SSL (?): an | aus

Bitte die Forenregeln und die Hinweise zu den Versuchen lesen!

Neue Antwort erstellen
My HPLC Diary [English Version]

Anmeldedatum: 02.01.2016
Beiträge: 557
Artikel: 3
Geschlecht: Geschlecht:Männlich
Alter: 20
Ich bin endlich mal dazu gekommen, mein HPLC-Tagebuch fertig zu übersetzen ... Smile

A while ago I told someone in the forum that I’ll write a “HPLC diary”, which describes the way of my second-hand HPLC. Well, here it is. It still doesn’t work now, in march 2017, but I’ll finish this article when it’s so far. As this was the basic idea, I’ll write the article like a diary. This also shows pretty well, how long this whole process needed and I think it will also be better structured. I’ll try to reproduce the dates as good as possible.
This article should be a lecture for interested people as well as a help for anyone who might have similar problems with an old HPLC system. The amount of information about this is, also in English language, very small. I found through my internet research not much more than a few questions in researchgate about specific problems, e.g. the connection of an old detector to a computer via GPIB. That’s the reason why I translated this article in English.
This article describes my specific work at a specific device. For a general introduction into HPLC I refer you to the article of R’-MgX, you can find it here.
Now, here’s a short introduction for those who didn’t get the story here in the forum. In June 2016 I got an old Merck Hitachi HPLC system from my school. It stood there for many years unused in the corner and should then be replaced by a new device. So I asked if I could take it home and, after consultation with the school administration, I was allowed to. As you may imagine, I was very happy about that.

24th June 2016
After a first inventory, I knew that the device was buildt by Hitachi (“Merck/Hitachi”) in 1986/87. It consists of a pump with a ternary low-pressure gradient mixture chamber and an injection valve with a 20 μL sample loop, a “LC-Controller”, an UV Monitor and a PDA detector. The exact names are: L-5000 LC Controller, 665A-11 Liquid Chromatograph (pump), 665A Variable Wavelength UV Monitor and L-3000 Photo Diode Array Detector.
Unfortunately, most of the cables had been disposed, only the buildt in cable for control of the proportioning valve and a GPIB cable to conect the controller to the pump were left. There also was buildt in a LiChrosorb® RP-18 (5 µm) Hibar® RT 250-4 column from Merck.

The HPLC system directly after its arrival

Because of the device should have been disposed, all the eluent and capillary tubing were cut through so I had to replace them. And I had to buy power cords.

27th June 2016
After the power cords had been bought and the devices were connected to the power supply (you need many sockets for that!), I tested the basic functionality of them. As things turned out, I could turn them on without any problems. Also the proportioning valve was acoustically tested and seemed to work. And I could control the pump with the LC-Controller (gradient, flow rate, start/stop and purge). A great first result!
I measured the outer diameter of the eluent and capillary tubes with a caliper and ordered each 3 meters eluent tube (3.2 mm OD, 1.6 mm ID) and capillary tube (1.6 mm OD, 0.26 mm ID), both in PTFE (as I found out visually afterwards, the old tubing wasn’t made from PTFE but another fluoroplastic, maybe FEP).
On closer inspection of the tubes I already thought that the connections will be a bigger problem. That was because the seal of them was a flattened tube ending (capillary tubes) or an everted tube ending (eluent tube). For the capillary tubes, these tube endings are mounted directly on the glass corpus of the flow cell in the UV and PDA detector.

Seal of the eluent and capillary tubing

14th July 2016
The new tubing arrived! But until now all attempts to get a tool for preparing the tube ends were unsuccessful. I’ve asked my teacher, here in the forum and also wrote to the closest university but no one even knew this type of sealing.
So I tried a tip from a friend who has been working with this type of seals in the 70s. He told me that the tubes could be widened with a hot cone (e.g. a nail). But that was a lot more difficult than I’ve thought before. The nails I had at home were either too small or a bit square. I ground one of the larger nails to a round form and tried it with that but every time the tube ripped on one side. I didn’t get it far enough widened so that the sealing would have been even roughly tight, I just burned my finger two times. That was really a disaster but at least I didn’t loose much material because I used the old capillary tubes for these tests.
Another try was to cut off the flattened end from an old tube and stick it to the new tube with a wider tube sticked over the connection. But that was as leaky as everything else. So there had to be new ideas …
The solution would have been plug-on flanges – but I couldn’t order them as a private person.

The tubes

21st July 2016
The idea came and I ordered again … this time adaptors (for a picture have a look here) to connect the eluent tubes to the proportioning valve and PTFE ferrules which I wanted to use as sealing for the capillary tubes, with the flat end to the glass corpus.
Allright, the parts came … and it turned out that the adapters didn’t fit into the valves – but also not in the tube, even if this was specified by the manufacturer. So there luckily was no problem to exchange it and I ordered mini tube connectors (made of PCTFE) instead. With these I could stick the new tubes to the remaining ends of the old ones at the proportioning valve. The first problem was solved!

The connection with the tube adaptor

The capillary tubes made more problems …
First of all I tightened the screw to much what left the ferrule as an (of course leaky) clump in the thread. I had to remove it very carefully with a bent needle for not scratching the glass beneath. At least it at ha hole in the middle…
Also at further tries with the ferrules the connection leaked every time. But I hold the solution in my hands already, though I didn’t know it at that time. As I found out, the ferrules didn’t leak at the front end but the back end, where the tube comes out. Differences in the strength of tightening didn’t bring the effort I hoped for, but at some time I got the Idea: I sealed the leak with a pretty small piece of Teflon tape (that usually usedfor water installations), which I wrapped around the tube. The tape was fixed by the screw when I screwed it into the measuring cell – and the connection was tight!

Now I just hope that there won’t be any sample carryover … ‘cause then I’ll have all the same once again.

05th August 2016
Today the HPLC solvents arrived: 2.5 L 2-propanol and 1 L acetonitrile. The first flush of the pump (of course without the column, in purge mode) had been done with “cheap” 2-propanol > 99.5 % and all tests where the column and the detectors where involved with MeOH LC-MS Grade of which I once got 250 mL very cheap.

22nd August 2016
After nothing happened for a while I can tell you some news now! I’ve contacted Hitachi, I think around two months ago, if they could send me the manual of the device parts. Theay answered that they forwarded the request to VWR because they are responsible for the support in Germany but since that I didn’t hear anything from them. So I called the number that the employee of Hitachi has sendt me and asked for it. It seems like the request was lost somewhere. But now I got the manuals of the LC-Controller and he PDA detector via E-Mail (If anyone is interested in the manuals, you can contact me via PM! I also have the manuals of some other parts now!)
For a better overview of the interfaces I made this nice drawing Wink

I’ve connected the devices already before to a “start signal chain” so when I started the controller program the detector started measuring as well. Now with the manual I could also connect the pressure sensor of the pump to the controller. I connected it firs in the wrong way but that was clearly visible because of the negative pressure shown on the display Wink.

I think now might also be the right time to look at the components in detail …

The LC-Controller:

The Pump:

The UV Monitor:

The PDA Detector:

Side-view of the pump with the column:

As you can see, the connection of the devices is made with these black screw-terminal strips (only the connection of the pump <-> LC-Controller and the proportioning valve are via GPIB). At least in my case … there is also the possibility to connect all parts (not the UV monitor!) via a so-called “PAN” interface, effectively nothing else than a BNC cable. With this interface it was also possible to connect the device to a computer and Hitachi sold a simple program for data aquision and controlling the device but in fact today both, the PC adapter and the software, aren’t available any more. And so we are already at the connection of the detector…

That’s a schematic drawing of the connection from the manual

The system was working and so I startet to think seriously about the connection of the detector … and I realized that the biggest problem was far away from being solved!
The PDA detector can send an analog signal (1 or 2 wavelengths), the mentioned PAN signal or a digital signal via RS-232 interface. For the last you can choose the time and wavelength interval that should be sendt. But if the chosen intervals are too small there are some values skipped on the devices screen (you can see this by the time) so I don’t know if there maybe is the processor performance of the detector as a limiting factor. I have to take a look on this when everything works.

11th September 2016
A little tip for people without experience in HPLC. If you have problems with your system pressure have a look on your solvent.
I recently found out that my 2-propanol creates with a flow rate of 0.2 mL/min already more than 90 bar in my system. Common flow rates in analytical HPLC are 0.5-2 ml/min, in preparative HPLC in general much higher.

To quote myself from somewhere alse in the forum (translated):
I looked for the viscosity of common HPLC solvents:
2-propanol: 2,43 mPa•s (dynamic, at 20 °C)
Water: 1,002 mPa•s (20 °C)
ACN: 0,369 mPa•s
MeOH: 0,593 mPa•s (20 °C)

Sources (only german):

Edit: And is there a formula to calculate the pressure from the viscosity of the eluent?

I’ve found something (quote translated, too):

The backpressure of a HPLC column is directly proportional to the linear flow rate u, the viscosity of the eluent η, the column length L and the specific column resistance φ.

That means that with acetonitrile I would only have a pressure of around 14 bar, is that correct?
With water it would be 38 bar and with methanol 22 bar.

So as you can see the system pressure of the HPLC depends highly on the composition of the eluent!

13th September 2016
I’ve turned to another problem once again and buildt some caps for the eluent bottles …

Homemade ventilation filters and an exhaust air filter for the waste bottle

Eluent bottle with preparated cap

Here you can see the waste bottle with a later bought cap

15th September 2016
Another crafts work has made from my rheodyne valve (injection valve) without “I” one almost with “i” Wink
Oh, yes, I’ll explain that of course … alright, the “i” in the name of the rheodyne valves means that they have an electrical contact inside that works as a switch when the small handle of the valve is turned for injection (more about the functioning here). As you turn the handle there is made a contact and electricity can flow through. The valve is connected to the HPLC system and so the device recognizes that the injection is done as soon as electricity flows. That isn’t only useful but also an important factor for the reproducibility of retention times and beause the valves with “i” are even on ebay at around 400 € I decided to build such a switch by myself – for almost nothing.

The injection valve with switch in load and injection position:

As you can see the contact is simply made between a fix strap and a small “bump” which moves with the turnable part of the valve at injection. The strap is made from a thin nickel plate which is soldered to another angled nickel piece that is then glued to the metal shield around the valve. The connection is isolated with a piece of PE foil between the contact and the shield. The “bump” is also made from nickel but here it’s “feets” were isolated by dipping into a PVC solution in THF before glued to the valve. The cables were simply soldered to the contacts, they are simple cables for model making. The other ends where screwed to the contacts on the back of the detector.
By the way, soldering the nickel plates wasn’t as easy as soldering copper wire but could be well done with normal solder.
But there are some bad news, too … it looks bad for the PAN connection. The VWR employee told me today that the software and parts aren’t available any more – and even not the knowledge about how it works (I know, I’ve told you that bevore already Very Happy ). So there is the other possibility left, connecting the PDA detector via analog signal path.
There is a free software called “CHROMuLAN” (for details see here) which is provided by the Czech company PIKRON. This company also sells, beneath a HPLC device and an amino acid analyser, an A/D converter (“ULAD32”) to connect analog devices to the software. The A/D converter costs, referring to an invoice made for me, 260 € plus 21 % tax and 50 € shipping costs, all in all more than 350 €, for only one wavelength. Really great …
There is also a plugin to connect a RS-232 device to the software but this works only with a specific signal converter which I can’t find on the internet. The man from PIKRON couldn’t help me with that because the plugin was written by the company of the signal converter and it seems like this was already a while ago …

27th September 2016
There is at least some good news!
I have to thank Mr. Pisa from PIKRON. It’s good to see that there are some people out there who aren’t only out to make a profit! He had a closer look on my problem and told me better not to buy the ULAD32. He adviced me to connect the PDA detector via RS232 to the computer instead and let someone develop a plugin for the open source chromatography software “OpenChrom”. He put me in contact with a student who would do that for 400-500 €.
That is, in fact, not much more expensive than the connection of only one wavelength via the A/D converter and I thereby get the full range of the DAD. Sounds like a good solution Wink!
There are also some more good news! I went to school today for measuring two samples at the GC and there I accidentially found a bunch of accessories for the HPLC. Two more manuals (for the UV monitor, those I didn’t have before, as well as for the Integrator that I don’t own) and, most notably, a bunch of spare parts!

I found the following parts: A tool to replace the seals in the HPLC pump, 4 replacement seals for the pump head, 1 fitting (metal), 1 ferrule (metal), some fuses (1/2 A (2x), 1 A (4x), 2 A (4x), 3 A (2x)), two green cables, a grey flat cable, a small washer, two small pieces of metal with a hole, a amber glass vial with an unknown liquid inside, accessories for the Rheodyne 7125 injection valve, (a metal capillary spiral, I think it’s the “coil mixer” as well as a multiply bent steel capillary with two fittings with ferrules attached), a cable with I think some sensor at it’s end, a fluoropolymer capillary with a plastic ferrule and some round plastic disc with 4 holes.

The found accessories

09th October 2016
The programming of the plugin begins ... I got a small Java program to test the connection between the computer and the detector – unfortunately it didn’t work ...

17th October 2016
A little digression to electronics buildt in 1987 ... Very Happy

The signal output parameters of the RS-232 interface could be changed with a tiny DIP switch. The switch is located on a removable board in the back part of the detector. With this switch it is also possible to switch between PAN and RS-232 output. The switching between analog and digital output is done in the menu of the detector.

So I opened the detector ...

Who can see the switch? Mr. Green

Maybe the drawing in the manual helps a little …

I switched it with a pointed pencil in the end Wink
And I didn’t open the detector only one time …

18th October 2016
The test program works but there still is no signal from the detector …
So we decided to test the connection with a “serial data logger”, a program that reads out the data communication of the serial interface of the PC. But there still is nothing.

20th October 2016
It seems like the first mistake is found … and the second one, too. I connected the detector with a 25 pin to 9 pin adapter. But as I just learned, those adapters weren’t standardized ... so we have no idea what’s inside. In fact, it doesn’t do what it should do.
So I used a 25 pin cable I’ve bought on Amazon to connect the detector to the 25 pin interface of the computer (luckily I have an old PC that has all those interfaces built in Wink ). But it still didn’t work ... but that’s hardly surprising, ‘cause an internet research told me that this interface isn’t a serial but a parallel interface.
I don’t know what’s the exact difference between those interface types, but in the end it doesn’t work.

There hast to be found another solution.

27th October 2016
And this solution ist to build my own adapter. If I would have had this Idea earlier I wouldn’t have bought the two unneeded cables ...

Jan, the guy who programms the plugin, sendt me the following information about the connection of the pins:
DTE(PC) - DCE(Detector)
2 - 2
3 - 3
8 - 6
7 - 5
5 - 7

or, If you don’t need the connection control data (CTS and RTS signal), the following option works as well:

DTE(PC) - DCE(Detector)
2 - 2
3 - 3

and “short-circuit” the two contacts at the detector side:

DCE(Detector) - DCE(Detector)

So I bought a cable and two RS-232 connectors from the radio and tv technician and soldered the adapter. With some basic soldering knowledge that is no problem.

Et Voilà, that‘s the result of my soldering work:

And suddenly it works! I think I’ve never before been so happy about a display full of numbers Very Happy

10th February 2017
I had to wait for a while but finally the beta version of the plugin is ready to test out!
But unfortunately it doesn’t work here …

18th February 2017
The old XP computer is broken ... okay, which modern PC will work for 15 years without problems? It was used for many years by my parents, then by my sister and now by me in my lab. Those old electronic was much more durable than today ...
But now it doesn’t boot any more, only the fans are still working. Maybe I broke it when I worked on it today.
Now I have to use the laptop first, I arranged with my family that I can use it for 4 weeks in the lab. Afterwards I need a new PC from somewhere.

21st February 2017
Today the adapter cable from RS-232 to USB arrived. Andi t works!
The data logger reads out the detector signal easily. So it works with the laptop for the moment. But the newest version of the plugin still doesn’t work correctly ...

18th March 2017
After a long time of testing, rework and testing again – without any real effort – we tried something new today. Jan worked directly on my pc via TeamViewer the solve the problem. I hope that this will bring us further!

10th August 2017
There are once again some news. But I'm still far away from the goal ...

The good news first: I got 18 used HPLC columns from someone here in the forum. Mainly RP-18 columns, the most of them from MN with spherical particles and endcapped, but also some from other producers. But there is also one RP-8 column encluded, also from MN, with 300 Å pore size and 7 µm particle size. That pore size is quite big! I wonder for what sort of separations this column could be used.
Of course I firstly have to have a look if the columns are still in a usable condition ...

I also got hands on a used PC, but it didn't work when I tried to start it first. So I'll replace the ELKOs, they don't look very well, and then I'll see if it works afterwards. I hope so, otherwise I have to look out for another used computer.

And, last but not least, I relaunched the plugin project. I got a new java test program from Jan, which worked (good?). I don't know what It exactly should do, in fact it read out the com port and wrote the communication data in two test files. But it only collected measuring data when I started the detector manually.
But Jan told me that he will be in Germany in september and suggested that he could finish the project here on site. He wants to finish it quickly now and to work on something like this from far away is certainly not easy, as I can imagine. So I'm confident that I can really use the HPLC relatively soon. I wish to do that for about one year now!

08th November 2017
Yes! The first Chromatogramm ist done Very Happy!
One and a half year after I got the HPLC from my school, it's finally done.

But I'll begin from the start. The weekend bevore the last Jan traveled all the way from the Czech Republic to Germany to finish the plugin here. And after two days of work it was time to try: Everything worked. I could see the baseline of my detector for the first time as a chromatogram Smile.
And we even found out that the sending conditions (wavelength range and intervall, time intervall) could be set from the computer.

And then, this morning, I got an e-mail with the compiled version of the program. And of course I tried it today, as soon as possible.
Here is a picture of the program:

On the left side is the overlook over the runs, on the right side is the monitor on the bottom and an alrady integrated peak from another run on the top.

The run you can see in the monitor was made under the following conditions:
Eluent: ACN/Water 30/70
Flowrate: 1,2 mL/Minute
Sampe volume: 20 μL
Column: MN EC 250/4 Nucleodur 100-5 C18ec
Sample: Caffeine solution 20 mg/L in water
Wavelength: 275 nm

As you can see there are many ghost peaks in the chromatogram. I have to look for the reason of that. There are some possible reasons: No degased solvents (no degaser and also no ultrasound), pulsation or a slight variation in the mixture ratio from the low pressure proportioning valve, pulsation from the pump or the used column.

A first test in terms of reproducability and linearity (done under the same conditions as you can see above:

Concentration Peak area
10 mg/L 5,5/5,6
20 mg/L 9,7/10,0
40 mg/L 13,3/15,4
80 mg/L 34,8/36,1

And in the end a nicer Chromatogramm Very Happy :

So now, the HPLC diary has come to an end. Alredy, I wanted to write - but that's not really correct. It's good that it's now finally done.

Chemistry is like cooking - just don't lick the spoon!

Sag' nicht, es funktioniert nicht, bevor du es nicht versucht hast!
Benutzer-Profile anzeigenAlle Beiträge von Cumarinderivat anzeigenPrivate Nachricht senden

Anmeldedatum: 02.01.2016
Beiträge: 557
Artikel: 3
Geschlecht: Geschlecht:Männlich
Alter: 20
Finally it's finished! What a great feeling! But there is still much to be done - and I'm really looking forward to some projects I've already planned!
Maybe you'll see some of theme here, soon Wink .

Chemistry is like cooking - just don't lick the spoon!

Sag' nicht, es funktioniert nicht, bevor du es nicht versucht hast!
Benutzer-Profile anzeigenAlle Beiträge von Cumarinderivat anzeigenPrivate Nachricht senden
My HPLC Diary [English Version]
Du kannst keine Beiträge in dieses Forum schreiben.
Du kannst auf Beiträge in diesem Forum nicht antworten.
Du kannst deine Beiträge in diesem Forum nicht bearbeiten.
Du kannst deine Beiträge in diesem Forum nicht löschen.
Du kannst an Umfragen in diesem Forum nicht mitmachen.
Alle Zeiten sind GMT + 1 Stunde  
Seite 1 von 1  

Vorheriges Thema anzeigen :: Nächstes Thema anzeigen  
  Neue Antwort erstellen