Freezing wood and parchment

Walter Henry <whenry [at] lindy__stanford__edu> writes

>We have a largish choir book (late c18), manuscript on parchment
>bound in wood boards and discovered signs of recent insect activity.
>We have a blast freezer and for many materials would deal with this
>by rapid freezing to about -50 deg F, but faced with this object, my
>confidence flags and would be most grateful for advice, anecdotal
>experience, cautions, exclamations of horror or delight, or failing
>that, pointers to literature.

It seems inevitable that Canadians are familiar with freezing our
things. I got asked about freezing even before pest control. Since
my CCI colleague Tom Strang has written extensively on low (and
high) temperature pest control (Collection Forum, vol 8, no 2, 1992,
pp41-67) we have, of course, double-checked our knowledge in this
area. I would like to briefly summarize the M.L. Florian article for
those unable to find it easily, and then build on it. (Leather
Conservation News, vol 3, no 1, Fall 1986, pp1-13,17).

This is a rather long response, so the executive summary is: chill
out most artifacts with confidence, especially those typically at
risk from pests but, coatings prone to cracklure, on wood, may
crackle a little bit more by -50 deg C.

As Florian points out, the most compelling evidence that freezing
will not harm most artifacts is anecdotal, and massive. (Bad
anecdotal evidence may bad, but good anecdotal evidence is still the
best.) No craft wisdom or modern observation that we know mentions
problems due solely to low temperature, not in the artifacts of
Europeans or those of our First Nations, many of which have seen -30
deg C routinely, -50 deg C in the high north. Michael Gates
(conservator, Parks Canada, Dawson City, Yukon) has monitored his
own humidistatically controlled storage (ambient temp. to a little
above) for two decades and reports no mysterious effects. (Yes,
vinyl siding shatters if you kick it in January, I will get to
that.)

As to scientific explanations, Florian provides some of the
references for moisture content work in collagen and wood, but I
think it fair to say that the following consensus has emerged in
several bodies of literature, both theoretical and applied, for the
behaviour of water in porous adsorptive materials below 0 deg C:

At waterlogged conditions, the liquid water simply freezes, i.e.
turns to ice. Some may migrate from high solute regions before doing
so, or from smaller pores (see below). In the process, macro- and
micro-structure may get damaged. This is the issue for foods,
floods, wet-site archaeology, and mortality, but it is not the issue
here.

At 75%-100% RH, something similar to freezing can still occur because
moisture is held in micropores as liquid (sort-of). The freezing point
is lower in smaller pores than larger pores (the "Kelvin" equation
based on pore diameter gives the depression). This gives rise to an
"icing-out" phenomenon known for both organic and inorganic pore
distributions: the water in larger pores will freeze first, and if they
aren't full, they suck up water from the smaller pores. Thus Litvan
proposes that the difference in frost resistance of things like bricks
or stone depends on a sufficient volume in the large pores to
accommodate all the small pore water without filling. But that isn't the
issue here.

Below about 75% RH (and aside from systems with lots of solutes) the
popular but simplistic phenomenological model of water menisci in
little pores gets into trouble since the adsorption models yield
water only 1 or 2 molecules thick, hardly enough to call a meniscus
(much of the water at these levels is ABsorbed within the various
polymer constituents). This is the area with the least data, but I
think this is so because experience has never shown it to be
relevant to material durability. The only two papers I ever found on
the state of water at these conditions were both NMR (nuclear
magnetic resonance) studies. This technique is very sensitive to any
change in the molecular neighbourhood of water molecules. In wood,
A.J. Nanassy concludes (Wood Science, vol 11, no 2, Oct 1978, pp.
86-90) that "starting at about 0 deg C water molecules are
approaching a rigid position. This occurs below about -30 deg C..".
This study went to -60 deg C and found nothing sudden. The Russians
Flyate and Grunin (Zhurnal Prikladnoi Khimii vol 47 no 12, Dec 1974,
pp 2739-2741, English transl. by Plenum, NY, 1975) studied low
temperature effects on paper pulps because they knew freezing of
high water content pulps reduced sheet strength. They looked at NMR
signals before, during, and after -21 deg C, and found definite
changes if equilibrium moisture content (EMC) was 16% or above, but
the 3.4% and 9.2% EMC samples were unchanged. Thus freezing, i.e.
absorbed moisture doing some kind of phase change (like ice) is not
the issue.

What about the endless discussion over the low temperature
correction to EMC at fixed RH, a perennial art in transit topic?
Stolow showed the wood and cotton data to our field 30 years ago
(Controlled Environment for Works of Art in Transit, Butterworths,
1966). To keep EMC constant, take the ratio of his beta/alpha
coefficients, for wood and cotton: 3% RH drop for each 10 deg C drop.
He noted that since silica gel has negligible correction, it held RH
well. Toishi pointed out we should stabilize an artifact's EMC, not
RH, so buffer like with like. Thomson found where the opposing
effects of dead air in the enclosure and the artifact meant the two
cancelled out. I pointed out oil paint appears to have no
temperature correction, so you can't keep both wood and paint at
constant EMC or RH, pick one only (chapter in Art in Transit,
National Gallery of Art, Washington DC, 1991, pp223-248). M.
McCormick-Goodhart has a nice variation on this recently (Journal of
the Society of Archivists, vol 17 no 1, 1996, 7-21. journal): given
gelatin's tendency to get soft and sticky at 80% RH at room
temperature, and its large temperature correction, then if 60% RH is
a safe upper limit at room temperature, then this becomes 40% RH in
cold storage. None of this is the issue here either. Why this
irritating preamble? Because these sidebars always rear their ugly
heads when "freezing", or more correctly, low temperature, is
discussed.

I think the issue is shrinkage and embrittlement. If this seems
little progress in twenty years, let me expand. Conclusions first:
Tom Strang and I are convinced that shrinkage and embrittlement does
*not* cause significant deterioration to the great bulk of historic
artifacts that need pest control, and any risk that may occur is
small compared to even a 1% probability of pest infestation if the
artifact is not cooled. (All full citations for most of the
following facts are in the article in Art in Transit cited above).
First, assume cooling in a closed bag, so we can ignore the RH
temperature thing, and just look at fixed EMC cooling. Shrinkage due
to temperature in amorphous polymer media below the glass transition
(varnish, oils, glue, gums) is roughly uniform in the glassy region:
about 0.7% for 70 deg C drop, sometimes less. Pigmented media will
experience about half this, 0.4% for 70 deg C. As given in the Wood
Handbook, USDA, Washington DC, 1987, or discussed by M. Richards in
his chapter in the Art in Transit book cited above (pp 279-297) wood
has shrinkage across the grain of about 0.25% for 70 deg C, along
the grain of only 0.03% (due to its crystalline cellulose). Estimate
paper and parchment as slightly less than crossgrain wood, so 0.2%
for 70 deg C. Given the structure of old books and most other
artifacts, a few significant differentials arise: paint and varnish
on stretched canvas, varnish vs wood along the grain, paint vs wood
along the grain, crossgrained wood. These all give upwards of the
full response of medium vs immovable boundary, i.e. 0.7%. To place
this in perspective, it is the same strain that is caused in these
same systems by a drop from 60% RH to about 30% RH. These worst
differentials are just at the verge of fracture strains in brittle
glassy amorphous polymers. (Mecklenburg et al (CAL) provide more
numerical detail on this cold temp. explanation for crossgrain
cracks in coatings in the Painted Wood Conference, and I believe the
proceedings are due out any day. M. Richards (NG, USA) also gave his
data and this model at the recent Getty Panel Painting symposium.
Patrick Albert at the Centre de Conservation de Quebec is planning
some experimental study of the low temperature effect on wood
varnish cracklure). For other material combinations, the shrinkage
differentials are much less risky: Paint on parchment, ink on
parchment will experience less than  0.2%. Even wood across the
grain vs brass battens seems OK: 0.25% - 0.14% = 0.11%. Luckily, most
artifacts in dire need of pest control experience negligible
differentials: textiles, fur, feathers, skins, single pieces of
wood, all possibly with lean paints (no medium no problem).

So, getting back to Walter's book: if the wood boards are not
battened by cross-grain wood strips or firmly attached metal, they
are not at risk, as long as the bagged book is wrapped a couple of
times with wool blanket or the equivalent to reduce temperature
gradients on the boards. (Bag and cuddle! but not so much the bugs
can adapt!) If the boards are heavily varnished, I have to say there
is a small but real risk cracklure will increase. There is only a
tiny risk that any illumination or ink on parchment or paper will
crackle more. (They will be at much higher risk in anoxia treatment
if the humidification fails.) They are certainly likely to crack and
delaminate if a page is bent when cold, but that would be silly.  If
you want to monitor a complex artifact such as this before and after
cooling, I would certainly suggest you focus on existing cracklure
and cross grain wood joints for careful documentation, e.g.
macrophotography. If anyone collects such data, please share the
information. I would be happy to pool and share the info. (I
research this stuff, and I'm coordinator of the ICOM-CC Preventive
Conservation Working Group. If it's embarrassing, send it to me
anonymously!

One final note, Mary Hough, who is currently preparing samples at
CCI for a research project on crack repair in contemporary
paintings, has made very nice glassy cracks in one week old lead
white oil paint on acrylic on cotton duck, simply by bringing them
down to -10 deg C, and hitting them from behind with a sharp edge.
These tightly stretched paintings did not crack from being chilled
alone, and the blow would not have done much at room temperature.
Being chilled and then being hit, was fatal, as it was supposed to
be.

                                  ***
                  Conservation DistList Instance 10:37
                 Distributed: Sunday, October 13, 1996
                       Message Id: cdl-10-37-001
                                  ***