From: gherbert@crl3.crl.com (George Herbert)
Newsgroups: sci.space.science
Subject: Re: Ice or Concrete on Moon?
Date: 9 Mar 1998 15:41:03 -0800

George Herbert <gherbert@crl.crl.com> wrote:
>Can anyone a) remember the references on the
>professor's work on lunar soil cements,

Aha.   Do enough altavista searching and even your own poor memory
can be overcome...  From the Space-tech mailing list digests a:
	http://www.cs.cmu.edu/afs/cs.cmu.edu/user/mnr/st/std135

#------------------------------
#
#To: Richard Schroeppel <rcs@cs.arizona.edu>
#Cc: space-tech@cs.cmu.edu, gwh@soda.berkeley.edu
#Subject: Re: lunar construction materials
#Date: Mon, 16 Nov 92 18:39:06 -0800
#From: George William Herbert <gwh@soda.berkeley.edu>
#
#
#Turning regolith into cement is a proven process.  The best
#method is to take regolith, put it in a pressuretight mold (low pressure)
#and inject steam.  Tests with regolith simulant were highly successful
#(in a vaccum chamber) and small-scale testing with several grams of
#actual lunar soil gave a verifying datapoint.  Regolith in fact
#makes better cement than earth-based materials due to vaccum
#welding of small grains to larger ones, forming a material
#very strong to bond with.
#
#The Japanese materials scientist who did the work is exceptional
#to listen to; he's onto something, if the cost of bringing
#water to the moon comes down (or we find it there), and he's
#very enthusiastic about every step he's taken.  A great lecturer.
#I can get name & references later (they're at home in a sedimentary
#filing system heap).
#
#-george
#
#------------------------------
#
#Date: Tue, 17 Nov 92 21:31:15 -0800
#From: George William Herbert <gwh@soda.berkeley.edu>
#To: space-tech@cs.cmu.edu
#Subject: Lunar Concrete info
#
#Ok, I dug through several piles of materials and found my references
#on Lunar concrete.  The guy who gave the lecture and did the experiments
#was T.D. Lin (affiliation unknown, look for a Japanese materials scientist
#or in American Concrete Institute publications).  I misspoke slightly in
#describing the process he suggested, so let's start from scratch.
#
#What he investegated was making cement from ground lunar basalt, and adding
#in more ground basalt as filler to make concrete.  The process involves
#heat-evaporating some of the basalt's constituents using solar thermal
#processes; a cook-out at 2000 K leaves a residual content of 40% CaO,
#49% Al2O3, and 11% SiO2.  At 2200 K the percentages are 43:53:5.  Both of
#these are within industry tolerances for the production of alumina
#cement on the earth.
#
#The production process for making concrete involves mixing the cement
#with whatever soil is around.  Using a 40 g sample from Apollo 16,
#a small sample was produced to compare the properties of earth and
#lunar material based cements.  The results were very positive;
#the lunar cement/concrete combination is stronger (up to 75 MPa
#compressive yield strength (about 10,000 PSI) versus 5,000 PSI for
#terrestrial concretes) and as tough and durable as earth-based
#concretes.  There was minimal strength loss when samples were exposed to
#vaccum for extended periods of time, which tailed off and is presumed
#to level out at some value about 80% of the maximum.
#
#The curing method was to use steam at mid pressures after drymixing
#the concrete.  Lin proposed generating water by mixing hydrogen with
#Ilmenite from the lunar surface at 800 C; the result is Titanium Oxide,
#Iron (which he thereafter referred to as "rebar" 8-) and water.
#The amount of hydrogen that has to be brought is about 0.3 % (three
#parts in 1000) of the final mass of concrete to be produced, quite a leverage.
#Especially if the iron coming out the side is useful 8-)
#
#Ok, is that a good enough summary? 8-)
#
#-george william herbert
#Retro Aerospace
#gwh@soda.berkeley.edu
#gwh@lurnix.com
#gwh@retro.com coming real soon now
#
#------------------------------
#
#To: George William Herbert <gwh@soda.berkeley.edu>
#Cc: space-tech@cs.cmu.edu
#Subject: Re: Lunar Concrete info
#Date: Tue, 17 Nov 92 21:34:02 -0800
#From: George William Herbert <gwh@soda.berkeley.edu>
#
#Slight clarification.  The vaccum strength loss is presumed to be less than
#20% ... with final strength about 80% of reference value.
#NOT a loss of 80% of the reference value...  8-)
#
#-george
#
#------------------------------
#
#End of Space-tech Digest #135
#*******************

So... it appears that to form optimal cement,
you want to bake out the soil first.  The question is, if you
do not bake it out, do you get materials that still will
hydrate and form (weaker) cements?  The second question is, what
are the reaction rates in the crater-bottom cold traps?

Lin's percentages of constituents in the baked soil
were:
	2000 K 40% CaO, 49% Al2O3, 11% SiO2
	2200 K 43:53:5

I am not a cement expert, but what I remember from my
materials science classes is that the Calcium Oxide is
the chemically active component of cement.  The other
materials (Al2O3 and SiO2) are inert filler in a composite
matrix formed of calcium compounds formed by hydrating the CaO.

Looking at some of the lunar soil compositions and simulant
compositions,  I find the following information on a low-titanium
mare soil (sample 14163) and the JSC-1 simulant modeled on it:
[From: _JSC-1: A NEW LUNAR SOIL SIMULANT_,
David S. McKay, James L. Carter, Walter W. Boles,
Carlton C. Allen, and Judith H. Allton;
http://www-sn.jsc.nasa.gov/explore/Data/Lib/DOCS/EIC050.HTML ]

	Oxide	JSC-1	14163
	SiO2	48%	47%
	TiO2	1.6%	1.6%
	Al2O3	15%	18%
	Fe2O3	3.4%	0
	FeO	7.4%	10.5%
	MgO	9.0%	9.6%
	CaO	10.4%	11.4%
	Na2O	2.7%	0.7%

That gives a roughly 10% CaO concentration in that type
of soil (not typical for crater bottoms, but work with me).
None of the other ingredients to my knowledge are
going to grab hydrogen that voraciously, so the primary
chemical activity should be CaO + H2O -> calcium hydrates.
The conclusion appears to be that at least with some
of the soil types to be found on the moon, the chemical
composition is correct for calcium hydration to form
weak, diluted cements if the thermal conditions and
reaction rates allow it.

As a quick hack of reaction rates, I'm going to assume
that the reaction rates don't have a particular
activation energy of note (note wild assumption)
and just depend on thermal energy.  Assuming that E ~= T^2
we have a data point at say 50 hrs for initial hardening
at 280 K, though reactions will continue for another 30 days (700+ hrs).
The cold traps are assumed to be at 40-50 K; if we assume the
median 45 K, so the reaction rates would be (280^2)/(45^2)
times slower or about 40 times slower.  This is still quite
high, only 2000 hrs (less than 3 months).

I don't remember the formula for determining the actual
energy spectrum at a given temperature, and don't know where to
find the activation energy of the hydration reactions
involving CaO off the top of my head, but I suspect that
the quick hack method above is not accurate and that the
more reasonable energy-spectrum approach is needed...


-george william herbert
Retro Aerospace
gherbert@crl.com