Centrifugation: Spin it Faster. More More.

posted by Dave Arnold

We asked for comments on centrifugation and got some really great responses—so many, in fact, that we wanted to do another post.

If you don’t know about centrifugation (what’s a rotor, what are g forces, why the hell should I care, etc., etc.) see our other posts here and here.

To recap, we have two useful centrifuges, the Sorvall RC-2B and the Jouan C412:

Our two fuges, Big Mama RC -2B on left, Baby Jouan C412 on right.
Our two fuges, Big Mama RC-2B on left, Baby Jouan C412 on right.

The RC-2B is an awesome workhorse. It is refrigerated, built like a bomb shelter, and, with the right rotor, can do  48,000 g’s as 20,000 rpm. The disadvantages are that it is very large, and with our current GSA rotor we are limited to 1.5 liter batches using little bottles with fairly narrow necks (they are hard to fill and clean out). We love it to death, but all in all, it is difficult to see a restaurant dealing with such a large piece of equipment.

Our C412 is not refrigerated, but is smaller, cheaper, slows down quicker, has swinging bucket rotors, and can do 3 liters at a crack. The maximum number of g’s it can pull is about 4000, but 4000 g’s has been sufficient for a lot of our work. We have been using plastic bags to line the rotor buckets. They are cheap and easy to load but aren’t as neat as bottles and tend not to work well on things that don’t form hard pucks like nuts do.

Kalamata spun in a bag. Not as clean as in a bottle.
Kalamata spun in a bag. Not as clean as in a bottle.

See what a mess the kalamata olives are in a bag? In a hard bottle you get three clean layers: brine (awesome), flavorless pulp (garbage), and cured kalamata oil (awesome). We need to figure out a way to make a hard, thin, cheap plastic liner for the 750 ml buckets in the Jouan (advice welcomed).

As for refrigeration, several readers suggested putting the fuge in the walk in fridge. A good idea. We just got a new walk-in recently, so I might just have the space. We’ll let you know how it works.  The reason you want refrigeration, by the way, is that the friction from spinning heats up the rotor, buckets and machine.  Refrigeration fixes that.  Another option is pulling a vacuum on the fuge. This keeps the friction down and, as schinderhannespointed out, allows you to do centrifugal evaporation.  Next fuge.

SPECIAL HELP NEEDED: Today our Jouan, which had been spinning at 4200 rpm nicely for a while, suddenly started topping out at 3200 rpm. Before I rip the machine apart, do any readers have suggestions?

We purchased the Jouan for several reasons:  We wanted a centrifuge we could take to demos.  We wanted a centrifuge that has a large capacity.  Most of all, we wanted a centrifuge that other chefs might find useful.  We also got a great side benefit: we can spin our RC-2B faster than before without worrying. If you have read any of our earlier posts on centrifuges you know we are concerned with metal fatigue. Our rotors are made of aluminum—old aluminum, which could break apart at any second if subjected to too many g’s. The GSA rotor in Big Mama fuge was rated for 27,000 g’s when it was new. We had been running it at 4000 g’s (5000 rpm) because we figured there was no chance of failure at that speed. If we ever DO have a failure the fuge would be destroyed. The rotor would fly apart and completely ruin the whole machine. Now that we have another fuge that can do 4000 g’s, we have pushed our RC-2B up to 10,000 g’s (aw yeah). 10,000 g’s is still well less than half what it was rated for new, so we feel safe. Even if we throw the rotor, the damage would be contained inside the fuge.  We would lose the fuge, but we would still have another one, and no one would get hurt.

So now, assuming we can fix the Jouan, we have a 4000 g 3 liter fuge and a 10,000 g 1.5 liter fuge. Now we have to think of applications. Our readers helped us with ideas.  Before we get serious let’s get stupid.

The Stupid:
One of our original tests was to spin avocado puree to see if we could get avocado oil. We couldn’t. It didn’t work at 4000 g’s or at 8000 g’s. All we got was really smooth guacamole. I mean REALLY smooth guacamole. I was very discouraged by the lack of oil and got in a weird mood, so for giggles we spun whole avocados just to see if they would crush

Whole avocados balanced for spinning. We know it's stupid
Whole avocados balanced for spinning. We know it's stupid
Whole avocados 4000 g's 20 minutes.  Stupid.
Whole avocados 4000 g's 20 minutes. Stupid.

Not as impressive as we had hoped.

The problem with making avocado oil is that the oil is contained within the cells and the fuge isn’t powerful enough to break the cell walls. Even the Vita-Prep isn’t sufficient to break the cell walls. Kent Kirshenbaum, our chemistry professor buddy from NYU and the Experimental Cuisine Collective, suggested putting the avocado through several freeze- thaw cycles.  The ice crystals that form each time you freeze the avocado make lots of tiny holes in the cell walls. Another idea Kent and reader Paul A. suggested was to hit the avocados with our 400-watt Branson ultrasonic homogenizer. The ultrasonic homogenizer (aka the sonic cell dismembrator—how cool is the term cell dismembrator?) is a machine that does two things. It sends people running from the room in agony because of the piercing noise it produces; and it shakes back and forth so fast and so hard that tiny bubble are formed in your food that break with enough force to mash things up very, very finely. We tried both. They both worked. Ultrasound won for yield (it was also a pain in the butt to freeze and thaw, freeze and thaw, freeze and thaw):

Ultrasonically homogenized (400 watts, 5 minutes, 500 grams) and spun (10,000 g's 20 minutes).
Ultrasonically homogenized (400 watts, 5 minutes, 500 grams) and spun (10,000 g's 20 minutes).

The problem with both techniques is that even when the avocados were heavily pre-treated with ascorbic acid to prevent oxidation, the act of breaking apart the cell walls produced some funky, nasty flavors. The oil looked great but tasted like old artichoke, not fresh avocado. Maybe if we cooked the avocado first?

Avocado oil.  Looks great. Taste, not so much.
Avocado oil. Looks great. Taste, not so much.


Herb oils:
Several people asked about cleaning up herb oils. We didn’t have much luck at cleaning up herb oils at 4000 g’s, but 8,000 g’s did some nice stuff to this basil oil:

Basil oil. Left side spun at 8000 g's 15 minutes, right un-spun.
Basil oil. Left side spun at 8000 g's 15 minutes, right un-spun.

The spun oil was not only clearer, it had an amazing ruby color and tasted much cleaner and less bitter, if slightly less complex than it’s un-spun counterpart.

Citrus and Other Essential Oils:
We have not had any luck with citrus oils.  I had mistakenly said we had tried to extract citrus oils using alcohol. As Paul A. pointed out, you don’t want to use alcohol, cause that will dissolve your oil.  I think he is right.  Looking back on my notes we DID use water. We ran orange peels blended with water at 4000 g’s for 20 minutes with no results.  The stuff tasted like it would make a great bitters so I made a note that we SHOULD try it with alcohol.  I still want to make a good citrus oil.  The best book on distillation I own is from 1871.  The bad news is it is outdated.  The good news is that it is free in the public domain on Google Books. On page 355 of A Treastise on the Manufacture and Distillation of Alcoholic Liquors by Pierre Duplais, the methods of obtaining essential oils are discussed: expression, which is what the fuge can do; distillation; and maceration in fat (fat washing/enfleurage). He then goes on to explain the differences in flavor between the techniques and when you should use them.  Turns out expression should work fine for citrus oil.  We should be able to zest a whole bunch of citrus and just press it out.  Again, I haven’t had any luck.  Maybe at 10,000 g’s with a homogenizer.   We’ll keep you posted.  The problem we are having with citrus would, I assume, translate to any other product that requires us to separate a very small amount of oil from a large amount of dry/liquid matter.

On essential oils without a centrifuge:

I have been wanting to do super-critical CO2 work for some time but have not had luck finding a unit or figuring out a safe way to jerry-rig one.  Any thoughts?

 I tried butane extraction like they do to make “honey oil” from swag pot, but didn’t feel safe and had no luck. 

I have also never set up my rotovap as an essential oil recovery machine.  Maybe some day. 

The most exciting thing with essential oils and flavors we are working on now is a really cool form of chromatography.  More on THAT soon. 

Nut Oils and Sugar/Water

Mindy mentioned in the last post that we were getting some interesting results from spinning nuts to recover nut oil.  We are planning on actually selling pecan oil out of the restaurant (if we are allowed to).  Why pecan? It is delicious, it is American, and we’ve never had it before. Anyway, we noticed that adding simple syrup to the mix before we spin drastically increases yield of oil. The oil we get is way is much clearer, sometimes less bitter, but always less intense than the oil we get without simple syrup.  Take a look at macadamia nut oil spun with and without simple syrup:

Macadamia oil without simple on left, with simple on right.
Macadamia oil without simple on left, with simple on right.

Wow. Big difference. 

Here is how this simple syrup thing happened:

When you spin nuts you first liquefy them in a Robot-Coupe then ultra-pulverize them in a Vita Prep.  This breaks up the nuts and also heats them up a lot. The nuts get so hot that we no longer toast them beforehand.  They are “blender toasted.” Hot nuts are more fluid than cold nuts, so we like to spin them warm.  When you spin pecans this way, you get a layer of oil, a layer of super-primo nut butter, and a layer of pecan paste+skins.  The skin layer is bitter so I separate it out to use in sweet pastries, etc.  One day, the primo butter got mixed in with the 2nd quality stuff.  That sucked.  I figured let’s re-spin it to try and separate it again.  It was too thick to spin so we added simple syrup.  The results are below:

Pecan on left without simple. Classic three layer pecan --oil, first quality butter, bitter chaff butter. Pecan on right with simple syrup, high yield of oil and a pecan brick.
Pecan on left without simple. Classic three layer pecan --oil, first quality butter, bitter chaff butter. Pecan on right with simple syrup, high yield of oil and a pecan brick.

Wow.  The one on the right had ALREADY been spunonce and still yielded more oil than the fresh-spun batch on left.  Look how clear the oil is!  Weird.  Look how solid the hockey puck of paste is. That extra oil that comes off of the pecans isn’t as heavily flavored as the regular stuff.  FoodPlayer asked if removing the extra oil from the paste makes the nut paste more flavorful.  I guess it does, but it is hard to say because the texture is so different.

We ran another experiment with cashews (I said pecans but one of the interns mis-heard me, oh well).  This time we wanted to measure the effect of simple syrup versus water and the effect of varying the amount of simple syrup. Here it is:

Cashew butter (oil drained). Left to right, 400 g cashew, 311 g cashew plus 89 g  simple 1:1, 350 g cashew plus 50 g simple 1:1, 350 g cashew plus 50 g water.
Cashew butter (oil drained). Left to right, 400 g cashew, 311 g cashew plus 89 g simple 1:1, 350 g cashew plus 50 g simple 1:1, 350 g cashew plus 50 g water.

The water didn’t incorporate as well as the simple syrup. Overall the simple syrup was more effective.  Some of our readers weighed in on this phenomenon. Shinderhannes wrote: 

The most effective additive to help break [a tough] emulsion is a few mls of methanol (stop, don´t try this but ethanol (go head) does the trick as well.

Therefore maybe adding a shot of booze to the nut puree will work like a charm!
As you certainly know, but maybe not all readers, glucose is a poly-alcohol, from a chemical standpoint, so maybe that does explain why SS is so effective….

Great advice. We plan on running more tests soon

Still to Try:
Foie Gras, suggested by Shinderhannes and Jeremiah (Shinderhannes also suggested liver sausage. Nice)

Chorizo (I think that would be good) suggested by Alex T.

Fish oil, suggested by Derek.  Don’t know how to do this one, but would be fun.

Chilled Ouzo, by Shinderhannes.   See the Ouzo Effect. Could we just water it down, precipitate the oil and spin it?

Stuff I don’t know how to do/ have other techniques for:

FoodPlayerasked about milk/cream.  That will definitely work but we have a small cream centrifuge for that.  If you have a specific application (like a really nice milk) it would be worthwhile. I dunno.  Hit me back with more ideas.

Derek asked about cereal grains/hops.  I can’t figure a way to do the grain.  The hops work great in a rotovap, and I’d bet we could do other forms of distillation/etc, but I think we’d run into problems in the fuge.

Thank you for your help. Please keep the comments coming.

30 thoughts on “Centrifugation: Spin it Faster. More More.

  1. Perhaps others would like to play along at home? I’m thinking hand-crank.

    No, not ice cream — hand cranked centrifuges.

    Although some applications may require high-speed, large volume ‘fuges, there may be some interesting products that can be obtained on the cheap. It seems intrepid individuals can dumpster-dive for low-volume, low-speed centrifuges. But the phrase “…bleaching the rabies out of it…”, while evocative, is rather alarming.

    Instead, for a low-tech approach, I’m wondering if people know that there are hand-cranked centrifuges. I used one back in the mid 20th Century, and they actually spin surprisingly fast. They look like they could do severe damage to personal extremities, which may have a certain appeal. They are also fucking cheap. New will run you around $400:

    I found some on e-bay for $85. And I didn’t even need to misspell the word centrifuge.

    Give it a whirl. Try it on your nuts.

    Regarding the ultrasound on avocodo oil, I am guessing you are forming epoxides by exposing unsaturated fatty acids to ultrasound in the presence of oxygen. I will dig up some journal articles that may be relevant. The ascorbate may not be sufficient as an anti-oxidant. Perhaps we could put an argon blanket over your guacomole before you blast it. I’m guessing you want minimal power as well. I’d set your machine to Sade, not Karen O. Anyhow, I’ll try to follow up on why ultrasound makes things taste nasty.

  2. Maybe try hot chiles. I know that you can run chiles in a rotovap to remove the oils and the heat, but it might be possible in a centrifuge. I’d imagine that a ghost chili with lowered or no heat might actually be tasty.

    1. I think you’d be much better off with a rotovap. I don’t think you’d get any separation of the capsaicin in the fuge. I have done many, many kilos of Habaneros and love the result. I also wonder if you could get a concentrated spiceless chili oil with steam distillation. Hmm. I also have high hopes for a new Liquid-Liquid chromatography technique that a company called Cherry Instruments is introducing us to. A post is coming on that soon.

      1. Hiya Dave,

        I agree that rotovap is still the way to extract heat-less chilli aroma.

        You should also be able to extract some heat-less chilli oil from steam distillation with water, since wikipedia states that the boiling point of capsaicin is >200°C. But don’t expect the aroma to be as good as the rotovap extract, as you are using high temperature.

        But then again, you can enhance the aroma of the steam distilled oil (for the taste) by adding back some of the rotovap extract.

        As for liquid-liquid extraction, I think it’s gonna take you a lot of time to get a concentrated enough sample. Besides, which food grade solvent are you gonna use for this method? Also, will the solvent you use extract capsaicin too?

        Can’t wait for your next post. 🙂

  3. Are you getting the 3200 RPM figure from the machine’s own readout, or confirming with a separate stroboscope or whatnot? Best-case scenario it’s just a display error.

    1. Howdy Paul,
      I wish. Although I have not verified it, my ear tells me we are not running full speed. hopefully I’m wrong. What’s odd is the speed is the same loaded or unloaded within 100 rpm. I’ll have time Monday to rip it apart.

  4. I’m wondering if the icky avocado oil is due to oxidation, or oil-soluble volatiles that accompany the oil. I imagine with all the agitation, there would be a lot of opportunity for the avocado to oxidize, vit-c or no. These guys (http://sci-toys.com/ingredients/ascorbic_acid.html) imply that ascorbic acid is not fat-soluble anyway.

    Have you considered doing the avocado oil in a vacuum? (I.e. toss it in the chamber, suck the air out, then abuse and `fuge all in the bag.) Not sure how much free O2 is in the flesh, nor what a proper chamber vac can do, but it may be able to rule out the oxidation hypothesis.

    Apparently the avocado oil you can buy in the store is done with mechanical agitation, settling, and then spinning. (http://www.olivado.com/studies4.htm) There’s lot of documentation out there, for those who are better than I at reading food science papers.

    1. Thanks for the link, that’s a very useful report on avocado oil. It confirms that oxidation may be a problem. It seems as though the presence of chlorophyll in the oil could also lead to oxidation reactions promoted by light. So in addition to low oxygen, perhaps work under low light.

      It makes me wonder about the discredited “avocado pit in the guacomole” experiment. Maybe the pit is not only blocking oxygen from contacting the avocado, but is blocking light, too.

      From the report:
      “High levels of chlorophyll in avocado oil (40-60ppm) can adversely affect the oxidative stability of the oil, causing rapid formation of oxidation products via the photosensitised singlet oxygen pathway (4) when stored under light. However, the emerald green colour of the oil (originating from high chlorophyll levels) has been identified by consumers as desirable, thus a novel approach to increasing the oxidative stability of this unique oil is currently being studied. This will include minimising oxygen and light exposure during handling of the oil (including shelf storage in dark glass bottles).”

  5. How bout adding dry ice to the avocado puree to keep eveythin under an inert atmosphere of CO2, might be more efeective than addin vit C….

    Also on the Ouzo issue, those droplets seem to bee extremly small and well stabilized thru the EtOH in the H2O. Them crystals in forzen ouzo are a lot bigger (and beautifyl shiny little tiles). I bet those are easier to fuge down.

    Good Luck!

  6. how about wine,can you spin red wine and get white wine as a result?

    i’d like to see a result for milk also, would you get a clear liquid without taste or a clear liquid that still taste like milk??

    1. The color in red wine is actually in solution,so I can’t spin it down. I could probably make milk clear with clarification techniques.

  7. when you spin in a bag, try inflating the bag before you seal to keep the bag from compressing as much when you spin.

  8. On the 3.2KRPM limit, I know bigger centrifuges with replacible rotors like your Sorval have indicators on the rotor that tell the fuge what their speed limit is. These are sometimes optical and gunk or scratches can make them fail.

    The benchtops usually have fixed rotors (just replacable buckets) so this probably isn’t the issue, but you might check.

    As to rotor failure–you are betting the life of anyone in the area that the centrifuge will contain a failed rotor. I have heard 8th-hand urban legend type stories that a failed rotor can escape the centrifuge, IOW I don’t know if they are true, but…. This is not an event you would be able to flee, it will basically be an explosion.

    You are supposed to log the hours on the rotor and get them checked every so many g-hours. Might be time to get yours checked.

  9. Hi guys, for the fuge buckets you might try Lineing the buckets with the plastic bags and filling the bags with an expandable styrofoam that you could carve out holes to recess delis in once the foam is set. Then you just place items in the deli and put a lid on it.

    As a side note, I’ve been playing at home with a branson B5200 ultrasonic bath. I’ve been able to sous vide in it as it heats to 69C. It also works well for a quick emulsion if you use an offset method, plus it will hold an emulsion all day. What I am trying to do in it now is use the ultrasonic waves to tenderize and marinade tougher cuts of meats in a shorter time frame. Are there any findings along these lines you could comment on?

    1. Howdy Rick,
      Tell us about your experiences with the ultrasound. We have one as well. Unfortunately with the fuge, deli containers are slightly too big. We tried them. they are also hard to shrink properly to make them fit.

  10. You may already know this, but it is important to make sure that the weight in each of the Jouan’s swinging buckets is equal and balanced so you can reach maximum speed. I used to use this model, and I found that balancing just two of the four buckets opposite each other is adequate. The manufacturer may recommend balancing all four.

    Also, what other uses have you found for your ultrasonic homogenizer?

    1. Howdy Eric.
      We keep our buckets balanced to within 2 tenths of a gram (the resolution of our trip balance). I think the bearings and brushes might need replacing. We don’t use the homogenizer for much except demos (re-emulsifying beurre-blanc and buttermilk into cream, etc). We also use it to unstick frozen glassware. Any suggestions would be appreciated.

  11. On the red wine pigments and other in solution, organic compounds, you might want to try buying some activated carbon, agitating for 10-20 mins (blender, stiring, sonicator) and then centrifuging. You don’t ned a lot, and if you use too much, it may strip out too much of the other flavors and aromas you want. Be careful, activated carbon likes to go every where.

    Might also be something to try with the oils, though I’ve never had a chance to play with those two together.

  12. I know this is pretty late, but my ochem lab book had a procedure for extracting essential oils where you put your spice in a filter paper cone that had the spice in it with a copper wire that had 2 loops in it that kept the paper off the bottom of the contraption and the end sticking up for a handle. You then fill the tapered portion of a plastic (very important!) centrifuge tube with dry ice, lowered your filter paper and handle thing into the tube, and filled the rest with ice. Cap the tube, then lower into a beaker filled with warm water. The dry ice will liquefy after a few minutes, then stay liquid for ~9 minutes, then start vaporizing. After it has all vaporized, you can open the tube and lift out the filter paper and spice. There should be some essential oil in the bottom of the tube.

    1. Hi Amy,
      I will have to look that up. How do you regulate the pressure so the CO2 stays liquid without blowing up the centrifuge tube?

  13. “We need to figure out a way to make a hard, thin, cheap plastic liner for the 750 ml buckets in the Jouan (advice welcomed).”

    This is an old post, but have you solved the issue yet? This might be a completely useless idea, but the first thing I thought of was vacuum forming.

    Is the “well” too deep for that? I’m not experienced with the technique, but I would think/hope there are ways to create air channels (even an air-permeable mold? Cast a replica bucket out of something semi-soft and stick a fine needle through it in several places?).

    If vacuum forming works at all, I’d guess the same plastic you used for your custom rotavap lid would be durable enough. I don’t know about thin, though.

    Anyway, that’s just what occurred to me. Maybe you already ruled it out.

    1. Hi Dave,
      I have not figured out this problem yet. It would be a deep draw for vacuum forming. Plus, you’d have to make a cast of the inside, then shrink that cast enough to account for the thickness of the plastic. I’m sure it is doable, but I haven’t figured out how to do it without access to a shop. I’d love it if someone could solve the problem.

      1. What if you side-step the issue of shrinking the cast by making a cast of the inside, and then making a cast of that? Ending up with a replica of the centrifuge bucket.

        Then you draw your plastic sheet down into the bucket.

        Now, that’s trading off one problem for another as I’m sure it’s not smooth sailing to do the deep draw that way. But I can imagine fooling around with air channels and whatnot… the plus side is if you get it to work, it WILL be the right size!

        This is sounding interesting, I wish I had a shopvac and some plastic lying around to try rigging it up.

  14. To extract essential oils using a rotovap or other glassware rig, stuffing a bump trap with herbal matter and attaching it after the boiling flask may be a cheap and easy solution. I plan to order one come springtime when oily things start growing again.

    Also, there are a few more modern books (from the 1920’s) by Parry that you might want to check out. They’re also public domain, but the Google Books scans are rubbish, unless you dig looking at some twat’s sausage fingers obscuring every other page. ScienceMadness forums has cleaner copies. The first volume is a compendium of oils and their properties, starting, of course, with Oak Moss Oil:



  15. If you are still looking for 750 ml liners for the centrifuge, try speaking to someone at the Uni in the biology/microbiology/genetics dept. I’m fairly sure there are 750ml wide mouthed bottles available for spinning down cells from batch preps of bacteria and the like, think I’ve used them myself. At least they could tell you where to order them from.

    1. Howdy CookAndSplash,
      750 wide-mouths do exist (Wylie uses them at WD50) but widemouth is a bit of a misnomer. Nothing beats the bucket so far.

Comments are closed.