Org Prep Daily

January 18, 2012

Replacement process solvents

Filed under: procedures — milkshake @ 1:45 pm

A recent Organic Process R&D editorial (thanks Chemjobber for pointing it out) publicizes Pfizer Process Group green solvent replacement chart that discourages chemists from using solvents that are either known to be toxic, dangerous to use on large scale or are expensive to dispose as waste. OPR&D makes it now a submission policy that if you used a problematic solvent in your work you have to demonstrate in your paper that you tried (and failed) to find more process-friendly alternatives. I think it is a sensible policy for a chemical industry process journal (and it probably makes the editors job of rejecting marginal manuscripts easier).

Also, Innocentive challenge was recently promising an award (up to 8k) to a winning proposal for replacing dipolar aprotic solvents like DMF, DMAc, NMP with less enviro-problematic alternatives.

I have few comments on the recommended solvent replacements in the table:

1) Acetonitrile is a perfectly good replacement of other dipolar aprotic solvents for things that dissolve in it, unfortunately MeCN dissolving power is quite poor. On the other hand, DMSO is famously bio-innocuous and it dissolves almost anything organic, and quite a few inorganic salts as well. But DMSO properties can complicate the workup, and DMSO can participate in quite a few unwanted sidereaction. I think overall DMSO is a pretty good media for alkylations that involve a reactive nucleophile. If the alkylating agent is highly reactive one might end up with S-alkylated DMSO-derived sideproducts although for many reactions this is not really a problem. Boiling DMSO has oxidizing properties and gives off Me2S funk so the reactions run in DMSO should not be heated above 140C. For acylations (where DMSO would interfere badly) an inexpensive eco-friendly solvent to try is 1,2-propylene carbonate, perhaps diluted with MeCN or DCM to cut down on this high-boiling solvent and to lower the viscosity. Propylene carbonate stability is quite remarkable – it tolerates alkali metals – but I would not heat it with alkoxides and reactive amines, the same limitation as with DMF and NMP. Another possibility for acylations is sulfolane-MeCN mixture. Adventurous eco-fanatic types may even go for triethylphosphate, another cheap degradable goo.

2) A suitable alternative for replacing DCM and DCE in many reactions (but not for AlCl3-promoted Friedel-Crafts) is trifluoromethylbenzene, bp. 102C.

3) For pyridine replacement the chart recommends NEt3 but I think N-methylmorpholine would be a closer surrogate/better alternative – NMM it is much less basic than NEt3 thus less prone to cause ketene-related dark impurities and racemizations during acylations, and it is a better solvent also. A strong fishy reek of NMM is a bit put-off though. If one so desires, Grignard reagents can be prepared in NMM.

4) One relatively underused process solvent is di-n-butyl ether. Its odor is annoying, the boiling point is quite high (142 C) and the dissolving power of Bu2O is not great but this solvent is cheap to buy and easy to dry. Room temperature lithiations with BuLi that require an etheral co-solvent might be a good pick for Bu2O (THF gets cleaved with BuLi at room temperature at appreciable rate, MTBE is pretty inefficient for solvating Li)


  1. What about CPME as an ether? Is this rubbish for solvating Li too?

    Comment by Erhuma — January 18, 2012 @ 2:35 pm

  2. I don’t know – It would be worth trying. A good suggestion. We have CPME and the smell is not bad (unlike Bu2O which reminds me a of bad flavor of a bubblegum toothpaste)

    Comment by milkshake — January 18, 2012 @ 2:41 pm

  3. What’s the solvation power of MTBE for Grignards? Got any good references on that?

    Comment by Chemjobber — January 18, 2012 @ 5:23 pm

    • I don’t have a reference but I faintly remember that there was some problem with making Grignards in MTBE. Also, Grignards can be done in Bu2O.

      Comment by milkshake — January 18, 2012 @ 5:50 pm

  4. Current solvent of choice for grignards, 2MeTHF (astounding, but you have to consider the effects of an asymmetric ether as a ligand for the reagent). Old industrial standard, toluene (you need a solvent that has a minimum azeotrope with water and a good boiling point). I share with Laird the penchant for isopropyl acetate, and about chlorinated solvents, 1,2-dichloropropane is my standard replacement for DCE/chloroform (it shares with 2MeTHF a long life cycle).
    The bad reputation of DIPE is due to incorrect use of this solvent: as long as you don’t distill it (obtaining unstabilzed DIPE), I would classify it as a safe solvent.
    Acetonitrile can’t replace DMF. Ok, DMF is not green, but you can’t get rid of it in many reactions, notably in Rosenmund-Von Braun and Ullman reactions: you can’t be always lucky and find Buchwald-like conditions for your amination/cyanation.

    I note that the Pfizer table doesn’t talk about water: phase transfer conditions are a standard for some scaled up acylations and nucleophilic substitutions, and water, if you can use it, is the cheapest and more avaiable solvent you can think about.

    Comment by processchemist — January 19, 2012 @ 5:14 am

    • I agree water seems attractive but aqueous waste has to be decontaminated or burned and if you have an untreated water waste stream going down the drain with enough quat ammonium salts or even with little bit of plain THF, it will wipe out the active microbial cultures in the waste water treatment plant and degrade its performance. (One reported incident like that and EPA or a similar state-level agency cites you and then keeps on your tail forever). And incinerating water will cost fortune because the heat capacity/evaporation enthalpy is so high. Its like with DCM, water costs nothing to buy but is difficult to recycle or dispose. And quat ammonium phase transfer reagents will have detergent properties so the phase separation or evaporation might not be fun.

      Ullman – I have actually done stoechiometric Wurtz-like cross-coupling reactions involving activated copper metal insertion and interestingly the best solvent for the reaction turned out to be anhydrous DMSO. It must be a bit expensive on scale though to work with DMSO.

      1,2-dichloropropane: this is an interesting suggestion, but I don’t think I would give it a try because the government regulators have particular concern about it, it has unpleasant toxicology and even though it used to be popular and cheap it has gotten very expensive just recently here in US. I am currently working on polymer-related projects, some of them are very process oriented, and the issues that typically come up are like: what would be the best solvent to run the reactions in it, to extract or to precipitate the product within a manageable volume or to re-dissolve the polymer and azeotrop it – and can it be done safely and cheaply enough on kilo scale? Some less common solvents for example that we were looking into included n-chlorobutane (popularized by Merck process) and dimethyl carbonate (VOC exempt). Most of the less-common solvents we looked into sucked for one reason or another.

      Comment by milkshake — January 19, 2012 @ 6:41 am

  5. 1,2-dichloropropane:
    LD50 Oral – rat – 1947 mg/kg
    LD50 Inalation – rat – 8 h – 14000 mg/m3

    And not carcinogenic. 1,2-DCP defects: quite high solubility of the water in the solvent (but this is a plus, if you’re quenching a reaction with POCl3). Anyway it has a low boiling azeotrope with water, so drying is not such a problem (honestly I never used it in conditions where ultra dryiness is required). But for final steps of anything for human use, you have to go for FDA class 3 solvents, so your choices are limited.

    I dislike DMSO, also if a couple of years ago we worked on PSD/polimorph optimization of the crystallization of an investigational API from DMSO/MTBE (the only possible solution!). Ok, is a class 3 solvent, but between DMSO and DMF, I always go for the last one if I can (from my experience I remember only a nucleophilic aromatic substitution that was great with DMSO and lousy in DMF).

    I never experienced problems in water treatment: the only golden rule for water streams was “no chlorinated solvents content”, but I’m not an expert in the field.

    Comment by processchemist — January 19, 2012 @ 7:23 am

  6. I forgot…. dimethyl carbonate can be a methylating agent, if used with bases…

    Comment by processchemist — January 19, 2012 @ 7:55 am

  7. Interestingly toluene has variable status (from ok to no way) depending on the country. Aqueous waste is an issue if there is high salt or the microbes in your bio filter can handle the trace components (like polar aprotic solvents….).

    Comment by Milo — January 19, 2012 @ 12:47 pm

    • I think Derek Lowe was joking once, that toluene is the greenest process solvent (tolerable tox, not too high volatility, easy recycling, low cost, inexpensive disposal by incineration).

      Milo, are you still in the fragrance & flavor business? What were the limits and preferred solvents when you were required to develop a new fragrance product with cost per pound below management-dictated cutoff (if I remember correctly you had mentioned something really low – like 50 USD per pound) free of solvent and other problematic impurities?

      Comment by milkshake — January 19, 2012 @ 1:28 pm

  8. I was wondering what you meant by DMSO being “bio-innocuous”. I’m a lousy organic chemist, but I remember McMurry (“Organic Chemistry: A Biological Approach”) warning “Dimethyl sulfoxide (DMSO) is a particularly well-known sulfoxide that is often used as a polar aprotic solvent. It must be handled with care, however, because it has a remarkable ability to penetrate the skin, carrying along whatever is dissolved in it” (pg. 528). I’m not qualified to say myself, but just wondered what you knew from your experience.

    Thank you for the post. I always learn a lot, particularly ones like these that have to do with techniques or general methodology.

    Comment by mwh — January 20, 2012 @ 12:06 pm

    • well DMSO helps things to soak through skin fast though several other solvents will have similar effect. DMSO is a naturally- occurring substance, it has famously low toxicity, people were even using it to rub it on their joints as arthritis relief. DMSO exposure gives you stinky breath as DMSO is metabolized to foul smelling volatile Me2S that is exhaled but this effect is fairly harmless. DMSO has been used in some drug formulations, even for IV injections.

      Comment by milkshake — January 20, 2012 @ 12:22 pm

  9. So if I’m understanding you correctly, it’s not the DMSO that’s a concern, but rather the substance dissolved in it that might be dangerous?

    Thank you.

    Comment by mwh — January 20, 2012 @ 12:38 pm

  10. What about fluorobenzene toxicity, versus the trifluorotoluene? I recall PhF bp being mid-75s I think? (ºC, of course!) 😀

    Comment by HPCC — January 24, 2012 @ 9:16 pm

    • I dont know, its a good point, but I never heard of using PhF as a solvent. I think PhCF3 is a bit cheaper than PhF – cheap enough to be used by pharma process groups – but that could be just a matter of production scale… I know that process folks are not opposed to the use of chlorobenzene

      Comment by milkshake — January 24, 2012 @ 9:36 pm

  11. The reason why I posted this was that I recall one of those Nicolaou methodology papers where PhF happened to be their best solvent – similar properties to dichloromethane. Used it once – smells just like benzene. Those fluorides really mimic hydrogens well!

    Comment by HPCC — January 24, 2012 @ 10:08 pm

  12. I meant famously infamous Nicolaou IBX papers…

    Comment by HPCC — January 24, 2012 @ 10:08 pm

  13. What about NMP as a replacement for DMF?

    Comment by Kevin — February 1, 2012 @ 2:15 pm

    • I think NMP is pretty good (I hate purifying/drying it/removing it from the reaction mix, but thats another issue). But for some reasons NMP seems to be an enviro concern also, it was one of the “solvents to find alternatives to” in the Innocentive challenge. The problem is, none of these “greener” solvents is going to be totally non-problematic once they get into the groundwater and re-optimizing a chemistry for some more degradable but chemically inferior solvent is much harder than preventing the actual spill.

      Comment by milkshake — February 1, 2012 @ 3:52 pm

  14. I heard washing away NMP from EtOAc phases with 1 M LiCl in H2O is the magic trick. But then again, is having lots of LiCl in waste water better or worse than having NMP around?…

    Comment by HPCC — February 1, 2012 @ 10:50 pm

  15. … by the way, is there a procedure to remove DMF comletely?
    Regards from Praha,

    Comment by Zdenek — February 3, 2012 @ 5:39 am

    • I think the answer depends on the scale on which you are doing the workup. In the lab, on a gram scale, you can usually drown the reaction mix with a large excess of water, extract it with solvent into which DMF has a poor partition against water – for example with ether or ethyl acetate – then then you would re-wash your organic extracts few times with water or half-saturated NaCl solution and you are ready to go. You can even add some LiCl to these aqueous washes because Li+ salts have high affinity to DMF and will help to keep it in water. On process scale it gets more tricky because you are constrained by the volumes that your equipment can handle, you would want to use smaller solvent volumes if possible and generate less water waste stream because incinerating water waste is very expensive (and you cannot send untreated water+DMF+EtOAc waste into sewer).

      Also, on really large scale process people often do not use real evaporators, they prefer “solvent replacement” by evaporating stuff directly into the reactor under vacuum and then adding a higher-boiling reaction solvent at the end of evaporation into the concentrated mix, to drive off the lower boiling solvent that was used for extraction. In such workup DMF carried over to the next step could be quite a challenge so one would need to perform a careful optimization of each workup operation

      Comment by milkshake — February 3, 2012 @ 12:51 pm

      • Somewhat old post but since you appear to be answering questions… The water + DMF + EtOAc problem appears somewhat simple: Heat with some base for some time, convert DMF to dimethylamine and formate; ethyl acetate (not that toxic anyway) gives ethanol and acetate. Adjust pH as necessary. I imagine the critters can deal with that?

        Comment by Andrew — March 1, 2012 @ 7:36 am

    • If you are doing a rection on a 100g scale – it helps to pump most of the DMF off before the workup. Have you tried to connect your rotovap to a common oil vacuum pump? It works quite well with Buchi rotovaps that have dry ice condenser finger, if the seals are good. I had a 3-way stopcock on a rotovap and after evacuating the mix with a normal teflon pump (to drive off the dissolved gases and very volatile components) I would switch the rotovap to oil pump vacuum and I could routinely get down to 0.5 Torr on my rotovap. This way I was evaporating DMF very fast, at room temperature, and DMAc from a 30C and DMSO at 40C water bath.

      You need to add one additional cold trap in between the rotovap and the pump though, to protect the oil

      Comment by milkshake — February 3, 2012 @ 1:09 pm

  16. Thank you for your detailed answer.

    Comment by Zdenek — February 4, 2012 @ 1:33 pm

  17. Hi MS,

    I am trying to prepare acyl derivative of an aliphatic amino alcohol and I am planning to use acetic acid (excess) and metal oxide (0.3 equiv) for this. I am not sure about the scalability of this. I am trying to scale this up.

    After the completion of reaction,work up involves filtration, followed by extraction using ethyl acetate and bicarbonate wash.

    My only worry is that the product (N-acylated amino alchol) has water miscibility.

    Alternatively, Distillation could be an option and bp of the product is <190 oC and the bp of the SM and acetic acid are well below that.

    It would be really helpful if you have any suggestions for the scale up



    Comment by marto — February 23, 2012 @ 12:43 pm

  18. I prepared some ethanolamine amides by direct dehydratation of ethanolamine excess with an acid without a solvent but these were large carboxylic acids (such as palmitic acid, 70% Y based on starting acid, after recrystallization) and the conditions for the dehydratation were harsh (reflux at 190-200C) so I would not recommend it.
    I remember that aminolysis of esters with unsubstituted ethanolamine proceeds fast – for example when you mix it with gamma butyrolactone without a solvent you get a thermal runaway but the aminolysis is clean. So maybe if your amine is reactive enough (aliphatic NH2) you should try to reflux it with excess of ethyl acetate or even better with methyl acetate. If this does not do work you can always heat up your mix higher in a pressure flask, or you can add a small amount of sodium methoxide as a transesterification catalyst (the O-acetyl intermediate will quickly migrate from O to N) and in case that you obtain O,N-diacylated impurities you can add MeOH with some catalytic MeONa at the end of the reaction to bring about O-deacylation by the transesterification onto MeOH. The advantage of this approach is that sideproducts are volatile so presumably you could avoid aqueous workup and just concentrate the mix and distill your product from the residue left in the reaction flask.

    Comment by milkshake — February 23, 2012 @ 3:50 pm

  19. I want to know, which slovent preferentially can be used in enolate chemistry? and also in a reaction the ether can be replaced by which solvent?

    Comment by shabbir — December 3, 2012 @ 7:58 am

    • it depends what kind of enolate chemistry. If you have an acidic substrate and the enolization is promoted by a Lewis acid like TiCl4 or Bu2BOTf and a tertiary amine like NEt3 is used as a base, then the typical solvent is dichloromethane and it can be replaced with a number of other (non-coordinating) solvents stable to Lewis acid.

      But if you are talking about enolization with a strong base like BuLi or NaH, THF and related ether solvents are your best bet (with NaH you can add co-solvents like DMSO and deprotonate with dimsyl anion, and also DMF as a coordinating additive also works but there are thermal runaway risks). tBuOK and tBuONa as a base could be used in toluene as a solvent. Additives like 18-crown-6 increase reactivity of potassium enolates. And so on. You would need to tell me more about what you are trying to do and what problems arre you trying to solve.

      Comment by milkshake — December 3, 2012 @ 2:48 pm

  20. Hello Milkshake! I was looking for replacements for DMF and I came across your site. I have a question regarding propylene carbonate, you said you wouldnt heat it with reactive amines I would appreciate it if you could give me a reason. Thanks!

    Comment by C.A — August 1, 2013 @ 1:29 pm

    • I would worry about aminolysis of the cyclic carbonate, with formation of urethane. Especially when using a reactive primary or secondary amine like benzylamine or pyrrolidine. Please note that DMF has this problem also to some degree – old mixtures of piperidine with DMF contain considerable quantity of formylpiperidine and heating with DMF is a known method for formylating primary amines. For what application are you trying to find this replacement? Can you use dimethylacetamide? Or even acetone?

      Comment by milkshake — August 1, 2013 @ 8:34 pm

      • HI MS! thanks for your reply. Im mixing lead halides with Methylammonium iodide using DMF or Proplyene carbonate via heating. Im trying to figure out whether propylene carbonate would make a good replacement solvent

        Comment by CA — September 6, 2013 @ 9:44 am

        • it might – but it has much higher boiling point so it is harder to remove by distillation, and it also has higher reactivity to nucleophiles. Please what is the application – and are you going to heat the mix? Can you also try acetone as a solvent?

          Comment by milkshake — September 6, 2013 @ 11:23 am

  21. Hii….In friedel crafts acylation, using alcl3 as catalyst which solvent other than DCM, carbontetrachloride and nitrobenzene can be used?? plz suggest.

    Comment by Gayatri — August 3, 2013 @ 10:10 am

    • 1,2-dichloroethane and carbon disulfide also work. (If you have a very electron rich partner like anisole, you may also use chlorobenzene or o-dichlorobenznzene). By the way, carbon tetrachloride is not a good solvent for Friedel-Crafts

      Comment by milkshake — August 5, 2013 @ 3:13 am

  22. I am Upashana (female, age 26, married), a PhD student at Indian Institute of Technology, Delhi, India.

    I will be working with DMF in lab. I have planned to coat textile material with thermoplastic polyurethane (30% solution) dissolved in DMF & then evaporation of DMF will be done in air oven & vacuum oven. I may need to work with 500 ml DMF per day. In our lab DMF (made by ‘Fisher Scientific’ & ‘Merck’ ) is available and I may need to continue the use DMF for more than three years.

    Is there any health risk associating with my work? Should I work with this much amount of DMF ? I am unable to get an idea of how much DMF I may inhale & what will be the effect. I am anxious too. Should I think for any alternative solvent like MEK?

    Kindly advise me.

    Comment by upashana — September 20, 2013 @ 7:03 am

    • I don’t know – I am not a professional toxicologist. DMF is in the “unhealthy” category rather than toxic or highly toxic. That is to say, occasional exposure to small quantities is unlikely to make you ill. But there could be some long-term cumulative toxicity on reproductive organs and fetus – especially if you work with it every day without taking precautions (protective gloves – DMF gets readily absorbed through skin) and the odor of pure DMF is quite faint and unremarkable. If you are likely to become pregnant during this time, then I would say that this is not a good work for you. Why to take the risk.

      Comment by milkshake — September 20, 2013 @ 11:39 am

  23. i m manager(r&d) in a paints ltd. company i am usinng DMF and DMSO for manufacturing of laquers for footwear soles to enhace ADHESION of laquer cause bethout using dmf and dmso we have adhesion problem.these products are alergic to skin during laquer application it reacts to human skin.can you suggest any replacement for this purpose can we use DMAC?

    Comment by mohit pathak — November 7, 2014 @ 2:42 am

    • I would try to re-formulate using DMAc or NMP, they have slightly lower toxicity than DMF. The skin reaction can be also caused by other factors though (i.e. other components present in the lacquers, that just get carried into skin in the presence of DMF or DMSO) so you may still need to skin test your newly reformulated lacquers.

      Comment by milkshake — November 7, 2014 @ 2:49 am

  24. I am research student in Sri lanka I am using DMF to make a quaci solid electrolyte for solar cell,I want to know about similar solvent for the DMF for replace the solvent because I have problem with evaporation
    Please help me
    Thank you

    Comment by srinathmadhushanka — January 1, 2015 @ 2:41 am

    • The closest ones would be dimethylacetamide (DMAc) and N-methylpyrrolidine (NMP). Other related solvents good to try would be sulfolane, tetramethylurea, 1,3-dimethyl-2-imidazolidinone (DMEU) and propylene carbonate

      For a solar cell, you probably need a fairly photo-stable solvent and DMF decomposes readily under UV light.

      Comment by milkshake — January 1, 2015 @ 6:13 am

  25. Hii…..
    why we can not use acetonitrile as a solvent in manufacturing process of API.

    Comment by anil gawali — December 12, 2016 @ 7:02 am

    • you can use it, if there is a good reason for it – and if the starting materials dissolve in it and the reaction is chemically compatible – but you have to be very thorough with removing acetonitrile solvent traces from API: as one of the more problematic class II solvents, the permissible MeCN residue in API must be below 400ppm, which is 0.04%

      This means if you are using MeCN in API manufacturing, the final polishing steps like re-crystallization or precipitation should be done in some more innocuous solvent that has high allowable limits for residue – i.e. 4000 ppm for EtOH

      Comment by milkshake — December 12, 2016 @ 8:49 am

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at

%d bloggers like this: