appears to be I less critical. It may be of importance with some media in which case the target profile can be determined accordingly. People packing similar media into similar columns subsequently can then get the benefit of that 15 initial empirical investigation on an automated basis.

skilled person will appreciate that for the present purposes it may be preferable to have an essentially progressive assessment of the rate of advance of the media front up most or all of the axial extent of 20 the column interior. '1' o this end a series or array of sonic trar-smission and detection elements e.g piezoelectric elements may be installed on the chromatography column, preferably on the outside of its the wall so as not to affect the uniformity of the 25 interior. The direction of transmission of the signals from transmitter to receiver is preferably substantially

( transverse to the direction of advance of the bed front, since this maximises the difference in effect of the transmission in front of and behind the front. However, other dispositions of sensors and transmitters may he 5 acceptable. For example, an emitter or receiver may be positioned at the end of the column opposite to the end where the bed initially accumulates. It may transmit to or receive from sensors or transmitters disposed on the side of the column. Or, such an arrangement may be 10 combined with a transverse (radial) system; the two systems may reference one another for reliability.

And/or, the rate of advance of the bed front may be determined by directing a transmission axially or with an axial component, onto the bed front and detecting the 15 back-reflection from the front. In the latter respect, we note that the use of ultrasonic transceivers to ctetermine the levels of materials in industrial vessels is established practice. Indeed, it has been used in the specialized context of a fluidised bed chromatography 2() column as a means of measuring the height of the particle bed in use. However, these transceivers project laterally inside the column which is acceptable in liquid-containing vessels, and in fluidised bed chromatography processes which are exceptional in that 25 the medium does not fill the column in use, but is not good practice in packed chromatography bed procedures.

( Also, the prior art uses of ultrasound transceivers in

vessels have not been used to control the rate of advance of a bed front on a feedback basis. There has not previously been any perceived reason for doing so.

5 The sonic/ultrasonic transmission apparatus may have separate emitters and receivers, or transceivers which combine the two functions. These technologies are in themselves well known, as are arrangements for positioning sonic emitters and receivers effectively on 10 the outside of the vessels so that they will work through the wall. The latter technology is well established in ultrasonic meters which measure fluid flow rates in pipes using Doppler-type effects.

Apparatus for carrying out the method is a further 15 aspect of the invention. In particular the apparatus may comprise a chromatography column adapted with suitable sensors and control circuitry operatively connected to a packing pump to carry out a method as described. A particular apparatus may include an array of sonic 20 scnsors/trarlsmitters mounted or (or adapted to be mounted on) a column wall and connected to electronic processing means programmed to determine a rate of advance on the basis of signals from these sensors, compare the rate of advance with desired targets or ranges and, in dependence 25 on the result of the comparison, send or adjust control signals to a packing pump.

( A further aspect of the invention is based on a further new finding we have made, which is that the presence of adventitious substances in a packed bed can affect the speed and/or attenuation of sonic s transmissions throug}l the bed. It is therefore possible to use such sonic transmission to detect the presence and/or the movement in the bed of such materials. In one aspect such materials might he contained in bands gradually elating through the bed during a 10 chromatographic process. Using one or more sensor arrangements to detect the passage of such a band at one or more corresponding regions of the column enables a "tracking" of the process which can be helpful to the operator in monitoring the procedure and collecting the 15 separated substances as they emerge. With sufficient sensors, e.q. a series or array as discussed above, the movement of a band of substance through the column can be tracked, and if desired visualized on a display outside the column.

20 Another use of this finding is as follows. In some processes the materials presented to the column for separation inclucie materials which will bind irreversibly to the chromatographic medium in the column. (generally these bind to the medium immediately or soon after 25 entering the bed. Their permanent binding reduces the transitory binding capacity of the medium which is the

foundation of the chromatographic process for the other components of the mixture. A region in which the bed is progressively less and less effective grows gradually adjacent that end of the column at which the starting 5 material is introduced. In practice there comes a point at which the affected band at the end of the bed is so large that the 'red as a whole is inadequate. This usually becomes apparent when sooner or latera product batch proves to be impure. The processing of that impure 10 batch is a substantial waste of time and materials. The present invention therefore provides a method in which, by means of sonic transmissions through the relevant part of the bed, it is determined from time to time whether such permanently-bound adventitious substances shave 15 reached a predetermined threshold position in the column corresponding to an operational limit at which the column needs to be emptied and repacked. This procedure promotes confidence and consistency which are of high importance with these techr-rologies.

20 A final aspect disclosed herein relates agair1 to packing rather than rurrr-ing the column. During the packing process, sonic transmissions according to any of the above proposals are used to identify a time at which the advancing bed front has nearly reached the top of the 25 bed space. In dependence on that detection, the control system switches the pump control to act in dependence on

l a detected packing pressure. When a dip in packing pressure characteristic of complete packing is observed, the pump is turned off. Correspondingly programmed apparatus is again an aspect protected herein. To 5 explain: it is weld known to those skilled in packing chromatography columns that a characteristic dip in packing pressure is seen just as the column becomes full.

It is undesirable to continue to apply the pump beyond this stage; a better pack is obtained if it is promptly 10 turned off. With an opaque column, the packing operator must keep a careful watch for this. A continuous pressure-sensitive control of the pump is not desirable, however, because pressure fluctuations of comparable sizes occur at other stages of the packing when the pump 15 should certainly not be turned off. By using a sonic transmission sensor to note when the bed is nearly complete and only then initiating the monitoring for a pressure dip, the virtues of these respective techniques are happily combined.

2() Tests underlying the present proposals, and examples of apparatus and procedure, are now described with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig l is a schematic view of a chromatography column 25 with an automated packing system; Fig 2 shows ultrasound speed measurements across the

( column during a trial pack; Fig 3 shows similar measurements but with a greater degree of temperature control) Figs 4,5,6 and 1 are data showing relations between 5 pack parameters and pack quality for a Sepharose (gel) TTlediUm; Figs 8,9 and 10 are packing data from 5 runs for determining a relation between packing rate and pack quality for ceramic media; 10 Fig 11 shows a UV trace at the product outlet of the column in a test run with a sample of albumin; Fig 12 is an ultrasound speed trace during the same run, showing a speed change through the albumin-occupied region of the column; 15 Fig 13 is a schematic view of a column packing system exploiting feed-hack from ultra sound sensors and having a display, and Fig 14 is a further schematic view illustrating the detection of permanently-bound impurities and of sample 20 bands passing through The column; Figs 15 to 19 are schematic views of apparatus set ups and procedures exploiting the ultrasound detection facility, and Fig 20 shows a packing 'skid' embodying the 25 invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

With reference to Fic3 1, a chromatography column 1 is provided with upper and lower packing valves vcl, vc2 of the kind described in our WO-A96/10451 to which 5 reference should be made. The packing flow system includes a larger pump pi for packing and for spraying the column fop where? unpacking, a smaller pump p2 also for packing and enabling suction from the bottom of the column when unpacking, tanks S. L to hold a slurry of 10 packing medium and particle-free liquid respectively, a four-way valve vl controlling the inlet to the pump pi, a four-way valve v2 controlling the outlet from the pump pl as between the top and bottom of the column, a three-way valve v3 controlling the inlet to the pump p2 from the 15 column or tank, a three-way valve v4 controlling the outlet from the pump p2 to the column or the tank, and a three-way v5 controlling a return from the column top to the tank or to a vent. For a description of a suitable

enclosed (injected) packing procedure, refer to e.q. WO 2() A-96/10451, GB-A-2258451 or WO-A-99/64130 as morltior-?ed previously. In the present work the column was a variable-length 600mm diameter column, operated at a bed height of 160mm.

TEST 1: RECOGNISING THE FRONT

25 In a first set of tests Sepha rose 4 medium was used, slurries in distilled water. By way of illustration, the

( following Table 1 shows the variation of various parameters as the pack proceeded. Ultrasound trar-rsceivers 2,2a (Fig 1) were secured to the outside surface of the column wall about a third of the way Up s the bed. The speed of sound through the column interior as measured between these is given in Table 1. 'I'he tabled values are purely illustrative. Figs 2 and Figs 3 show plots of the sound speed between the ultrasonic transceivers 2,2a and time in two different runs. Fig ? 10 shows an initial phase I during which the sound speed increased gradually, followed by a discontinuity at which the advancing bed front rose past the transceivers - the column used was a transparent-walled-one for reference, so this was seen to happen - and a second phase II where 15 the sound speed ttlrough the medium bed was appreciably higher. Sound speed is significantly dependent on temperature. Investigations showed that the gradual rise in phase was due more to temperature changes than slurry density phenomena. Indeed, more scrupulous 20 temperature control of all the materials resulted in the Fig:3 plot wI-rcrc the speed of sour- d through tile urrpacked region of the column remained essentially constant until the clear discontinuity when the bed front, rose past the trar-rsceivers 2, 2a. These experiments confirmed the ease 25 of identifying the bed front using ultrasour-rd.

( TEST 2: QUALITY OF PACK, AND ACCUMULATION RATE

_ A further set of tests was then carried out using a ceramic medium. The study and its results were as follows. 5 SUMMARY

Ceramic medium (Bi.o',epra Q Hyper DF) was packed into a 400mm diameter column to a 200mm bed height. ':he buffer, packing pressure and slurry concentration were varied. The two best repeated results were obtained by 10 0.()25M Tri.s HC1 adjusted to Ph adjusted 1.5 with NaOH.! The slurry concentration was 50'.; and the end packing pressure in the column was 1. 5 bar. The flow rate at the start was initially l000cm/hr.

In the light of our work, it appears that at the 15 start of the pack the bed height build should be at say from 25-30mm/minute for the first third of the total packing time. After this stage the rate should increase to for example 75-80mm/minute until the column is filled.

For a 200mm bed hei.gilt, near the completion of the pack 20 of the flow may bc pulsed between 300 and 90()c-m/hr. The pack Carl be ccnsiclereci c:omplcte when the pump:' stalls, usually after 3-4 minutes for a POOmm bed height.

It appears from the work (described below) that there its a packing rate "corridor" in Lerms of rate of 25 bed height accumulation which has an appreciable width and which if kept to can lead to a better pack. An actual packing rate/time patio following Such a corridor

( can be regarded as a fingerprint of an optimised packing process. That is to say, if the packing operative keeps the packing parameters so that the rate of bed accumulation is withir- the upper and lower bounds of the prescribed "corridor" then a good pack can be predicted.

This "fir-gerprint" or pack profile Carl then be used in production as a practical aid via automation and as a validation parameter for a column.

DESCRIPTION OF WORK

_ 10 The column used was a variable bed height Mark II Euroflow column (EQ400-V-EQ911, with steel meshes). The medium has already been identified...DTD: Before packing the medium was de-fined three times in slurries at about 30 -O concentration. The slightly lS cloudy supernatant was pumped away. The columr- was fully primed with buffer before all packs. 'slurry was pumped in via the top nozzle (it can be done through the bottom) and the slurry liquid left via the mobile phase path at the bottom of the column. The packing pressure was 20 measured at the top mobile phase port. Tree slurry was pumped by a Husky 715-diaphragm air-driven pump.

Red quality evaluation was by injecting one lithe of a lo acetone solution in water under the column using up flow and down flow at a variety of flow rates, before and 2.s after an 18 hour test period for bed stability. The same was done with mobile phase in 0.025 M Tris HCl adjusted

to pH 7.5 with NaOH.

Table 2 details the 13-staye test conditions used.

Table 3 displays the 100cm/hr flow rate (21/minute) results based on the lowest HETP, being those obtained 5 nearer the 100cm/hr results. Results are shown for before and after the stability test delay Fig 4 of the drawing shows the packing profile for each of the five packs indicating the quality (in terms of plates) of the pack achieved. The graph shows the 10 rate of build-up of the bed during packing; this rate of build up is of course a complex result of flow and pressure conditions in the column during packing at the relevant stage.

The results indicate that the ceramic medium could IS be packed in the 400mm diameter column to a 200mm bed heighL to a quality of over 4000 plates per metre with asymmetries of about 1.4. Two repeats gave similar results. Fig 5 adds to Fig 4 by including linear regressive 2() approximations to tile respective curves, and their radiance. Fig 6 shows that there is a correlation between the gradient (i.e. the rate of bed build-up in mm/minutes) and the plate number (quality) of the resulting bed.

25 Of course these straight-line regressions are crude approximations. 'I'able 4 below gives a better analysis by

( dividing the packing procedure into 3 stages.

Observations from Table 4 are listed below.

The 'good' packs (Pack 2 and 3) share very similar rates at start middle and end. The first build up 5 of the bed is the slowest part. Ther1 the rate at which the bed builds increases at Phase 2 perhaps because the supernatant slurry concentration increases. This higher rate is held constant for Phase:3. 10 2. The 'satisfactory' packs (Packs 1 and 4) deviate from the successful profiles significantly at the later part of the pack. Their initial build up rates are quicker than the 'good' packs. Their 2'0 and 3" Phases are very different yet they yield 15 similar results.

3. The 'poor' pack (Pack 5) profile deviates the most from the others. The initial build up is low or similar but the 2"' phase is 4 to 8 times slower.

Phase 3 increases but is still half that of the 20 Phase 3 rates of the good packs.

Some conclusions based on these observations are listed

below: 1. During the first phase where media initially builds 25 up on the bed support the rate needs to be slower than the later Phases. Perhaps to avoid blinding

( the mesh with high velocity media that has little or no back pressure.

2. Ii the 2'i'i Phase is equal or slower than the 1;' Phase the result is very poor (Pack 5). Perhaps the 5 slower rate allows the layer against the bed support to mix and lose its packing density or homogeneity.

Titus a quicker 2' Phase rate is needed to hold down the 1; layer.

3. Tile aim Phase can be very different to the 3rd Phase 10 and still yield satisfactory packs (Packs 1 and 4) it is not catastrophic.

4. To achieve better packs the 2n'i and 3ri Phases are preferably about three times faster than the 1St Please. They may be at about the same rate. This IS may helps homogeneity throughout the rest of the bed. Based on these results it appears that there is a packing rate 'corridor' of varying width that can be 2() foTlowecl to achieve a 'good pack'.

Figs 11,12 show the result of an experiment in which a column set-up as shown in Fig 1 had applied to it a sample pulse of 10t albumirl. Fig 12 indicates that th albumin passed as a band or pulse through the column as 25 detected by UV detector at the column exit. Fig ll shows the interesting results from the ultrasound sensor,

namely that the passage of the band past the sensor correlated with a band of increased ultrasound transmission speed through the column. This is marked B. Iigs 13,14 show schematically an apparatus set up 5 embodying the invention, with a linear array 20 of numerous piezoelectric transceivers applied up one side of the column, with a corresponding array 21 on the opposite side. They need not be exactly opposite; in particular a slight offset helps to avoid difficulties 10 adjacent the ends of the column if there is a projecting central packing nozzle. The drawings show schematically the front 5 of a bed 16 rising up the column as packing proceeds. A programmed control unit 4 - a conventional microprocessor - is fed with the inputs from the sensors IS and programmed with desired target data for the target rate profile. The packing pump P is controlled accordingly. An external display 3 is provided which may show the sound impedance or speed in bar form against the height up the column. Fig 14 shows a similar apparatus 2() being used at a different stage, when the column has been packed and is in use. Ore aspect of the use is shown at the top of the bed. band of accumulating permanently bound contamination is gradually extending down into the bed from the top. This affects the ultrasound 25 transmission from the top sensors and is therefore shown on the display at 'A'. When it reaches a critical level

( CL the program issues a warning to the user that the column is effectively spent.

A hand of material 6 is also shown, progressing down the column. This is material being purified. Despite 5 the opaque column wall, its progress can be followed (peak B) on the visual readout of the ultrasound data.

Figs 15 to 19 show schematically these and other functionalities of the proposed column arrangement having the array of ultrasound sensors extending axially.

ID Fig 15: column 1, sensor array 20, control processor 4, data logger 31, pump P. Basic 'fingerprinting' of a pack profile, plus pump control, using feedback.

Fig 16: packing, with tracking of the accumulating bed on display 3.

15 Fig 17: packing method using additional control parameter of packing pressure at meter 25. [)etect when column nearly full, open time window of sensitivity to pressure drop, pressure drop signals pack complete, pump stopped and valves moved to 'run' positions for 20 c.}lromatography. Fig 18: display 3 indicating void or inhomogeneity 62 in the column contents.

Fig 19: tracking a band 6 of valuable component through the column display 3 or band 'B'.

25 Fig 20: shows a packing station or 'skid' i.e. a movable trolley having a packing pump P and the

! associated valve connectors V, operatively controlled by processor 4 adapted to receive inputs from ultrasound detectors, initiate operation of transmitters, receive packing profile data and,orograrred to control the pump P accordingly.

Table I

SI,IJRRY 18EC GOLIJMN 19EC AMBIENT 19EC

LIQUID

_ _ _

PAC'KING l'ROFILE Packing lYir Slurry ' Slurry Pressure through through Baseline, valve only valve only bar I itres/min bar PACKIN(1 Fl,()W PRIES SURE [3F.I) MOTIVE SLURRY Speed TIME I IEIGI IT AIR INLET of PRESSURE PRESSURE Sound minutes I/min bar mm _ bar = bar =.._ 0 36 0. 25 0 3 1478.6

_. _ 1 36 () 15 15 3 1479.3

_. 2 27 0.2 45 2.5 1479.9

_ 3 25 0.4 85 2.5 1480.2

4 17.2 0.55 105 2.5 1480.3

_. _ 5 14 8 ().65 130 2.5 1479 8

_ _ _

6 14 ().65 155 _ 2.5 _ 1490.6

7, _ 14 0.65 160_ 2.5 1490.X

Tahle 2 Sl IOWINC; 'I'IJE 'I'EST (:ONL)ITIONS USED Volumetric flow rate l/min Lincar flow rate cm/hr Flow direction __.._ _...

3 150 _ Down . I ()O Llown _. _... ... _ __.

I 50 L)own 1 5 () _1 JP

2 _. 100 _.. _ _1 Jp 3 150 1 Jp _ _. _. ..

318 hour stability run 150 L)own . _... _.. _..

150 L)own _.., _

2 100 L)own . _ _

I 50 Down _ fJp _ _. _

2 _.__ 100 lJp 3 _ 150 tJp _._ _

Tablc 3 Pack Results Bel'ore Bekre After After Mean Packing, Number Stability: Stahility: Stability: Stability: Dil'f'erences L) ownflow Upflow Downflow Upflow to B2 and B3 I'l/m 3384 n/a 3100 n/a 3242 Packed in Asymmetry 1.5 n/a 1.25 n/a I. 375 P04 bul'! cr at 1.3 her _ _

B2 1'1/m 4881 3842 3857 377() 4088 _._ _ Asymmctry 1.4 I.4 I 4 I 3 1.4 B3 _, 4308 4328 4267 4147 4263 _ _

Asymmctry 1.4 1.4 I.5 1.6 I.5 _ 1

B4 1'1/m_ 3485 3449 3142 _ _ 3112 3297 5 minutes I Asymmetry 1.4 1.4 1. 6 1.8 1.6 to pack 1.4 bar. Before stability downflow at 50 cm/hr B5 131/m 2443 2161 2597 2398 _2400 Took 7 Asymmetry 1. I 1.8 1.5 I.4 I.5 minutes (stopped & started) to pack at 2.2 _ _ bar I'ablc 4: C.'omparill the Pack Profiles as C)bservations from (Irapl1 Pack NumUcr I'lates/metre Initial'l'hird of I Middle Third of linal 'I'hird of (2 sig l lg.) Pack Phase I I Pack Phase 2 Pack Phase 3 .._ P acking Rate mm/m n 1 3200 33 40 60

_ _ 2 4100 26 82 65

3 4200 27 77 77

_ _. _

4 _ 3300 41 5() 26

5. 2400 24 = 39

Claims (10)

( CLAIMS:

1. Chromatography process carried out in a chromatography column having a column housing defining a 5 bed space containing a packed bed of particulate chromatography medium, the process comprising passing a process liquid containing components to be separated through the packed bed to separate the components chromatographically, 10 characterized by detecting the presence and/or position of a said component in the packed bed by means of transmitting one or more ultrasound signals through the bed space and detecting the transmitted signal, the speed and/or attenuation of the signal being affected by 15 the presence of said component in the bed on the path of a said transmission.

2. (chromatography process according to claim 1 in which the detected component elates through the bed as a band.

3. Chromatography process according to claim 2 in which ultrasound transmissions are made at plural locations distributed along the packed bed, to track the progress of the band of component as it passes through the bed.

4. Chromatography process according to claim 2 or 3 dependent thereon in which the movement of said component band is visualized or-i a display outside the column.

5. Chromatography process according to any one of the preceding claims in which plural ultrasound transmitters and/or plural ultrasound detectors therefor are r) distributed along tile column to enable ultrasound transmissions along a plurality of paths distributed along the packed bed.

6. Chromatography process according to any one of the 10 preceding claims in which the column housing wall is opaque.

7. Chromatography process according to any one of the preceding claims in which the speed of transmission of the ultrasound signal is monitored.

8. (chromatography process according to any one of the preceding claims in which the attenuation of the ultrasound transmission is monitored.

9. (chromatography process according to any one of the preceding claims in which said ultrasound transmissions through the packed bed adjacent an inlet for the process liquid are used to determine the extent of encroachment 25 of bound impurity into the bed from that end.

10. Chromatography process according to any one of the preceding claims in which the column housing is a

( vertical cylinder and ultrasound transmissions are made radially across the cylinder.