High Molecular Weight PEG for Basketry


Introduction: PEG treatments for waterlogged archaeological basketry at the Alaska State Museum and published treatments for such basketry suggest that use of mostly low molecular weight PEG was not enough to impart the stability needed for study and exhibition.  This experiment investigated the use of various concentrations of high molecular weight PEG.  Testing included impregnation in both room-temperature solutions and solutions kept in a 60degree F oven.  Informed by these results, the Alaska State Musuem decided to proceed with treatment using 55% PEG 3350.

Catalog No:  95-12-6

Object: Basketry fragments

Culture:  Northwest Coast Native, possibly Tlingit.

 Description:  Sixteen samples from 95-12-6, a large group of approximately 300 fragments (mostly knots.)  Artifact described by archaeologists as spruce root webbing or semi-rigid netting.  Reconstruction of this artifact is unlikely, and the large number of similar small fragments give good comparative study samples.  Material is most likely spruce root.















ASM 95-12-6 semi-rigid knotted netting

ASM 95-12-6 semi-rigid knotted netting









Condition:  All fragments stored in distilled water in a refrigerator since their discovery in 1995. Biological growth has occurred in the past. Water was rarely changed. Since 2006, little biological growth has been noted. All fragments are fragile.

Background/ Reason for Report:  Found on South Baranof Island in 1995.  Another basket from this site, 95-12-1, was already treated.  It was C-14 dated at 4,450 years BP and the material was (with some controversy) identified as hemlock and not spruce root.  20% PEG 400 and 5% PEG 4000 was the protocol used by the ASM to treat 95-12-1 as well as the ancient Thorne River spruce root basket from a different site.  The physical appearance of both those treated baskets is pleasing, but they are still too fragile to be exhibited or handled easily.  They are very flexible and shed fibers readily.  Perhaps the PEG concentrations were too conservative?  This experiment intends to test the theory that an increased amount of higher molecular weight PEG might yield better results.


Low molecular weight PEG (PEG 200-600) is thought to penetrate more deeply into the secondary cell wall and the smaller spaces in the wood than higher molecular weight PEG.  It is also more mobile and hygroscopic.  If too much is used, the surface will look wet, feel soft, attract dust, and be humidity-sensitive.  High molecular weight PEG (PEG 1500-6000) does not penetrate the secondary cell wall because the molecule is too large, but it acts like a filler, impregnating the lumens and interstices between the cells.  Addition of higher molecular weight PEG is thought to be helpful if the wood is more degraded.  Too much high molecular weight PEG can leave white crusts on the surface, result in a heavy artifact, and be harder to dry.  A combination of high and low molecular weights of PEG is often the solution, but it can be tricky to determine the right mixture for solid wood, and basketry is even more challenging.  Higher molecular weight PEG is thought to cause damage if used on wood with fairly intact cell wall structure, perhaps from the force as the hygroscopic PEG pulls water out of the smaller structures where the PEG molecule cannot penetrate? (Astrup 1994, Grattan 1996.)

Low mw PEG is a liquid, and high mw PEG is a solid powder

Low mw PEG is a liquid, and high mw PEG is a solid powder











PEG 540 blend was used in the ASM lab on the Tawah Creek basket at 50% concentration for a month at 60°C in the lab oven with good results.  Heating may speed and enhance penetration as well as the solubility of high molecular weight PEG (Grattan and Clark 1987.)  However, heating was thought to contribute to undesirable darkening for the objects treated at the Ozette site (Cooke, Cooke and Grattan 1993.)  Heat is an accelerant to deterioration and PEG treatments for leather in the literature have mostly eliminated heat altogether for that reason.  Heat is also thought to break down the PEG molecule, and some sources have suggested not heating PEG during the impregnation (Bilz et al 1994.)  Since most of our fragments are small enough to put in the oven, it would be worth knowing if heating provides an advantage. 

Goals of Treatment:  The treatment for these 16 fragments will give data that will set the protocol for stabilization treatment of the other waterlogged basketry fragments in group 95-12.  Several variables affect PEG treatment: deterioration of the wood, species of wood, anatomical structure of the wood (bark, trunk, root etc,) concentration of PEG used, molecular weight of PEG used, duration of soaking, and heating during impregnation.  The proposal aims to address the following questions:

  1. Can we develop a PEG protocol that will make the waterlogged artifacts stable enough for study and exhibition?
  2. What are the optimum concentrations of PEG for this basketry?
  3. What are the optimum molecular weights of PEG for this basketry?
  4. Will increasing the amount of high molecular weight PEG specifically help?
  5. Does heating during treatment provide a benefit?
  6. Will the treated artifact be vulnerable to high humidity?

Stabilization will prevent ongoing deterioration and allow future study and exhibit as part of the permanent collection.  After treatment, additional research into adhesives for mending broken fragments may be undertaken.  The results of the PEG basketry treatments here over the past 15 years would be of interest to the conservation field and should probably be published.  There is very little in the conservation literature about successful treatment of waterlogged basketry.  Additional consolidation of the two ancient baskets already treated in the collection is needed for added stability, which might be achievable with nebulizer mist application.  Adhesives and consolidants will be explored in separate future treatment proposals.

Sample knot, waterlogged, before treatment

Sample knot, waterlogged, before treatment









 Treatment Methodology and Rationale:  Samples fell into three main groups.  Each group had five fragments treated with various concentrations and molecular weights:

1. Samples treated at room temperature

2. Samples treated at 60°C

3. Samples treated briefly at 160°C and then at 60°C.  (This grouping was the result of an error that led the samples to be overheated for approximately 12 hours.)

20% PEG 400, then 20% PEG 3350

Rationale: PEG 400 will enter the secondary cell wall and bond there, while the 3350 will fill in the larger voids and give strength.  This is slightly higher than the concentration of 3350 PEG used previously in the lab, but those examples did not give enough strength. PEG 400 is kept at 20% to hopefully prevent excess from oozing out.

20% PEG 400, then 35% PEG 3350

Rationale: High molecular weight PEG is supposed to perform well on highly degraded wood.  Our basketry is very old and treatment with mostly low molecular weight PEG was not fully successful.  This suggests the basketry may be more degraded than predicted, and may respond better to high molecular weight PEG.

20% PEG 400, then 55% PEG 3350

Rationale: Some references suggest avoiding the eutectic, but others (Jensen et al 2000) seem to suggest that aiming for the eutectic is desirable.  Theoretically, ice crystals form in a way that blocks even distribution of the PEG unless the eutectic is used.  Apparently, concentrations lower than the eutectic also expand on freezing, causing cracks.  At the eutectic, the 9% expansion of ice is counterbalanced by 7% volumetric contraction of PEG.  A medieval log house in Oslo was treated successfully with 50-55% PEG 4000 (Astrup 1994.)  The successful Tawah Creek basket treatment by Scott Carrlee (unpublished, Alaska State Museum) used PEG 540 near the eutectic.

55% PEG 3350 alone 

Rationale: Since the Jensen et al article (2000) seems to suggest PEG near the eutectic is optimal even though other articles specifically indicate the opposite, it would be worthwhile to isolate the PEG 3350 to test this.  Perhaps it simplifies the phase diagram to only use one molecular weight of PEG.  Astrup (1994) and Hoffman (1990) found some success with 50% PEG 4000 in degraded softwoods.

20% PEG 400, then 75% PEG 3500 

Rationale: Several sources report success with high concentrations of high molecular weight PEG for highly degraded wood (DeWitte et al 1984, Kaenel 1994.)










Fragments of similar size with no obvious joins to other fragments were selected for testing and photographed.  Each sample was sewn between layers of nylon mesh screening with polyester thread to hold the fragment securely, allow good circulation around the fragment, and permit handling.  Each PEG concentration was increased incrementally approximately every two weeks.  PEG 400 was increased in 5% increments, PEG 3350 was increased either 5 or 10% increments.  In each case, the concentrations were increased gradually to minimize the risk of osmotic shock from pressure differentials between the fluid inside the fragile wood and the fluid in the container.  The time to reach desired concentration took between 3 and 6 months.  PEG 3350 was supplied as a powder and was dissolved in a bit of the test solution using a hotplate before adding it to the unheated and heated sample containers.  For the unheated samples, the addition of the warm PEG 3350 caused them to be cloudy for two to three minutes before becoming clear again.  All samples were removed from solution at the same time.  Each sample was rinsed in a beaker of distilled water to rinse excess PEG from the surface and gently tamped with KimWipes to remove as much water as possible before freezing.  Fragments were weighed and placed in the freezer (-35°C) to drive off the excess water through sublimation (solid ice directly to water vapor.)  Air drying without the freezer would send liquid water to water vapor, and the strong surface tension of the liquid water contributes to collapse of cell structure as the water evaporates.  Samples were regularly weighed to determine the end point of drying (when weight no longer decreased,) and fragments were all removed at the same time.  After removal from the freezer, fragments were taken out of the nylon mesh and photographed on graph paper while still cold.  They were photographed on graph paper again a week later to check for possible distortion from the evaporation of residual water at room temperature (none observed.)  A month after they were removed from the freezer, they were subjected to 80% humidity for 12 hours. 









Front and back of fragment cold out of freezer, and here fully dry

Front and back of fragment cold out of freezer, and here fully dry










Results and Interpretation: While waterlogged, the wood was dark brown in color.  After treatment, the room temperature samples all turned out very pale beige-gray driftwood-like color, with no obvious color difference with higher concentrations.  60°C oven samples were all pale yellow ochre-grayish in color, but still much paler than most historical basketry.  160°C overheated samples were a rich brown burnt umber color, ironically more like historical spruce root basketry in color, and the higher concentrations were darker.  The untreated air-dried control sample was the darkest of all (dark burnt umber) and extremely brittle, shrunken, and deformed.  Almost all samples had some whitish powdery PEG residue/crusts in the crevices, and this did not seem to increase with concentration, but it was more pronounced on the samples impregnated at room temperature.  The waxy PEG could not easily be brushed from the surface (the brush tended to drag it around) but localized application of ethanol with a brush under magnification seemed to drive the PEG below the surface and improved the appearance.  All impregnated samples were placed in the humidity chamber and raised to 80% RH for 12 hours to evaluate effect of high RH on the concentrations and molecular weights of PEG used.  No oozing or surface changes were observed on the samples or on the blotter paper below them.  While the Alaska State Museum has stable RH, there are less stable locations in Alaska that may wish to exhibit the artifacts after treatment.


UPDATE: From a suggestion by Dana Senge, I repeated the RH test.  The 75% concentrations of the unheated and the overheated samples both oozed, and their surfaces got very dark and waxy.  This did not revert back upon stabilization of the RH, and did not seem to get worse with repeated RH tests.  The heated 75% sample did not ooze or get dark, even on repeated RH fluctuations.


Himdity test at 80% RH

Humdity test at 80% RH









20% PEG 400, 20% PEG 3350

Room Temperature: Weight loss stabilized after 20 days in the freezer, sample lost 22% of its wet weight

Heated: 65 days, lost 22%

Overheated: 40 days, lost 33%

20% PEG 400, 35% PEG 3350

Room Temperature: 40 days, lost 23%

Heated: 90 days, lost 21% 

Overheated: 40 days, lost 25% 

20% PEG 400, 55% PEG 3350

Room Temperature: 65 days, lost 23%

Heated: 90 days, lost 23%

Overheated: 20 days, lost 22% 

55% PEG 3350 

Room Temperature: 90 days, lost 23%

Heated: 65 days, lost 18% 

Overheated: 20 days, lost 22% 

20% PEG 400, 75% PEG 3350 

Room Temperature: 90 days, lost 20% 

Heated: 20 days, lost 10% 

Overheated: 65 days, lost 18% 

Samples after treatment: unheated rather gray, overheated brown, heated yellowish.  Control in upper corner

Samples after treatment: unheated rather gray, overheated brown, heated yellowish. Control in upper corner










Assumptions: a lower percent weight loss during drying means less water was lost and there was less water to lose in the first place because it had been replaced by PEG.  If this is true, all the room temperature fragments had about the same amount of water loss, and therefore a similar amount of PEG penetration.  But at higher concentrations, the heated fragments had less water loss, and thus better penetration of PEG 3350 than the room temperature samples.   Explanations could include better penetration of molecules that may have some size variation through thermal breakdown, better diffusion, possible expansion of wood structure with heat to allow better penetration, or the enhanced solubility of heated PEG (Grattan and Clark 1987.)

The samples with lowest concentrations were distinctly spongier to the touch than those with 55% 3350 and higher, in spite of similar amounts of weight loss.  Surprisingly, results for 20% PEG 400 with 55% PEG 3350 were the same as the results for 55 % 3350 alone. 

Flaws: I may be misinterpreting the Jessup et al (2000) article about the eutectic.  Varying geometry of the shape of the knots made my qualitative comparisons about flexibility and brittleness too subjective.  Sample size was not statistically significant.  Have not done cobalt thiocyanate staining to see which samples have more PEG in them under microscopic examination.   Determining degree of deterioration from the molecular weight of PEG that works is a backwards method.  I began an experiment in July 2009 to try to determine the moisture content of the archaeological basketry compared to historical basketry samples waterlogged artificially in the lab.  Also need to spend some time examining the samples structurally under the microscope.

Conclusions:  The idea that high molecular weight PEG around 55% concentration is useful for highly degraded softwoods is in harmony with the conclusions found by others (Astrup 1994 and Hoffmann 1990.)  The superior performance of the heated 75% solution was based on a single fragment losing a total of only 0.1g of water in the freezer after only 20 days.  Could this be an anomaly?  If 55% PEG 3350 does not give adequate stability for the first basket I treat, I may attempt the higher concentration.  The oozing of the 75% fragments at room temp and with voerheating makes me nervous.

Annotated Bibliography:

Astrup, E.E. “A Medieval Log House in Oslo – Conservation of Waterlogged Softwoods with Polyethylene Glycol.”  Proceedings of the 5th ICOM Group on Wet Organic Archaeological Materials Conference, Portland, Maine. 16- 2- August 1993.  Pub 1994 pp 41-50

Ellen Carrlee’s Notes: Wood was identified as pine (Pinus silvestris) and spruce (Picea abies).  For both highly degraded and less degraded softwoods, 50-55% PEG 4000 was most useful.  She heated at 60°C adding 10% every 8 weeks.  She was getting 3-5 % shrinkage in the 50-55% range, but more above and below it.  Waterlogged wood is already somewhat swollen, however.  Results similar to Hoffmann’s 1990 article on softwoods.  Wood with fairly intact cell wall structure might suffer from shrinkage with low mw PEG alone, perhaps the hygroscopic PEG pulls the water out of tyhe small spaces the molecule cannot fir into?  Mentions that the two-step method is really intended for wood like oak that tends to have areas of both low and high degredation.

Barbour, R.J. and L. Leney.  “Shrinkage and Collapse in Waterlogged Archaeological Wood: Contribution III Hoko River Series.  In book. Proceedings of the ICOM Waterlogged Wood Working Group conference: Ottawa, 15-18 September 1981. ICOM Waterlogged Wood Working Group (1982), pp. 209-225.

Ellen Carrlee’s notes: Shrinkage from capillary tension happens when the free water is pulled by evaporation from the void structure of the wood.  Shrinkage is also caused from desorption of bound water from the cell wall.  Article gives extensive info about which cells collapse in which direction.

Bilz, Malcolm, Lesley Dean, David W. Grattan, J. Clifford McCawley, and Leslie McMillen.  “A Study of the Thermal Breakdown of Polyethylene Glycol.”  Proceedings of the 5th ICOM Group on Wet Organic Archaeological Materials Conference.  Portland, Maine. 1993.  Ed Per Hoffmann.  1994.

Ellen Carrlee’s Notes: PEG from ten year old artifact did not decrease in molecular weight.  Keeping oxygen or air out slows PEG degradation.  Thermal aging increases PEG degradation.  Slightly more degradation of PEG 3350 in presence of iron salts, but not so much with PEG 200.  Less degradation of PEG 3350 in oak than by itself.  BHA (not very soluble in water) helps reduce oxidation degradation of PEG 3350; propyl gallate does not, and also forms colored complexes with iron.  No degradation of PEG with natural aging either in artifacts or by itself.  Impregnation should take place at room temperature to avoid degradation.  Store PEG at its most concentrated.  PEG is reasonably reversible and can be leached from artifacts.  Lower concentrations of PEG 400 (below 40%) oxidize more.  In degradation, are we getting smaller molecules, do they go into the wood, or do they evaporate? 

Cooke, Vincent, Deborah Cooke, and David W. Grattan.  “Reversing Old PEG Treatments of Objects from the Ozette Site.”  Proceedings from the 5th ICOM Group on Wet Organic Archaeological Materials Conference, Portland, Maine 16-20 August. 1993. 1994. pg 92-109

Ellen Carrlee’s notes: Treated with PEG 540 blend c. 1973, hygroscopic, dark, heat used.  Extracted PEG, retreated with 2-step PEG 200 at 15% for 40 days, and PEG 3350 at 9.5% for 20 days, then freeze dried.  Acetone Rosin was found irreversible.  They also consolidated with Polyox 2.5%.

Cook, Clifford and David Grattan.  “A Method of Calculating the Concentration of PEG for waterlogged Wood.”  Proceedings of the 4th ICOM Group on Wet Organic Archaeological Materials.  Bremerhaven 1990.

Ellen Carrlee’s notes: Liquid grade, such as PEG 200 is used to control cell wall shrinkage and waxy solid PEG 3350 is used to give some structural strength to the wood.  Important to know:

  1. Wood species
  2. Actual density of the wood
  3. The normal density of undeteriorated wood
  4. The Moisture content at the fiber saturation point of the undeteriorated wood.

These are the parameters used by PEGCON.  Lower MW PEG such as PEG 200-600 penetrate into the cell wall better than higher MW.  Higher MW PEG is meant to fill in the lumens and has been shown to work better on very deteriorated wood.  Too much PEG 400 results in weeping , soft, humidity-sensitive wood.

Too much PEG 3350 results in heavy wood, harder to dry. Umax is the maximum moisture content, increases with deterioration.  Some woods have a higher Umax naturally than others, so that’s why you have to know the species.  Over 55% PEG mixture might not freeze in the drying process.

Christensen, B. Brorson.  The Conservation of Waterlogged Wood in the National Museum of Denmark.  National Museum of Denmark, Copenhagen, 1970. 

Ellen Carrlee’s notes: Includes history of treatments 1859-1962, and reports on work thusfar on Viking ships from Roskilde Fjord.   Includes PEG 4000 in various solvents but having difficulty with the issue of oak penetration.  Methylene chloride is the best solvent for removing PEG after treatment.  Water is too slow and results in too much rubbing which obscures fine details and rounds sharp edges.  Cold treatment with PEG 4000 as opposed to warm treatment resulted in less collapse of the oak.  Began with 25% and increased to 50% for a year total

DeWitte, Eddy, Alfred Terfve, Jozef Vynckier.  “The Consolidation of the Waterlogged Wood from the Gallo-Roman Boats of Pommeroeul.”  Studies in Conservation Vol 20 no 2 1984.  pp 77-83

Ellen Carrlee’s notes: PEG impregnation with PEG 4000 up to 85% heated to 65C in a steel tank.  Oak with no solid core.

Florian, Mary-Lou and Richard Renshaw-Beauchamp.  “Anomalous Wood Structure: A Reason for Failure of PEG in Freeze-Drying Treatments of Some Waterlogged Wood from the Ozette Site.”  Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981. pub 1982 pp. 85-98 

Ellen Carrlee’s notes: Factors that decrease the penetration of the PEG include:

  • Few and small hardwood cells
  • Inclusions (gum and tyloses) in hardwood vessels
  • Abundant hardwood thick walled fiber cells
  • Short soft wood longitudinal tracheids
  • Absence of ray tracheids
  • Absence of radial and longitudinal resin canals
  • Aspirated and encrusted bordered pits
  • Blind simple pits
  • High specific gravity
  • Ray parenchyma containing resin or extraneous material
  • Reduced tangential wall pitting
  • Extractives in heartwood

Picea sp. (spruces) and Thuja pliacata (Western red cedar) have no tangential permeability due to little tangential wall pitting, blind pits, and resin in the ray parenchyma cells.  Spruce more permeable than red cedar due to resin canals and ray tracheids.  Cedar is vulnerable to collapse of radial walls of the tracheids.  In some woods, large thin-walled early wood tracheids adjacent to thick walled later wood tracheids can collapse and cause uneven grain appearance.  In this case, spruce showed this result.  Roots have many parenchyma cells and can have excessive longitudinal shrinkage.  List of history-chemical tests given.  Many softwoods samples show open enlarged tracheid bordered pits which make the wood abnormally porous. 

Grant, Tara and Malcolm Bilz.  “Conservation of Waterlogged Cedar Basketry and Cordage.”  Proceedings of the 6th ICOM Group on Wet Organic Archaeological Materials York, 1996.

Ellen Carrlee’s notes: Parylene coating on cedar “bark” baskets from Scowlitz site (500-1200 years old)  treated with modified PEG treatment. 20% PEG 400 for 3 months and then freeze-dried.  Came out still too fragile.  Parylene is irreversible but worked well.  Need a special vacuum pyrolysis chamber to apply it.  NOTE: Bilz reported at the 11th ICOM-CC Triennial in Edinburgh Sept. 1996  that further study of Parylene shows that it does not age as well as they thought and it is not recommended for long term conservation.  Also, personal communication with Kathryn Bernick (April 9, 2009) she mentions that the Scowlitz baskets are all made of cedar withes (wood splints) and not bark.  Some of the cordage was indeed bark.

Grattan, D.W. and R.W. Clarke.  “Conservation of Waterlogged Wood.”  In, Conservation of Marine Archaeological Objects. Ed Colin Pearson.  Butterworth. London and Boston. 1987. 164-206   

Ellen Carrlee’s Notes: This is really the most comprehensive description of PEG use and especially its history, but of course it ends in 1987 AND this book is out of print and can be hard to come by.

  If PEG 3350 is too bulky to penetrate the cell wall, osmotic pressure could build and cause collapse.  Microcapillaries in the cell wall thought to be around 10nm in deteriorated wood. Idea that bulking the cell wall without bulking the lumina has better results: less PEG oozes out, more natural appearance.  Freeze drying gives a more even distribution of PEG.  Grades of PEG over 600 need to be warmed in order to be completely dissolved, and the complete solubility of PEG in water is very important.  Beginning impregnation with a low percentage of PEG is important to prevent osmotic collapse.  Up to 30% PEG 400 is needed to bulk the cell wall.  Freeze drying helps to keep it in there.  Very deteriorated wood doesn’t have much cell wall and so the low mw PEG can’t stay in there as well.  At the eutectic, PEG can’t freeze so you have a drying from with solid on one side and a mushy PEG solution on the other side. 

Grattan, David W.  “International Comparative Study; Report”  .  ICOM Committee For Conservation Working Group on Wet Organic Archaeological Materials Newsletter. no 14. Feb 1986. 

Ellen Carrlee’s Notes:  I don’t remember seeing this published as a final thing elsewhere?  Grattan described his work at CCI.  He treated some samples with equal parts PEG 400 and PEG 3350 raised in 10% increments up to 30% v/v.  Frozen at -40C and then freeze dried at -20C.  results were good, but the very deteriorated samples should have had less 400 more 3350.   Second batch placed in PEG 3350 in sequential batches of 10%, 20%, 30%m 40%, and 50% v/v and then freeze dried.  Grattan feels the PEG reach the eutectic and did not freeze, so unfrozen PEG exuded onto the surfaces during freeze drying and some surface collapse.

Grattan, David W.   “Some Observations on the Conservation of Waterlogged Wooden Shipwrecks.”  AICCM  Vol. 12 No 3 and 4. 1986.

Ellen Carrlee’s Notes: Diagram of meniscus for water in a capillary, discussion of capillary tension induced collapse.  Issue of PEG penetration into white oak timbers.  Air pockets might prevent penetration.  For more deteriorated wood, concentrations over 50% are better, and fill the lumens in addition to bulking the cell wall.  Less deteriorated wood does all right with 10-30% which only bulks the cell wall.  Capillary action during drying concentrates PEG at the surface and depletes the core…this can be overcome using concentration over 50%. (ellen: also, if you freeze, doesn’t ice crystal formation influence where the PEG ends up?)  For humidity about 80% all PEGs are affected.  Lower concentrations of PEG can be used with freeze drying than with air drying.

Grattan, David W.  “A Practical Comparative Study of Several Treatments for Waterlogged Wood”  Studies in Conservation Vol 27 No 3 Aug 1982 pp 124-136

Ellen Carrlee’s notes: Water content is a rough guide to the state of deterioration.  For smaller sample, no advantage to heating the PEG 400.  Didn’t help with penetration, shrinkage or cracking, but it did make end result darker.  In larger objects, heat may speed up the diffusion.  PEG 400 can end up soapy and attract dust.  Can melt some PEG 4000 on surface to make it stronger and less absorbent.  PEG 540 is a 50/50 blend of 300 and 1540 MW ave of 500-600…tends to creep out of the wood, looks heavy, dark, waxy.  Can be introduced unheated and unstirred in 10% increments over 18months.  Final concentration 64% v/v. 

Grattan, David.  “A Practical Comparative Study of Treatments for Waterlogged Wood part II The Effect of Humidity on Waterlogged Wood.”  Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981. pub 1982 pp. 243-252.

Ellen Carrlee’s notes: Wood treated with low MW PEG (400 or 540 blend) should not be exposed to RH above 60%.  Below that shouldn’t matter.  Union Carbide PEG 4000 has been renamed PEG 3350.

Hamilton, Donny L. Methods of Conserving Archaeological Material from Underwater Sites.  Nautical Archaeology Program Department of Anthropology Texas A&M University.  1998.

Ellen Carrlee’s notes: Amount of water in waterlogged wood is calculated: weight of wet wood minus weight of oven dried wood divided by weight of the oven dried wood and multiplied by 100 to give % water.  Anything over 200% is considered degraded.  Explanations of other treatments for waterlogged wood are also given.

Hoffmann, Per, Adya Singh, Yoon Soo Kim, Seung Gon Wi, Ik-Joo Kim, Uwe Schmitt.  “The Bremen Cog of 1380: An Electron Microscopic Study of it Degraded Wood Before and After Stabilization.”  In Holzforschung Vol 58 No 3 2004 pp 211-218

Ellen Carrlee’s notes: Two-step PEG treatment for wood degraded primarily by erosion bacteria.  SEM/TEM investigation.  In degraded tissues, all cell types were filled with PEG 3000.   Non-degraded tissues are impermeable to PEG 3000 and impregnated only with PEG 200.  Confirms that PEG 200 goes into the cell walls.

Hoffman, Per.  “On the Stabilization of Waterlogged Softwoods with Polyethylene Glycol (PEG).  Four species from China and Korea.”  Holzforschung. Vol. 44 No 2.  1990.  pp 87-93.

Ellen Carrlee’s notes: Best stabilization of degraded softwoods is 50% PEG 4000 with cross section shrinkage of only 2-4%. 

Hoffmann, Per.  “On the Stabilization of Waterlogged Oakwood with PEG II Designing a Two-Step Treatment for Multi-Quality Timbers.”  Studies in Conservation 31 (1986) pp.103-113

Ellen Carrlee’s notes: Bremen Cog PEG 200 and then PEG 3000.  Two step for thicker woods with outer layer of degraded wood and a core of less degraded wood like oak and pine.  Can even do simple air drying afterwards.  Results with intermediate MW not as good as two step.  PEG 200 stabilizes slightly degraded wood better than PEG 300.  Maybe some lower MW PEG diffuses out during the second step, but some is trapped or adsorbed in the capillaries.  Doesn’t weep until RH is 86%.  Can be air dried, too big for a freeze dryer.

Hoffmann, Per. “On the Stabilization of Waterlogged Oak with PEG – Molecular Size Versus Degree of Degradation.”  Waterlogged Wood Study and Conservation, Proceedings of the 2nd ICOM Waterlogged Wood Working Group Conference, Grenoble, France.  1984.  pp. 243-252.

Ellen Carrlee’s notes:  More degraded oak does better with PEG 3350, while less degraded does better with PEG 200.  PEG 1450 is poorer for both.  A 2-step method is more effective.  For a long time people were afraid to use anything under PEG 1500 for fear it would be too hygroscopic.

Jakes, K.A. and L.R. Sibley.  “Survival of Cellulosic Fibres in the Archaeological Context.” In Science and Archaeology. No. 25, pg 31-38. 1983.* Jeberien, Alexandra and Malcolm Bilz.  “Comparison of Air Dried and Freeze Dried Solutions of Polyethylene Glycol 3350”  In ICOM WOAM Newsletter No. 31 June 2000.

Ellen Carrlee’s notes: They expected all to line the walls and bottom of beakers.  Air dried samples had consistency of beeswax and occupied only the bottom the beaker, might have lined walls if they were more porous?  Freeze dried samples had a powdery matrix appearance and occupied the full volume that the frozen water/PEG mix had originally occupied.  For the freeze-dried, when they were returned to low temperature, higher concentrations that had been freeze-dried (20% and above) showed some concentrated PEG solution at the bottom of the beaker that had not been able to get past plugs of PEG.  5% did not have that, so water must have escaped.  

Jensen, Poul; Grethe Jorgensen, Ulrich Schnell.  “Dynamic LV-SEM Analyses of Freeze Drying Processes for Waterlogged Wood”  Proceedings of the 8th ICOM Group on Wet Organic Archaeological Materials Conference, Stockholm, 11-15 June 2001 Pub.  2002 pp.319-333

Ellen Carrlee’s Notes: Looking at distribution of PEG in wood with Low Vacuum Scanning Electron Microscopy (can avoid sputter-coating of SEM and the bias it introduces.)  Machine itself acts like a freeze-dryer with wet samples.  Phase diagrams for PEG done by different authors usually don’t agree, but they all agree the eutectic is around 55% (w/w) for all the mw of PEG.  Eutectic temperatures are not in agreement.  Below the eutectic temperature, a solid lamellar eutectic phase forms between the ice crystals.  Phase diagram they suggest for PEG 6000 shows solid PEG as well as a solid PEG/ice mixture at freezing temperatures with a concentration above 55% (eutectic.)  Below the eutectic, all PEG solutions result in 3-9% expansion at temps below freezing.  Ice has 9% expansion, PEG has 7% volumetric contraction.  Is this a primary function of the PEG?  PEG solutions below 55% all expand on solidifying.  PEG concentrated in the later wood and distributed irregularly in the early wood.  Collapse can be avoided if we stay below the eutectic temperature.  When ice forms, the PEG gets more concentrated.  Formation of large ice crystals contributes to uneven distribution of PEG.  Even distribution of PEG is only possible for eutectic concentrations.  PEG has a low affinity for the secondary cell wall (Jensen’s PhD thesis from 1995.)  Heating after freeze drying causes the PEG to aggregate on surfaces of cell wall and give better distribution.  They suggest investigating methods to nucleate smaller ice crystals and thus better distribution.

Kaenel, Gilbert.  “PEG Conservation of a Gallo-Roman Barge from Yverdon-les-Baines (Canton of  Vaud, Switzerland.)”  Proceedings of the Fifth ICOM Waterlogged Archaeological Materials Conference.  Portland, Maine 16-20 August 1993.  Pub 1994.  pp. 143-165.

Ellen Carrlee’s notes: Mostly oak, treated in a tank with PEG 4000 raised from 15% to 85% at 60C over 18months of impregnation.  AD 400.  Slow drying, natural looking finish reported.

Keene, Suzanne.  “Waterlogged Wood from the City of London.”  Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981. pub 1982 pp. 177-180

Ellen Carrlee’s notes: Variety of different hardwoods and softwoods.  PEG 4000 was used up to 90% at 50-60C, but wood came out dry and splintery so they switched to PEG 1500.  Rinsed with hot water, wrapped in towels and dried slowly in the dark.  Objects come out heavy, brittle, but not much shrinkage.  A little darker, and fine detail is preserved.  For freeze drying, PEG 400 left the wood too fragile, 40-50% PEG 1500 or 4000 worked better.  She’s down on acetone/rosin because while other colleagues have had luck with metal handled tools, she’s not had the best luck and working with hot acetone is difficult.

Lindblad, Cecilia and Ingmar Persson. “Polyethylene Glycol/ Polyethylene Oxide: An Overview of the Physical-Chemical properties of PEG/PEO”  Presented at the ICOM-CC Working Group on Wet Organic Archaeological Materials in Amsterdam, 2007.  Not yet published.

Ellen Carrlee’s notes:  Persson generously sent me a copy.  Describes a lot of chemistry of the polymer.  There’s no clear distinction between PEG/PEO.  PEG extremely soluble in water, but solubility of PEO actually decreases with high temps, salt, and strong stirring.  Exponential increase in viscosity with higher molecular weight.  Bonding discussed.  Hydration radius for PEG 4000 at 30C is 50 nm.  High mw PEGs  above PEG 1000 form random coils and low mw PEGS below PEG 600 maybe form stretched aggregate chains?  PEG on objects treated with high temps show signs of decomposition. 

Masuzawa, Fumitake and Yoichi Nishiyama.  “Experiments on the Impregnation of Waterlogged Wood with PEG Part II”  Conservation Science Bulletin.  Hozon Kagaku Kentyushitsu Kiyo.  Vol. 3 1974 pp39-46

Ellen Carrlee’s notes: this is from a literature review in ICOM-CC WOAM Newsletter No 8 Nov 1982.  Size greatly affects time for saturation: 50-60 days for 3cm wood, but more than 120 days for oak of 7.5cm

Masuzawa, Fumitake.  “Change of Waterlogged Wood Impregnated with PEG Along the Lapse of Time.”  Conservation Science Bulletin.  Hozon Kagaku Kentyushitsu Kiyo.  Vol. 3 1974 pp. 52-58

Ellen Carrlee’s notes: this is from a literature review in ICOM-CC WOAM Newsletter No 8 Nov 1982.    Reviewing wood treated with PEG after one year.  Wood treated with PEG 540 blend absorbed moisture, oozed out and there was shrinkage.  PEG 2000 showed radial cracks after 3 months and some expansion.  PEG 3300 showed no great change.  Based on other articles, they were probably looking at oak.

Murray, Howard.  “The Conservation of Artifacts from the Mary Rose.” Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981. pub 1982 pp. 12-19

Ellen Carrlee’s notes: Wood treated was poplar, oak, ash and softwoods.  Passive holding included spraying with boric acid and double wrapping in sealed polyethylene sheeting (dried out somewhat and got cracking and salt efflorescence within 6 months) or submersion in water (resulting in additional expansion and softening.)  Pre-treatment with 5% EDTA as well as ultrasonic cleaning to remove salts and iron corrosion products, although more than 48 hours in EDTA resulted in notable softening.  PEG 3400 up to 50%.  Mostly doing freeze drying, although getting good results with drying only the surface in the freeze drier and then allowing it to air dry.  Excess PEG removed with hot air, IR lamps, and ethanol swabs.  Reconstruction done with wooden dowels, nitrocellulose adhesive and melted PEG 6000 at 50% as a surface coating.

Rodgers, Bradley.  ECU Conservator’s Cookbook: A Methodological Approach to the Conservation of Water Soaked Artifacts.   Chapter 2: Waterlogged Wood.  Herbert P. Paschal Memorial Fund Publication.  East Carolina University.  1992. 

Ellen Carrlee’s notes: ECU= East Carolina University.  Dr Bradley Rodgers teaches conservation of material from underwater environments. Don’t put items from saltwater directly into fresh water, needs to be gradual or the salt water moves out faster than freshwater can move in, causing collapse (ie pruny fingertips in the bath.)  Specific gravity of several woods are given.  Pits in cell walls have valves, and if these are blocked it is harder to impregnate the wood.  Hard to see w/o SEM.  PEG solution is heated, and you weigh it weekly after you achieve 50% PEG.  Should stabilize at about 20% weight gain.  Might go up to 35%.  If you’re using PEG 3350, maybe even 45%  Wood in good condition sometimes as little as 15%.  Sucrose treatment described.  Brush coating of 50% PEG 3350 to consolidate the outer layer.  Ellen’s note: Bradley’s treatments don’t always match up to AIC standards and ethics, and there was a big dustup over one of his recent publications, an update and expansion of the material in this book.

Singley, Katherine.  The Conservation of Archaeological Artifacts from Freshwater Environments  Lake Michigan Maritime Museum, South Haven, Michigan 1988. 

Ellen Carrlee’s notes: She tends to use PEG 300 and 400 with 4500 applied later hot.  Recommends pretreating wood contaminated with iron salts with 3% perchloric acid, soaked for about a week.  This dissolves iron and opens up the pore structure to increase the penetration of PEG.  PEG 300 at room temp 10% for six months, then removed and wrapped in plastic with regular surface brushing of higher concentrations 3X a week increasing the concentration every two months until 80% reached.  Slow drying at high RH was used.  50% PEG 4500 paste was then applied with hot air.  This layer was thought to prevent oozing out of lower MW PEG.  She also uses 10% PEG 300, soaked for a year and then freeze dried. 

Viduka, Andrew.  Survey of Methods Used by Some Large institutions Specializing in the Conservation of Wet Organic Archaeological Materials.  Report as the 2002 Churchill Fellow: Winston Churchill Memorial Trust of Australia.  2002. 


Ellen Carrlee’s Notes: Viduka visited several labs and wrote up his observations.  Texas CRL sees, among other things, darkening of surfaces with high PEG concentrations as a disadvantage.  Parks Canada was looking at whether PEG modifies the crystal size of water.  Tara Grant at CCI has generally found air-drying to be unsuccessful.  Wooden material from Nydam Mose, 5th C Iron Age site treated in Denmark, is highly deteriorated with only lignin remaining.  NMD is looking into sucrose, wood flexibility post-treatment, Cellusolve method, lots of vacuum freeze drying, review of old treatments, deacidification of low mw PEG timbers, experimental cold air drying.  On the Vasa, acid affected areas have white or yellow build-up and generally appear on areas treated with PEG 400 and not PEG 4000.  Remedial treatment is a spray with sodium bicarbonate and soda with a pH >10 and covered with plastic wrap.  Skuldelev ships were some of the earliest to be conserved with PEG 4000, and 40 years later, they still have a mw of 3900.  Using too high a mw on only slightly degraded wood is a problem.  Too much PEG would give a purple or blue sheen on the surface.  Longer immersion made bigger timbers darker than smaller pieces.  Heating for re-shaping caused darkening of the surface. Glenn McConnachie’s PhD reveals that even well-preserved waterlogged oak from the Mary Rose has 30% volumetric shrinkage after air drying. 

Vynckier, Jozef.  “Examination and Conservation of Basketwork from a Roman Well at Destelbergen”  Bulletin de’Institut Royal du Patrimoine Artistique Vol 8.  1965  (Flemish?) 

Ellen Carrlee’s notes: Linden bark basketry fragments immersed 4 months in carbowax 4000

Watson, Jacqui. “The Application of Freeze-Drying on British Hardwoods from Archaeological Excavations.”  . Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981. pub 1982

Ellen Carrlee’s notes: Fungal degradation in the cell wall contributes to splintering when freeze-dried.  Mineral deposition or replacement in outer layers is common in archaeological wood, especially with iron.  Sometimes these can be seen as casts of material that is now lost.  Chelating agents etc remove the iron.  But iron also interacts badly with PEG.  Early warning in this article of problems with iron salts and PEG. Even if the soil has low iron content, there is sometimes high iron content in the wood. Oak freeze dried after 10% PEG 400 for 4-6 weeks had deep cracks next to the rays, and some distortion and collapse in some small fibers.  Using 10% PEG 400 and 15% PEG 4000 resulted in less cross checking on the surface.  Wood that underwent freeze-drying without impregnation collapsed, and showed crystals (calcium and sulphur) that precipitated out that did not do so when PEG was used.  Pitting details in the vessels that are useful for ID were obscured after treatment with PEG 4000.  Mixed PEG solutions work best when using freeze-drying.

Young, Gregory S.  “Microscopy and Archaeological Waterlogged Wood Conservation.”  CCI Newsletter, No. 6, September 1990.  pp 9-11.

Ellen Carrlee’s notes: General overview of PEG treatment/ research.  Deterioration greatly improves the ability of PEG to penetrate and treat wood successfully, but most excavated waterlogged wood is only moderately deteriorated.  White oak, various cedars, and white ash are hard to penetrate, while aspen, cottonwood, alder and spruce allow greater penetration.  Caution with determining degree of deterioration from thin 3mm cross sectioned wafers of wood examined under the microscope, as the sample has a lot of disrupted wood cells and suggests more cell wall accessibility than there really is. 

Young, Gregory S. and Richie Sims. “Microscopical Determination of PEG in Treated Wood – the Effect of Distribution on Dimensional Stabilization.”  Conservation of Wood and metal: Proceedings of the ICOM Conservation Working Group on Wet Organic Archaeological Material and Metals.  Freemantle, Western Australia Museum. 1987  Pub 1989. pp109-140

Ellen Carrlee’s Notes:  Fluorescence microscopy.  25-35% PEG 200 and 3350.  Found less penetration with larger molecular weights.  Dimensional stabilization correlates with full impregnation of secondary cell wall.  Eight different species tested showed broad range of access to cell wall. 

Young, Gregory S. and Ian N.M. Wainwright.  “Polyethylene Glycol Treatments for Waterlogged Wood at the Cell Level.”  Proceedings of the ICOM Waterlogged Wood Working Group Conference Ottawa 1981.  pub 1982 pp.107-116

Ellen Carrlee’s notes: Shrinkage happens when hygroscopically bound water within the cell walls is lost when wood is below the fiber saturation point (FSP.)  Lower MW PEG is expected to stay in the cell wall, while higher MW PEG bulks the lumina.  Cobalt thiocyanate is a microscopical stain when dissolved in non-aqueous histological grade cedarwood oil.  The oil (cedrene and cedrol) is immiscible with both water and PEG.  It  dissolves the cobalt thiocyanate, but readily gives it up to the PEG in wood sections. 

PEG 400 (35%) in cell wall capillary matrix

PEG 3350 (60%) bulked cell lumina

PEG 540 blend (64%) in both cell wall and lumina  540=1450 + 300 together equal MW mixture

Over 3000 shouldn’t get into cell wall, above 2000, should be impeded somewhat.  MW below 400 penetrate better.  PEG must penetrate the cell wall for best dimensional stability.  Double bulking is good, though.


Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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

%d bloggers like this: