Page 1 of 1

Column-making protocols

Posted: Tue Aug 09, 2011 1:52 pm
by Neil
I'm trying to familiarize myself with different column making protocols, as well as the benefits of different columns (differences in resins, frits vs. no frits, column diameters, etc.). Can anyone explain the benefits of using certain setups and columns over others?

Posted: Tue Aug 09, 2011 9:50 pm
by Doug
I have used three different procedures which I can describe here. For all three methods I used a precolumn for fast loading and an analytical column for separation.

1) The (mostly) store bought method

We purchased C18 cartridges and a peek holder that we used for precolumns. I can't remember who we purcahsed them from but it could have been upchurch. Analytical columns were pico frit columns purchased from new objective. They were 50-76 um ID with 10 um tips and had an integrated frit in the tip which was nice. I then packed them using a bomb with C18 packing material. 5 um particle size with 300 angstrom pores. I think we purchased the packing material from michrom (Magic C18). But packing material could probably be a whole other other thread all on its own. We resuspended the packing material in 50:50 ACN and MeOH and packed the columns to abotu 25 cm.

This method actually seemed to work pretty well. The columns almost never closed or closed up. But it was very expensive. If you can afford it this is a good way to go but you will be spending a lot of money. I guess if you dont have access to a laser puller it might be reasonable.

2) The bottle neck method.

I later used a method that is described in thispaper from the Hunt lab. Making the analytical column involved pulling a silica capillary using a laser puller just enough to make a bottle neck through which packing material could not pass. And then packing your column behind it. It was sort of a delicate method but it worked pretty well. I think there are easier ways to do this now though and I wouldn't recommend it. Some people just pull tips and pack right behind them, no frit or bottle neck, and this seems to work fine. I have no experience with that method but it seems like it might be a really good option. Hopefully someone else will post about it.

3) The Silicate Method

This one involves making frits by polymerization of Lithium Silicate inside the capillary. It is patented by Jarrod Marto. I am not sure about the legality of sharing this method but you can check out the patent here. It works great. Occasionally these tips did start to emit bubbles which disrupted the spray a tiny bit (could be a problem for quantitation). Also, pulling your own tips can be a bit delicate. This is the same for method #2. They have a tendency to close up and clog. The picofrit tips from method #1 were pretty robust to those problems.

Anyway, that is just my experience. Hopefully that help a little bit.

Posted: Wed Aug 10, 2011 7:08 am
I have also used all three procedures listed above. Of those three, technique 3 is the best compromise in terms of price (a few dollars per column), durability, and ease of fabrication. Of course, technique 3 (detailed here) will work best for analytical columns with integrated emitters if you have access to a laser puller. Read the supporting information in this Ficarro/Marto paper for details. In my experience, trying to pull a tip with the cast frit in place by heating followed by manually pulling a tip is a recipe for disaster. This polymerized silicate-based frit works much better for precolumns compared to making sintered frits because no capillary coating is removed for fabrication and no flame or spark is required for polymerization, just gentle heating.

Doug is correct about the tendency to emit bubbles with technique 3. In fact, this was often so severe that I abandoned the use of this frit technique altogether a few years ago. My first few months in grad school I would often watch LC-MS/MS runs through their entirety via the camera we had focused on the emitter tip at the MS capillary inlet (I know, such a dedicated student). There were portions of the gradient that were bubble free (usually the beginning) and others where bubbles were emitted at 5-10 Hz. This bubble problem became much worse as I tried larger capillary inner diameter (up to 200 um i.d.). This sort of problem (bubbling/outgassing at column frits) has plagued electrochromatographers for decades and is not unique to this style of frit fabrication. A combination of the pressure drop at the frit, the high surface areas of the frit providing lots of bubble nucleation sites, and maybe some minor Joule heating at the frit from the application of ESI voltage, all encourage bubbling. The only real solution to this bubbling issue is to completely degas buffers and maintain a blanket of helium over the mobile phases at all times. This is not practical for our lab because we have more mass spectrometers than LCs so we are constantly moving our LC systems (Waters nanoACQUITY) around the lab and short of setting up helium stations at each mass spec, continuous degassing simply is not a practical option.

Waters nanoACQUITY systems have vacuum degasser systems, but are bypassed for nanoflow applications. A Waters service engineer explained to me that the cylinder volume of each pump head is ~ 50 uL, so cylinders are not going to refill very often at 300 uL/min. Meanwhile, mobile phase sits in the degasser chamber under vacuum for hours. This probably isn't a big deal if you use all aqueous or all organic mobile phases, but the concentration of mobile phase mixtures will definitely be affected, as will the concentration of volatile mobile phase additives, being under vacuum for so long. Sparging/degassing without blanketing the bottles with helium is a temporary fix as Henry's Law dictates that the concentration of gas dissolved in solution is directly proportional to the partial pressure of the gas above the solution. Helium has incredibly low solubility (relative to N2, O2, and CO2) in water and common organic solvents for LC and will sweep those gasses out. As long as helium blankets the mobile phase after sparging the amount of dissolved gases in the mobile phase will be very, very low and bubble formation should be minimized. If you sparge/degas and then let the mobile phase sit exposed to the atmosphere for a few days, additional sparging/degassing will be required otherwise you will again suffer the consequences of Henry's Law.

Okay - so my preferred technique for making analytical columns with integrated ESI emitters comes with a disclaimer...I use hydrofluoric acid and this stuff is really, really nasty. Read the MSDS and be very knowledgeable of the risks of using HF. I always keep a tube of this stuff handy in case HF should happen to spill on my skin. I clean capillaries very thoroughly with methanol, pull a tip with a laser puller making ~ 1 um diameter tip. I immerse the tip into a polypropylene tube containing 50 uL of HF for 1-2 minutes and then flush methanol through the capillary for 2-3 minutes. HF etching has a couple of advantages, 1) it increases the tip diameter to about 5-10 um making packing much easier and greatly reduces the incidences of tips closing/clogging, 2) it thins the exterior bulk of the tips making a fine tip promoting more stable electrospray, 3) the tip can be salvaged when the laser puller is not pulling ideal tips (with 15-20 users, our laser puller is not always is perfect pulling condition).

Could I optimize tip diameter/geometry this with a laser puller? Probably. Would I have to spend a day each weak tweaking the laser puller settings? Nope, probably two days. After etching and cleaning the capillary I pack with my favorite stationary phase using a device similar to this product. At 800 psi, I can pack an acid-etched column (75 um i.d.) to 15 cm in about 30-45 minutes. All of this work gets me a column with an emitter that almost never clogs, has robust bubble-free spray, and awesome chromatography (zero dead volume).

Column making for nanoLC is an art. There are lots of little tricks, superstitions, and unusual protocols out there, so your mileage may vary.

Posted: Wed Aug 10, 2011 5:14 pm
by daniswan
I used to use the silicate method (#3 above), then I stopped doing that and just started pulling tips with a laser puller and packing directly. It seems to be working really well for me, I think I have only had one clogged tip in the past year. Also, for some perspective I'm packing 50-75um ID columns with 3um material.

What are people's packing material preferences? Right now I'm using: Maccel, C18 AQ, 200Å, 3µm, Highly Aqueous RPC, Nest Group Chromatography, Columns
It gives nice separation for more hydrophilic peptides, like phosphopeptides, but peaks broaden as the gradient progresses.

Posted: Thu Aug 11, 2011 11:04 am
My personal favorite reversed-phase material for peptide separations is Magic C18 AQ, 200 Å, 3µm, from Michrom (acquired by Bruker), beautiful unmodified peptide chromatography, although I am uncertain of how it performs for modified (phosphorylated) peptides. It can be used with 100% aqueous mobile phases, has very low column bleed/chemical background, is relatively inexpensive at $150/g, and is mechanically strong to at least 5000 psi in my experience. I prefer the 200 Å pore size because of the lower back pressure, but in terms of practical performance (peptide IDs) it is nearly identical to 100 Å material.

I have also tried these C18 materials: Waters BEH/Xbridge, Alltech Alltima HP, Phenomenex Jupiter, Dionex Acclaim and a few others. They all perform very similarly for complex separations (thousands of peptides per injection) on homemade capillary columns. I do think materials with embedded polar groups that permit 100% aqueous conditions without suffering stationary phase collapse do offer very real advantages for retention of hydrophilic peptide species. But, in terms of first separating a proteome into a dozen or so fractions (SCX, HILIC, ERLIC, high-pH) and then throwing it on a reversed-phase column, nearly all modern reversed-phase materials from well-known vendors will probably work for large-scale, low-pH peptide separations.

I've heard really good things about a material called TARGA distributed by vendors like the Nest Group and Western Analytical. Has anyone tried this material that has also tried other 100% aqueous compatible C18 materials?

Posted: Mon Sep 26, 2011 2:28 pm
by Toxic
JDRCHEM wrote:I clean capillaries very thoroughly with methanol, pull a tip with a laser puller making ~ 1 um diameter tip. I immerse the tip into a polypropylene tube containing 50 uL of HF for 1-2 minutes and then flush methanol through the capillary for 2-3 minutes.

Can I just clarify something? After pulling the tip with the laser puller, is the tip open or are you using the HF to both open the tip and "clean up" the tip so its nice and flat? I have been thinking about doing a similar thing with a butane flame pulled tip.


Posted: Tue Sep 27, 2011 5:25 am
The majority of the time the emitter tip will be open after being pulled on the laser puller. But, there are periods when the laser puller will pull closed tips requiring one of us to clean the mirrors or make fine adjustments to the tip-pulling method. HF is particularly advantageous when the tip is closed as it almost always opens the tip salvaging the emitter. I also like to use HF to modify the tip geometry. Occasionally, long filamentous tips are pulled that do not spray well for me so I use HF to eat away the excess material creating a shorter more cone-like geometry. Some folks will use a fused silica cleaving tool to remove the excess tip material, but I have better luck using HF.

As a side note, many of the emitter fabrication methods using HF I have seen in the literature will bubble some inert gas through the capillary as it is being etched. I have found this unnecessary. Of course, the internal diameter of the fused silica will be slightly increased, but in my hands it is only by a few micrometers.

So, HF can be used to both open/widen fused-silica emitter tips and alter the tip geometry.

Posted: Tue Sep 27, 2011 3:47 pm
by Toxic
Awesome. Thanks for that. I will try it out.

Related to this, it was thinking of combining this with the point heat source silica frit method found in this paper from the Marto lab to make something like a PicoFrit.

Posted: Tue Nov 29, 2011 4:32 pm
by daniswan
I just built a little column oven and I want to try using some sub-2um packing material. Does anyone know of a US vendor who sells this type of bulk media? All I have found so far is Reprosil (which appears to have no US vendors) or purchasing a pre-packed column and dumping it out.

Posted: Wed Nov 30, 2011 4:15 pm
by sesf43
I would like your opinion regarding analytical/trap column "storage". I currently use the Magic matrix you mentioned above, and would like to know how you store a used column/trap long and short term. I would like to leave a used column and trap plumbed in and ready for a week at a time without drying the matrix by stopping flows, or removing them for storage. What solution could I flow over the matrix without compromising its integrity? I am currently leaving the column flows at 50% ACN and reduced flow rates (200nL/min, which is one half of our analysis rates), for a few days at a time and have noticed peak broadening and reduced sensitivity, which I am guessing is due to stationary phase breakdown of some sort.

Posted: Thu Dec 01, 2011 9:15 am
If I am not actively running samples I generally flow at 50% ACN (w/ 0.2% formic acid). I usually keep the flow at the same flow rate as my analysis, but decreasing the flow should not make much of a difference for column lifetime. I have not noticed a change in column performance after a couple of days at constant flow. It is rare for a nanoLC to be idle for more than a day in our lab so I have never run a column at constant flow for that long, but I cannot imagine that a week at constant flow would cause a dramatic loss in performance. For long-term storage I store both my trap and analytical columns dry. This may not be the best method for long-term storage, but it works just fine for me. I have read that some people will store their capillary columns in vial with some buffer, but I find this unnecessary.

What could cause the performance issues you described? You could be sloughing off stationary phase into the tee (assuming you are doing vented-style trapping). This happens from time to time with homemade frits. If I suspect this is happening I measure the length of the stationary phase in the column to determine if it is decreasing over time. However, I can usually tell by a decrease in system pressure that I am losing stationary phase. Your system may have low-level contaminants that build up and are not easily removed affecting chromatographic performance. We have seen this with PEG contaminants from poor quality mobile phase inlet filters. Sometimes other crud makes its way into the system. I often examine the stationary phase at the inlet end of the trap column under a microscope. The stationary phase can be brown to black from the buildup of stuff from samples and the system (usually rotor seal wear). The mechanical strength of silica-based Magic stationary phases is 6500 psi, so if you are above this you might be seeing degradation of the material. The mobile phase pH is also somewhat of a concern. If your pH is around 2 or 7.5 with Magic materials you are slowly degrading the material, but this should manifest itself over a longer time than a week.

The number one cause for the type of problem you describe (peak broadening, decrease in sensitivity) in our lab is from leaking nano-fittings or improperly made connections. Leaks are typically so small that they are difficult to see, but are large enough to impact chromatographic performance. I check the flow rate and system pressure with the trap/vent line connected to the tee and then I replace it with a tee plug and repeat those measurements. Small leaks can gradually become worse if you are using PEEK finger-tight fittings and you exceed 4000 psi. Finger-tight means different things to different people and it takes about a few months for new people in our lab to be able to make these connections routinely so that they are tight enough to prevent leaks but not so tight that they crush the fused-silica capillaries. If the leak is in the trap/waste line then during analytical flow part of your sample will go to waste causing both a decrease in sensitivity and noticeable peak broadening.

We do not work in a regulated lab so we do not have to qualify each batch of columns for performance which can be timely. So, if I were in your situation where performance was good one day and then poor a week later I would first check all of the connections and then my next step would be to put on a new column (ours cost ~2$ US they take 10 minutes to fabricate) to minimize downtime.

Good luck!

Posted: Thu Dec 01, 2011 9:32 am
I have not had luck with finding sub-3 um fully porous stationary phases from US vendors. Some vendors are willing to sell small particle bulk material normally reserved for packed columns if you are willing to pester them enough. Other vendors will sell the bulk material but at about the same cost as buying the entire column due to their cost of packing, testing/qualifying, and then unpacking a column. At that point, the cost of them testing the column is more than the cost of the column housing. As more column manufacturers begin making capillary columns (and profiting from them) I think buying small particle bulk material from them is going to become even more difficult.

Posted: Fri Mar 30, 2012 1:10 am
by yuva2004
Great suggestions

Posted: Tue Jun 26, 2012 10:07 am
by zhen
Hi All, I found this thread highly informative and inspired me to start trying to make my own cap columns. We are using a nano-acquity system, and I am wondering how everyone is coupling the integrated tip and column to the LC/MS? Thanks

Posted: Tue Jun 26, 2012 11:51 pm
by Toxic
Upchurch (now IDEX) make a nanoflow voltage junction with a mouting bracket. I use it where the inlet capillary from my CapTrap connects to the inlet of the packed PicoFrit. Very simple, easy to fit together and totally reliable.

Oh and I mount the PicoFrit in the MicroIonSpray II head that came with my QSTAR Elite because you still need nitrogen to get stable spray with that instrument. There is another round voltage junction with a wider through hole that comes in the spare parts. The column fits all the way through and out the back and is held in place with a ferrule-swaged orange Upchurch sleeve.

Posted: Wed Jun 27, 2012 5:01 am
We used a vented-style trapping configuration where we apply the ESI voltage to the waste line through a stainless-steel union. We also purchase our nanoflow fittings from mostly from IDEX.

I'll break down every connection and give a part number for each.

Connection from injection valve to capillary line: stainless-steel nut (IDEX U-313X, check valve thread compatibility), PEEK tubing sleeve (IDEX, F-230), PEEK high-pressure ferrule (VICI-Valco, ZF1-PK10)

Connection from injection line to capillary precolumn: PEEK zero dead volume union with sleeves and ferrule (IDEX, P-720)

Connection from precolumn to analytical column and waste/trapping line (with ferrules and sleeves): PEEK microtee (IDEX, P-775)

Connection from tee to union in the waste line: stainless-steel union with nuts (IDEX U-435), PEEK tubing sleeve (IDEX, F-230), PEEK high-pressure ferrule (VICI-Valco, ZF1-PK10)

I routinely run at 5000 psi with these fittings and have been able to go as high as 6000 psi (sustained) with no leaks. The PEEK fittings are rated for 4000 psi. When I teach new students how to plumb LC lines and connections I ask them to assemble the components and then disassemble them and inspect the capillary ends under a microscope. There is a fine line between getting a leak-free connection and fracturing the fused-silica capillary - I think this is an under appreciated aspect of plumbing up the LC by novice users.

We have custom nanoESI sources that we use on Thermo mass spectrometers. In my opinion, simpler is better. I know this is dependent on the instrument vendor, but one does not need Star Wars Death Star technology for a nanoESI source. All one really needs is a source that allows fine positioning of the emitter at the mass spectrometer inlet. Our source is different from this one, but similar in concept - simple, but effective (everyone should check out this site, it has a lot of useful information).

Posted: Wed Jun 27, 2012 12:31 pm
by zhen
Thanks, JDRCHEM and Toxic for the replies. I am glad I found this forum and can't wait to contribute soon!

Posted: Mon Oct 27, 2014 7:44 am
by proteomics

Thank you for all the information above, it is very helpful. I had one question.. my apologies if it is obvious. If I am trying to use the laser puller for tips to pack my nano columns, do I need to have frits on the capillaries before filling them? or once I pull the tips using a puller I can directly, fill it with my packing material.

Thank you

Posted: Mon Oct 27, 2014 8:41 am
by daniswan
You can directly fill with packing material after pulling the tip, or you can make a frit, pull tip, then fill with packing material.

I like to use 1.8um particles for my columns and I worry about those not being well retained in the tip if I don't have a frit. So to be cautious, I typically pack a few mm of larger particles at the tip (3um or 5um), then fill the rest with 1.8um.

Posted: Mon Oct 27, 2014 9:12 am
by proteomics
In that case my question would be, how do you make frits when pulling tips for nano-columns? I usually use the Kasil method to make my trap columns.

Any suggestions for improving packing speed

Posted: Thu Sep 17, 2015 3:02 am
by roberthardt

just stumbled upon this discussion. I'm curious if any one of you has good suggestions to improve our packing speed. Before we were mainly packing 5 µm C18 material into 100µm ID self-pulled fused-silica capillaries. Now we want to switch to 3µm material, and the packing process takes forever if it completes at all. Problem is that we want to pack 40 cm (length) columns so I expect to have some amount of backpressure building up during the packing process. Is there any other option to increase the velocity besides applying a higher pressure. We currently use 50-60 bar.
Do you, for example, think a change of the solvent we use for making the bead slurry could help? Currently we are using acetone.
Any suggestions are greatly appreciated.